balkian/soil
1
0
Fork 0
mirror of https://github.com/gsi-upm/soil synced 2024-11-23 11:42:27 +00:00
Code Issues Releases Wiki Activity

Merge branch 'mesa'

Browse Source
This commit is contained in:
J. Fernando Sánchez 2023-04-21 15:19:21 +02:00
commit 4e296e0cf1
113 changed files with 8457 additions and 12300 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
.gitlab-ci.yml
View File

@ -20,7 +20,7 @@ docker:
test:
tags:
- docker
image: python:3.7
image: python:3.8
stage: test
script:
- pip install -r requirements.txt -r test-requirements.txt
@ -31,7 +31,7 @@ push_pypi:
- tags
tags:
- docker
image: python:3.7
image: python:3.8
stage: publish
script:
- echo $CI_COMMIT_TAG > soil/VERSION
@ -44,7 +44,7 @@ check_pypi:
- tags
tags:
- docker
image: python:3.7
image: python:3.8
stage: check_published
script:
- pip install soil==$CI_COMMIT_TAG

25
CHANGELOG.md
View File

@ -3,7 +3,30 @@ All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## [UNRELEASED]
## [1.0 UNRELEASED]
Version 1.0 introduced multiple changes, especially on the `Simulation` class and anything related to how configuration is handled.
For an explanation of the general changes in version 1.0, please refer to the file `docs/notes_v1.0.rst`.
### Added
* A modular set of classes for environments/models. Now the ability to configure the agents through an agent definition and a topology through a network configuration is split into two classes (`soil.agents.BaseEnvironment` for agents, `soil.agents.NetworkEnvironment` to add topology).
* Environments now have a class method to make them easier to use without a simulation`.run`. Notice that this is different from `run_model`, which is an instance method.
* Ability to run simulations using mesa models
* The `soil.exporters` module to export the results of datacollectors (`model.datacollector`) into files at the end of trials/simulations
* Agents can now have generators as a step function or a state. They work similar to normal functions, with one caveat in the case of `FSM`: only `time` values (or None) can be yielded, not a state. This is because the state will not change, it will be resumed after the yield, at the appropriate time. The return value *can* be a state, or a `(state, time)` tuple, just like in normal states.
* Simulations can now specify a `matrix` with possible values for every simulation parameter. The final parameters will be calculated based on the `parameters` used and a cartesian product (i.e., all possible combinations) of each parameter.
* Simple debugging capabilities in `soil.debugging`, with a custom `pdb.Debugger` subclass that exposes commands to list agents and their status and set breakpoints on states (for FSM agents). Try it with `soil --debug <simulation file>`
### Changed
* Configuration schema (`Simulation`) is very simplified. All simulations should be checked
* Model / environment variables are expected (but not enforced) to be a single value. This is done to more closely align with mesa
* `Exporter.iteration_end` now takes two parameters: `env` (same as before) and `params` (specific parameters for this environment). We considered including a `parameters` attribute in the environment, but this would not be compatible with mesa.
* `num_trials` renamed to `iterations`
* General renaming of `trial` to `iteration`, to work better with `mesa`
* `model_parameters` renamed to `parameters` in simulation
* Simulation results for every iteration of a simulation with the same name are stored in a single `sqlite` database
### Removed
* Any `tsih` and `History` integration in the main classes. To record the state of environments/agents, just use a datacollector. In some cases this may be slower or consume more memory than the previous system. However, few cases actually used the full potential of the history, and it came at the cost of unnecessary complexity and worse performance for the majority of cases.
## [0.20.8]
### Changed

73
README.md
View File

@ -1,10 +1,65 @@
# [SOIL](https://github.com/gsi-upm/soil)
Soil is an extensible and user-friendly Agent-based Social Simulator for Social Networks.
Learn how to run your own simulations with our [documentation](http://soilsim.readthedocs.io).
Follow our [tutorial](examples/tutorial/soil_tutorial.ipynb) to develop your own agent models.
> **Warning**
> Soil 1.0 introduced many fundamental changes. Check the [documention on how to update your simulations to work with newer versions](docs/notes_v1.0.rst)
## Features
* Integration with (social) networks (through `networkx`)
* Convenience functions and methods to easily assign agents to your model (and optionally to its network):
* Following a given distribution (e.g., 2 agents of type `Foo`, 10% of the network should be agents of type `Bar`)
* Based on the topology of the network
* **Several types of abstractions for agents**:
* Finite state machine, where methods can be turned into a state
* Network agents, which have convenience methods to access the model's topology
* Generator-based agents, whose state is paused though a `yield` and resumed on the next step
* **Reporting and data collection**:
* Soil models include data collection and record some data by default (# of agents, state of each agent, etc.)
* All data collected are exported by default to a SQLite database and a description file
* Options to export to other formats, such as CSV, or defining your own exporters
* A summary of the data collected is shown in the command line, for easy inspection
* **An event-based scheduler**
* Agents can be explicit about when their next time/step should be, and not all agents run in every step. This avoids unnecessary computation.
* Time intervals between each step are flexible.
* There are primitives to specify when the next execution of an agent should be (or conditions)
* **Actor-inspired** message-passing
* A simulation runner (`soil.Simulation`) that can:
* Run models in parallel
* Save results to different formats
* Simulation configuration files
* A command line interface (`soil`), to quickly run simulations with different parameters
* An integrated debugger (`soil --debug`) with custom functions to print agent states and break at specific states
## Mesa compatibility
SOIL has been redesigned to integrate well with [Mesa](https://github.com/projectmesa/mesa).
For instance, it should be possible to run a `mesa.Model` models using a `soil.Simulation` and the `soil` CLI, or to integrate the `soil.TimedActivation` scheduler on a `mesa.Model`.
Note that some combinations of `mesa` and `soil` components, while technically possible, are much less useful or might yield surprising results.
For instance, you may add any `soil.agent` agent on a regular `mesa.Model` with a vanilla scheduler from `mesa.time`.
But in that case the agents will not get any of the advanced event-based scheduling, and most agent behaviors that depend on that may not work.
## Changes in version 0.3
Version 0.3 came packed with many changes to provide much better integration with MESA.
For a long time, we tried to keep soil backwards-compatible, but it turned out to be a big endeavour and the resulting code was less readable.
This translates to harder maintenance and a worse experience for newcomers.
In the end, we decided to make some breaking changes.
If you have an older Soil simulation, you have two options:
* Update the necessary configuration files and code. You may use the examples in the `examples` folder for reference, as well as the documentation.
* Keep using a previous `soil` version.
## Citation
@ -31,24 +86,6 @@ If you use Soil in your research, don't forget to cite this paper:
```
## Mesa compatibility
Soil is in the process of becoming fully compatible with MESA.
As of this writing,
This is a non-exhaustive list of tasks to achieve compatibility:
* Environments.agents and mesa.Agent.agents are not the same. env is a property, and it only takes into account network and environment agents. Might rename environment_agents to other_agents or sth like that
- [ ] Integrate `soil.Simulation` with mesa's runners:
- [ ] `soil.Simulation` could mimic/become a `mesa.batchrunner`
- [ ] Integrate `soil.Environment` with `mesa.Model`:
- [x] `Soil.Environment` inherits from `mesa.Model`
- [x] `Soil.Environment` includes a Mesa-like Scheduler (see the `soil.time` module.
- [ ] Integrate `soil.Agent` with `mesa.Agent`:
- [x] Rename agent.id to unique_id?
- [x] mesa agents can be used in soil simulations (see `examples/mesa`)
- [ ] Document the new APIs and usage
@Copyright GSI - Universidad Politécnica de Madrid 2017-2021
[![SOIL](logo_gsi.png)](https://www.gsi.upm.es)

241
docs/configuration.rst
View File

@ -1,241 +0,0 @@
Configuring a simulation
------------------------
There are two ways to configure a simulation: programmatically and with a configuration file.
In both cases, the parameters used are the same.
The advantage of a configuration file is that it is a clean declarative description, and it makes it easier to reproduce.
Simulation configuration files can be formatted in ``json`` or ``yaml`` and they define all the parameters of a simulation.
Here's an example (``example.yml``).
.. literalinclude:: example.yml
:language: yaml
This example configuration will run three trials (``num_trials``) of a simulation containing a randomly generated network (``network_params``).
The 100 nodes in the network will be SISaModel agents (``network_agents.agent_type``), which is an agent behavior that is included in Soil.
10% of the agents (``weight=1``) will start in the content state, 10% in the discontent state, and the remaining 80% (``weight=8``) in the neutral state.
All agents will have access to the environment (``environment_params``), which only contains one variable, ``prob_infected``.
The state of the agents will be updated every 2 seconds (``interval``).
Now run the simulation with the command line tool:
.. code:: bash
soil example.yml
Once the simulation finishes, its results will be stored in a folder named ``MyExampleSimulation``.
Three types of objects are saved by default: a pickle of the simulation; a ``YAML`` representation of the simulation (which can be used to re-launch it); and for every trial, a ``sqlite`` file with the content of the state of every network node and the environment parameters at every step of the simulation.
.. code::
soil_output
└── MyExampleSimulation
├── MyExampleSimulation.dumped.yml
├── MyExampleSimulation.simulation.pickle
├── MyExampleSimulation_trial_0.db.sqlite
├── MyExampleSimulation_trial_1.db.sqlite
└── MyExampleSimulation_trial_2.db.sqlite
You may also ask soil to export the states in a ``csv`` file, and the network in gephi format (``gexf``).
Network
=======
The network topology for the simulation can be loaded from an existing network file or generated with one of the random network generation methods from networkx.
Loading a network
#################
To load an existing network, specify its path in the configuration:
.. code:: yaml
---
network_params:
path: /tmp/mynetwork.gexf
Soil will try to guess what networkx method to use to read the file based on its extension.
However, we only test using ``gexf`` files.
For simple networks, you may also include them in the configuration itself using , using the ``topology`` parameter like so:
.. code:: yaml
---
topology:
nodes:
- id: First
- id: Second
links:
- source: First
target: Second
Generating a random network
###########################
To generate a random network using one of networkx's built-in methods, specify the `graph generation algorithm <https://networkx.github.io/documentation/development/reference/generators.html>`_ and other parameters.
For example, the following configuration is equivalent to :code:`nx.complete_graph(n=100)`:
.. code:: yaml
network_params:
generator: complete_graph
n: 100
Environment
============
The environment is the place where the shared state of the simulation is stored.
For instance, the probability of disease outbreak.
The configuration file may specify the initial value of the environment parameters:
.. code:: yaml
environment_params:
daily_probability_of_earthquake: 0.001
number_of_earthquakes: 0
All agents have access to the environment parameters.
In some scenarios, it is useful to have a custom environment, to provide additional methods or to control the way agents update environment state.
For example, if our agents play the lottery, the environment could provide a method to decide whether the agent wins, instead of leaving it to the agent.
Agents
======
Agents are a way of modelling behavior.
Agents can be characterized with two variables: agent type (``agent_type``) and state.
Only one agent is executed at a time (generally, every ``interval`` seconds), and it has access to its state and the environment parameters.
Through the environment, it can access the network topology and the state of other agents.
There are three three types of agents according to how they are added to the simulation: network agents and environment agent.
Network Agents
##############
Network agents are attached to a node in the topology.
The configuration file allows you to specify how agents will be mapped to topology nodes.
The simplest way is to specify a single type of agent.
Hence, every node in the network will be associated to an agent of that type.
.. code:: yaml
agent_type: SISaModel
It is also possible to add more than one type of agent to the simulation, and to control the ratio of each type (using the ``weight`` property).
For instance, with following configuration, it is five times more likely for a node to be assigned a CounterModel type than a SISaModel type.
.. code:: yaml
network_agents:
- agent_type: SISaModel
weight: 1
- agent_type: CounterModel
weight: 5
The third option is to specify the type of agent on the node itself, e.g.:
.. code:: yaml
topology:
nodes:
- id: first
agent_type: BaseAgent
states:
first:
agent_type: SISaModel
This would also work with a randomly generated network:
.. code:: yaml
network:
generator: complete
n: 5
agent_type: BaseAgent
states:
- agent_type: SISaModel
In addition to agent type, you may add a custom initial state to the distribution.
This is very useful to add the same agent type with different states.
e.g., to populate the network with SISaModel, roughly 10% of them with a discontent state:
.. code:: yaml
network_agents:
- agent_type: SISaModel
weight: 9
state:
id: neutral
- agent_type: SISaModel
weight: 1
state:
id: discontent
Lastly, the configuration may include initial state for one or more nodes.
For instance, to add a state for the two nodes in this configuration:
.. code:: yaml
agent_type: SISaModel
network:
generator: complete_graph
n: 2
states:
- id: content
- id: discontent
Or to add state only to specific nodes (by ``id``).
For example, to apply special skills to Linux Torvalds in a simulation:
.. literalinclude:: ../examples/torvalds.yml
:language: yaml
Environment Agents
##################
In addition to network agents, more agents can be added to the simulation.
These agents are programmed in much the same way as network agents, the only difference is that they will not be assigned to network nodes.
.. code::
environment_agents:
- agent_type: MyAgent
state:
mood: happy
- agent_type: DummyAgent
You may use environment agents to model events that a normal agent cannot control, such as natural disasters or chance.
They are also useful to add behavior that has little to do with the network and the interactions within that network.
Templating
==========
Sometimes, it is useful to parameterize a simulation and run it over a range of values in order to compare each run and measure the effect of those parameters in the simulation.
For instance, you may want to run a simulation with different agent distributions.
This can be done in Soil using **templates**.
A template is a configuration where some of the values are specified with a variable.
e.g., ``weight: "{{ var1 }}"`` instead of ``weight: 1``.
There are two types of variables, depending on how their values are decided:
* Fixed. A list of values is provided, and a new simulation is run for each possible value. If more than a variable is given, a new simulation will be run per combination of values.
* Bounded/Sampled. The bounds of the variable are provided, along with a sampler method, which will be used to compute all the configuration combinations.
When fixed and bounded variables are mixed, Soil generates a new configuration per combination of fixed values and bounded values.
Here is an example with a single fixed variable and two bounded variable:
.. literalinclude:: ../examples/template.yml
:language: yaml

29
docs/example.yml
View File

@ -3,24 +3,29 @@ name: MyExampleSimulation
max_time: 50
num_trials: 3
interval: 2
network_params:
model_params:
topology:
params:
generator: barabasi_albert_graph
n: 100
m: 2
network_agents:
- agent_type: SISaModel
weight: 1
agents:
distribution:
- agent_class: SISaModel
topology: True
ratio: 0.1
state:
id: content
- agent_type: SISaModel
weight: 1
state_id: content
- agent_class: SISaModel
topology: True
ratio: .1
state:
id: discontent
- agent_type: SISaModel
weight: 8
state_id: discontent
- agent_class: SISaModel
topology: True
ratio: 0.8
state:
id: neutral
environment_params:
state_id: neutral
prob_infect: 0.075
neutral_discontent_spon_prob: 0.1
neutral_discontent_infected_prob: 0.3

22
docs/index.rst
View File

@ -1,12 +1,20 @@
.. Soil documentation master file, created by
sphinx-quickstart on Tue Apr 25 12:48:56 2017.
You can adapt this file completely to your liking, but it should at least
contain the root `toctree` directive.
Welcome to Soil's documentation!
================================
Soil is an Agent-based Social Simulator in Python focused on Social Networks.
Soil is an opinionated Agent-based Social Simulator in Python focused on Social Networks.
.. image:: soil.png
:width: 80%
:align: center
Soil can be installed through pip (see more details in the :doc:`installation` page):
.. code:: bash
pip install soil
To get started developing your own simulations and agent behaviors, check out our :doc:`Tutorial <soil_tutorial>` and the `examples on GitHub <https://github.com/gsi-upm/soil/tree/master/examples>.
If you use Soil in your research, do not forget to cite this paper:
@ -38,8 +46,6 @@ If you use Soil in your research, do not forget to cite this paper:
:caption: Learn more about soil:
installation
quickstart
configuration
Tutorial <soil_tutorial>
..

45
docs/installation.rst
View File

@ -1,7 +1,10 @@
Installation
------------
The easiest way to install Soil is through pip, with Python >= 3.4:
Through pip
===========
The easiest way to install Soil is through pip, with Python >= 3.8:
.. code:: bash
@ -14,6 +17,10 @@ Now test that it worked by running the command line tool
soil --help
#or
python -m soil --help
Or, if you're using using soil programmatically:
.. code:: python
@ -21,4 +28,38 @@ Or, if you're using using soil programmatically:
import soil
print(soil.__version__)
The latest version can be installed through `GitLab <https://lab.gsi.upm.es/soil/soil.git>`_ or `GitHub <https://github.com/gsi-upm/soil>`_.
Web UI
======
Soil also includes a web server that allows you to upload your simulations, change parameters, and visualize the results, including a timeline of the network.
To make it work, you have to install soil like this:
.. code::
pip install soil[web]
Once installed, the soil web UI can be run in two ways:
.. code::
soil-web
# OR
python -m soil.web
Development
===========
The latest version can be downloaded from `GitHub <https://github.com/gsi-upm/soil>`_ and installed manually:
.. code:: bash
git clone https://github.com/gsi-upm/soil
cd soil
python -m venv .venv
source .venv/bin/activate
pip install --editable .

2
docs/make.bat
View File

@ -12,7 +12,7 @@ set BUILDDIR=_build
set SPHINXPROJ=Soil
if "%1" == "" goto help
eE
%SPHINXBUILD% >NUL 2>NUL
if errorlevel 9009 (
echo.

22
docs/mesa.rst Normal file
View File

@ -0,0 +1,22 @@
Mesa compatibility
------------------
Soil is in the process of becoming fully compatible with MESA.
The idea is to provide a set of modular classes and functions that extend the functionality of mesa, whilst staying compatible.
In the end, it should be possible to add regular mesa agents to a soil simulation, or use a soil agent within a mesa simulation/model.
This is a non-exhaustive list of tasks to achieve compatibility:
- [ ] Integrate `soil.Simulation` with mesa's runners:
- [ ] `soil.Simulation` could mimic/become a `mesa.batchrunner`
- [ ] Integrate `soil.Environment` with `mesa.Model`:
- [x] `Soil.Environment` inherits from `mesa.Model`
- [x] `Soil.Environment` includes a Mesa-like Scheduler (see the `soil.time` module.
- [ ] Allow for `mesa.Model` to be used in a simulation.
- [ ] Integrate `soil.Agent` with `mesa.Agent`:
- [x] Rename agent.id to unique_id?
- [x] mesa agents can be used in soil simulations (see `examples/mesa`)
- [ ] Provide examples
- [ ] Using mesa modules in a soil simulation
- [ ] Using soil modules in a mesa simulation
- [ ] Document the new APIs and usage

35
docs/notes_v1.0.rst Normal file
View File

@ -0,0 +1,35 @@
What are the main changes in version 1.0?
#########################################
Version 1.0 is a major rewrite of the Soil system, focused on simplifying the API, aligning it with Mesa, and making it easier to use.
Unfortunately, this comes at the cost of backwards compatibility.
We drew several lessons from the previous version of Soil, and tried to address them in this version.
Mainly:
- The split between simulation configuration and simulation code was overly complicated for most use cases. As a result, most users ended up reusing configuration.
- Storing **all** the simulation data in a database is costly and unnecessary for most use cases. For most use cases, only a handful of variables need to be stored. This fits nicely with Mesa's data collection system.
- The API was too complex, and it was difficult to understand how to use it.
- Most parts of the API were not aligned with Mesa, which made it difficult to use Mesa's features or to integrate Soil modules with Mesa code, especially for newcomers.
- Many parts of the API were tightly coupled, which made it difficult to find bugs, test the system and add new features.
The 0.30 rewrite should provide a middle ground between Soil's opinionated approach and Mesa's flexibility.
The new Soil is less configuration-centric.
It aims to provide more modular and convenient functions, most of which can be used in vanilla Mesa.
How are agents assigned to nodes in the network
###############################################
The constructor of the `NetworkAgent` class has two arguments: `node_id` and `topology`.
If `topology` is not provided, it will default to `self.model.topology`.
This assignment might err if the model does not have a `topology` attribute, but most Soil environments derive from `NetworkEnvironment`, so they include a topology by default.
If `node_id` is not provided, a random node will be selected from the topology, until a node with no agent is found.
Then, the `node_id` of that node is assigned to the agent.
If no node with no agent is found, a new node is automatically added to the topology.
Can Soil environments include more than one network / topology?
###############################################################
Yes, but each network has to be included manually.
Somewhere between 0.20 and 0.30 we included the ability to include multiple networks, but it was deemed too complex and was removed.

BIN
docs/output_21_0.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 8.3 KiB

BIN
docs/output_30_0.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 4.7 KiB

BIN
docs/output_34_0.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 11 KiB

BIN
docs/output_49_0.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 26 KiB

BIN
docs/output_50_0.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 31 KiB

BIN
docs/output_55_0.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 29 KiB

BIN
docs/output_56_0.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 33 KiB

BIN
docs/output_72_0.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 23 KiB

93
docs/quickstart.rst
View File

@ -1,93 +0,0 @@
Quickstart
----------
This section shows how to run your first simulation with Soil.
For installation instructions, see :doc:`installation`.
There are mainly two parts in a simulation: agent classes and simulation configuration.
An agent class defines how the agent will behave throughout the simulation.
The configuration includes things such as number of agents to use and their type, network topology to use, etc.
.. image:: soil.png
:width: 80%
:align: center
Soil includes several agent classes in the ``soil.agents`` module, and we will use them in this quickstart.
If you are interested in developing your own agents classes, see :doc:`soil_tutorial`.
Configuration
=============
To get you started, we will use this configuration (:download:`download the file <quickstart.yml>` directly):
.. literalinclude:: quickstart.yml
:language: yaml
The agent type used, SISa, is a very simple model.
It only has three states (neutral, content and discontent),
Its parameters are the probabilities to change from one state to another, either spontaneously or because of contagion from neighboring agents.
Running the simulation
======================
To see the simulation in action, simply point soil to the configuration, and tell it to store the graph and the history of agent states and environment parameters at every point.
.. code::
❯ soil --graph --csv quickstart.yml [13:35:29]
INFO:soil:Using config(s): quickstart
INFO:soil:Dumping results to soil_output/quickstart : ['csv', 'gexf']
INFO:soil:Starting simulation quickstart at 13:35:30.
INFO:soil:Starting Simulation quickstart trial 0 at 13:35:30.
INFO:soil:Finished Simulation quickstart trial 0 at 13:35:49 in 19.43677067756653 seconds
INFO:soil:Starting Dumping simulation quickstart trial 0 at 13:35:49.
INFO:soil:Finished Dumping simulation quickstart trial 0 at 13:35:51 in 1.7733407020568848 seconds
INFO:soil:Dumping results to soil_output/quickstart
INFO:soil:Finished simulation quickstart at 13:35:51 in 21.29862952232361 seconds
The ``CSV`` file should look like this:
.. code::
agent_id,t_step,key,value
env,0,neutral_discontent_spon_prob,0.05
env,0,neutral_discontent_infected_prob,0.1
env,0,neutral_content_spon_prob,0.2
env,0,neutral_content_infected_prob,0.4
env,0,discontent_neutral,0.2
env,0,discontent_content,0.05
env,0,content_discontent,0.05
env,0,variance_d_c,0.05
env,0,variance_c_d,0.1
Results and visualization
=========================
The environment variables are marked as ``agent_id`` env.
Th exported values are only stored when they change.
To find out how to get every key and value at every point in the simulation, check out the :doc:`soil_tutorial`.
The dynamic graph is exported as a .gexf file which could be visualized with
`Gephi <https://gephi.org/users/download/>`__.
Now it is your turn to experiment with the simulation.
Change some of the parameters, such as the number of agents, the probability of becoming content, or the type of network, and see how the results change.
Soil also includes a web server that allows you to upload your simulations, change parameters, and visualize the results, including a timeline of the network.
To make it work, you have to install soil like this:
.. code::
pip install soil[web]
Once installed, the soil web UI can be run in two ways:
.. code::
soil-web
# OR
python -m soil.web

30
docs/quickstart.yml
View File

@ -1,30 +0,0 @@
---
name: quickstart
num_trials: 1
max_time: 1000
network_agents:
- agent_type: SISaModel
state:
id: neutral
weight: 1
- agent_type: SISaModel
state:
id: content
weight: 2
network_params:
n: 100
k: 5
p: 0.2
generator: newman_watts_strogatz_graph
environment_params:
neutral_discontent_spon_prob: 0.05
neutral_discontent_infected_prob: 0.1
neutral_content_spon_prob: 0.2
neutral_content_infected_prob: 0.4
discontent_neutral: 0.2
discontent_content: 0.05
content_discontent: 0.05
variance_d_c: 0.05
variance_c_d: 0.1
content_neutral: 0.1
standard_variance: 0.1

2
docs/requirements.txt
View File

@ -1 +1 @@
ipython==7.31.1
ipython>=7.31.1

12
docs/soil-vs.rst Normal file
View File

@ -0,0 +1,12 @@
### MESA
Starting with version 0.3, Soil has been redesigned to complement Mesa, while remaining compatible with it.
That means that every component in Soil (i.e., Models, Environments, etc.) can be mixed with existing mesa components.
In fact, there are examples that show how that integration may be used, in the `examples/mesa` folder in the repository.
Here are some reasons to use Soil instead of plain mesa:
- Less boilerplate for common scenarios (by some definitions of common)
- Functions to automatically populate a topology with an agent distribution (i.e., different ratios of agent class and state)
- The `soil.Simulation` class allows you to run multiple instances of the same experiment (i.e., multiple trials with the same parameters but a different randomness seed)
- Reporting functions that aggregate multiple

4548
docs/soil_tutorial.rst
View File

File diff suppressed because it is too large Load Diff

27
examples/complete.yml
View File

@ -1,27 +0,0 @@
---
name: simple
group: tests
dir_path: "/tmp/"
num_trials: 3
max_time: 100
interval: 1
seed: "CompleteSeed!"
network_params:
generator: complete_graph
n: 10
network_agents:
- agent_type: CounterModel
weight: 1
state:
state_id: 0
- agent_type: AggregatedCounter
weight: 0.2
environment_agents: []
environment_class: Environment
environment_params:
am_i_complete: true
default_state:
incidents: 0
states:
- name: 'The first node'
- name: 'The second node'

16
examples/custom_generator/custom_generator.yml
View File

@ -1,16 +0,0 @@
---
name: custom-generator
description: Using a custom generator for the network
num_trials: 3
max_time: 100
interval: 1
network_params:
generator: mymodule.mygenerator
# These are custom parameters
n: 10
n_edges: 5
network_agents:
- agent_type: CounterModel
weight: 1
state:
state_id: 0

39
examples/custom_generator/generator_sim.py Normal file
View File

@ -0,0 +1,39 @@
from networkx import Graph
import random
import networkx as nx
from soil import Simulation, Environment, CounterModel, parameters
def mygenerator(n=5, n_edges=5):
"""
Just a simple generator that creates a network with n nodes and
n_edges edges. Edges are assigned randomly, only avoiding self loops.
"""
G = nx.Graph()
for i in range(n):
G.add_node(i)
for i in range(n_edges):
nodes = list(G.nodes)
n_in = random.choice(nodes)
nodes.remove(n_in) # Avoid loops
n_out = random.choice(nodes)
G.add_edge(n_in, n_out)
return G
class GeneratorEnv(Environment):
"""Using a custom generator for the network"""
generator: parameters.function = staticmethod(mygenerator)
def init(self):
self.create_network(generator=self.generator, n=10, n_edges=5)
self.add_agents(CounterModel)
sim = Simulation(model=GeneratorEnv, max_steps=10, interval=1)
if __name__ == '__main__':
sim.run(dump=False)

27
examples/custom_generator/mymodule.py
View File

@ -1,27 +0,0 @@
from networkx import Graph
import networkx as nx
from random import choice
def mygenerator(n=5, n_edges=5):
'''
Just a simple generator that creates a network with n nodes and
n_edges edges. Edges are assigned randomly, only avoiding self loops.
'''
G = nx.Graph()
for i in range(n):
G.add_node(i)
for i in range(n_edges):
nodes = list(G.nodes)
n_in = choice(nodes)
nodes.remove(n_in) # Avoid loops
n_out = choice(nodes)
G.add_edge(n_in, n_out)
return G

35
examples/custom_timeouts/custom_timeouts.py
View File

@ -1,35 +0,0 @@
from soil.agents import FSM, state, default_state
class Fibonacci(FSM):
'''Agent that only executes in t_steps that are Fibonacci numbers'''
defaults = {
'prev': 1
}
@default_state
@state
def counting(self):
self.log('Stopping at {}'.format(self.now))
prev, self['prev'] = self['prev'], max([self.now, self['prev']])
return None, self.env.timeout(prev)
class Odds(FSM):
'''Agent that only executes in odd t_steps'''
@default_state
@state
def odds(self):
self.log('Stopping at {}'.format(self.now))
return None, self.env.timeout(1+self.now%2)
if __name__ == '__main__':
import logging
logging.basicConfig(level=logging.INFO)
from soil import Simulation
s = Simulation(network_agents=[{'ids': [0], 'agent_type': Fibonacci},
{'ids': [1], 'agent_type': Odds}],
network_params={"generator": "complete_graph", "n": 2},
max_time=100,
)
s.run(dry_run=True)

41
examples/custom_timeouts/custom_timeouts_sim.py Normal file
View File

@ -0,0 +1,41 @@
from soil.agents import FSM, state, default_state
from soil.time import Delta
class Fibonacci(FSM):
"""Agent that only executes in t_steps that are Fibonacci numbers"""
prev = 1
@default_state
@state
def counting(self):
self.log("Stopping at {}".format(self.now))
prev, self["prev"] = self["prev"], max([self.now, self["prev"]])
return None, Delta(prev)
class Odds(FSM):
"""Agent that only executes in odd t_steps"""
@default_state
@state
def odds(self):
self.log("Stopping at {}".format(self.now))
return None, Delta(1 + self.now % 2)
from soil import Environment, Simulation
from networkx import complete_graph
class TimeoutsEnv(Environment):
def init(self):
self.create_network(generator=complete_graph, n=2)
self.add_agent(agent_class=Fibonacci, node_id=0)
self.add_agent(agent_class=Odds, node_id=1)
sim = Simulation(model=TimeoutsEnv, max_steps=10, interval=1)
if __name__ == "__main__":
sim.run(dump=False)

9
examples/events_and_messages/README.md Normal file
View File

@ -0,0 +1,9 @@
This example can be run like with command-line options, like this:
```bash
python cars.py --level DEBUG -e summary --csv
#or
soil cars.py -e summary
```
This will set the `CSV` (save the agent and model data to a CSV) and `summary` (print the a summary of the data to stdout) exporters, and set the log level to DEBUG.

231
examples/events_and_messages/cars_sim.py Normal file
View File

@ -0,0 +1,231 @@
"""
This is an example of a simplified city, where there are Passengers and Drivers that can take those passengers
from their location to their desired location.
An example scenario could play like the following:
- Drivers start in the `wandering` state, where they wander around the city until they have been assigned a journey
- Passenger(1) tells every driver that it wants to request a Journey.
- Each driver receives the request.
If Driver(2) is interested in providing the Journey, it asks Passenger(1) to confirm that it accepts Driver(2)'s request
- When Passenger(1) accepts the request, two things happen:
- Passenger(1) changes its state to `driving_home`
- Driver(2) starts moving towards the origin of the Journey
- Once Driver(2) reaches the origin, it starts moving itself and Passenger(1) to the destination of the Journey
- When Driver(2) reaches the destination (carrying Passenger(1) along):
- Driver(2) starts wondering again
- Passenger(1) dies, and is removed from the simulation
- If there are no more passengers available in the simulation, Drivers die
"""
from __future__ import annotations
from typing import Optional
from soil import *
from soil import events
from mesa.space import MultiGrid
# More complex scenarios may use more than one type of message between objects.
# A common pattern is to use `enum.Enum` to represent state changes in a request.
@dataclass
class Journey:
"""
This represents a request for a journey. Passengers and drivers exchange this object.
A journey may have a driver assigned or not. If the driver has not been assigned, this
object is considered a "request for a journey".
"""
origin: (int, int)
destination: (int, int)
tip: float
passenger: Passenger
driver: Optional[Driver] = None
class City(EventedEnvironment):
"""
An environment with a grid where drivers and passengers will be placed.
The number of drivers and riders is configurable through its parameters:
:param str n_cars: The total number of drivers to add
:param str n_passengers: The number of passengers in the simulation
:param list agents: Specific agents to use in the simulation. It overrides the `n_passengers`
and `n_cars` params.
:param int height: Height of the internal grid
:param int width: Width of the internal grid
"""
n_cars = 1
n_passengers = 10
height = 100
width = 100
def init(self):
self.grid = MultiGrid(width=self.width, height=self.height, torus=False)
if not self.agents:
self.add_agents(Driver, k=self.n_cars)
self.add_agents(Passenger, k=self.n_passengers)
for agent in self.agents:
self.grid.place_agent(agent, (0, 0))
self.grid.move_to_empty(agent)
self.total_earnings = 0
self.add_model_reporter("total_earnings")
@report
@property
def number_passengers(self):
return self.count_agents(agent_class=Passenger)
class Driver(Evented, FSM):
pos = None
journey = None
earnings = 0
def on_receive(self, msg, sender):
"""This is not a state. It will run (and block) every time check_messages is invoked"""
if self.journey is None and isinstance(msg, Journey) and msg.driver is None:
msg.driver = self
self.journey = msg
def check_passengers(self):
"""If there are no more passengers, stop forever"""
c = self.count_agents(agent_class=Passenger)
self.debug(f"Passengers left {c}")
if not c:
self.die("No more passengers")
@default_state
@state
def wandering(self):
"""Move around the city until a journey is accepted"""
target = None
self.check_passengers()
self.journey = None
while self.journey is None: # No potential journeys detected (see on_receive)
if target is None or not self.move_towards(target):
target = self.random.choice(
self.model.grid.get_neighborhood(self.pos, moore=False)
)
self.check_passengers()
# This will call on_receive behind the scenes, and the agent's status will be updated
self.check_messages()
yield Delta(30) # Wait at least 30 seconds before checking again
try:
# Re-send the journey to the passenger, to confirm that we have been selected
self.journey = yield self.journey.passenger.ask(self.journey, timeout=60)
except events.TimedOut:
# No journey has been accepted. Try again
self.journey = None
return
return self.driving
@state
def driving(self):
"""The journey has been accepted. Pick them up and take them to their destination"""
self.info(f"Driving towards Passenger {self.journey.passenger.unique_id}")
while self.move_towards(self.journey.origin):
yield
self.info(f"Driving {self.journey.passenger.unique_id} from {self.journey.origin} to {self.journey.destination}")
while self.move_towards(self.journey.destination, with_passenger=True):
yield
self.info("Arrived at destination")
self.earnings += self.journey.tip
self.model.total_earnings += self.journey.tip
self.check_passengers()
return self.wandering
def move_towards(self, target, with_passenger=False):
"""Move one cell at a time towards a target"""
self.debug(f"Moving { self.pos } -> { target }")
if target[0] == self.pos[0] and target[1] == self.pos[1]:
return False
next_pos = [self.pos[0], self.pos[1]]
for idx in [0, 1]:
if self.pos[idx] < target[idx]:
next_pos[idx] += 1
break
if self.pos[idx] > target[idx]:
next_pos[idx] -= 1
break
self.model.grid.move_agent(self, tuple(next_pos))
if with_passenger:
self.journey.passenger.pos = (
self.pos
) # This could be communicated through messages
return True
class Passenger(Evented, FSM):
pos = None
def on_receive(self, msg, sender):
"""This is not a state. It will be run synchronously every time `check_messages` is run"""
if isinstance(msg, Journey):
self.journey = msg
return msg
@default_state
@state
def asking(self):
destination = (
self.random.randint(0, self.model.grid.height-1),
self.random.randint(0, self.model.grid.width-1),
)
self.journey = None
journey = Journey(
origin=self.pos,
destination=destination,
tip=self.random.randint(10, 100),
passenger=self,
)
timeout = 60
expiration = self.now + timeout
self.info(f"Asking for journey at: { self.pos }")
self.model.broadcast(journey, ttl=timeout, sender=self, agent_class=Driver)
while not self.journey:
self.debug(f"Waiting for responses at: { self.pos }")
try:
# This will call check_messages behind the scenes, and the agent's status will be updated
# If you want to avoid that, you can call it with: check=False
yield self.received(expiration=expiration)
except events.TimedOut:
self.info(f"Still no response. Waiting at: { self.pos }")
self.model.broadcast(
journey, ttl=timeout, sender=self, agent_class=Driver
)
expiration = self.now + timeout
self.info(f"Got a response! Waiting for driver")
return self.driving_home
@state
def driving_home(self):
while (
self.pos[0] != self.journey.destination[0]
or self.pos[1] != self.journey.destination[1]
):
try:
yield self.received(timeout=60)
except events.TimedOut:
pass
self.die("Got home safe!")
simulation = Simulation(name="RideHailing",
model=City,
seed="carsSeed",
max_time=1000,
parameters=dict(n_passengers=2))
if __name__ == "__main__":
easy(simulation)

21
examples/mesa/mesa.yml
View File

@ -1,21 +0,0 @@
---
name: mesa_sim
group: tests
dir_path: "/tmp"
num_trials: 3
max_time: 100
interval: 1
seed: '1'
network_params:
generator: social_wealth.graph_generator
n: 5
network_agents:
- agent_type: social_wealth.SocialMoneyAgent
weight: 1
environment_class: social_wealth.MoneyEnv
environment_params:
num_mesa_agents: 5
mesa_agent_type: social_wealth.MoneyAgent
N: 10
width: 50
height: 50

7
examples/mesa/mesa_sim.py Normal file
View File

@ -0,0 +1,7 @@
from soil import Simulation
from social_wealth import MoneyEnv, graph_generator
sim = Simulation(name="mesa_sim", dump=False, max_steps=10, interval=2, model=MoneyEnv, parameters=dict(generator=graph_generator, N=10, width=50, height=50))
if __name__ == "__main__":
sim.run()

56
examples/mesa/server.py
View File

@ -1,7 +1,8 @@
from mesa.visualization.ModularVisualization import ModularServer
from soil.visualization import UserSettableParameter
from mesa.visualization.UserParam import Slider, Choice
from mesa.visualization.modules import ChartModule, NetworkModule, CanvasGrid
from social_wealth import MoneyEnv, graph_generator, SocialMoneyAgent
import networkx as nx
class MyNetwork(NetworkModule):
@ -13,15 +14,18 @@ def network_portrayal(env):
# The model ensures there is 0 or 1 agent per node
portrayal = dict()
wealths = {
node_id: data["agent"].wealth for (node_id, data) in env.G.nodes(data=True)
}
portrayal["nodes"] = [
{
"id": agent_id,
"size": env.get_agent(agent_id).wealth,
# "color": "#CC0000" if not agents or agents[0].wealth == 0 else "#007959",
"color": "#CC0000",
"label": f"{agent_id}: {env.get_agent(agent_id).wealth}",
"id": node_id,
"size": 2 * (wealth + 1),
"color": "#CC0000" if wealth == 0 else "#007959",
# "color": "#CC0000",
"label": f"{node_id}: {wealth}",
}
for (agent_id) in env.G.nodes
for (node_id, wealth) in wealths.items()
]
portrayal["edges"] = [
@ -29,7 +33,6 @@ def network_portrayal(env):
for edge_id, (source, target) in enumerate(env.G.edges)
]
return portrayal
@ -40,7 +43,7 @@ def gridPortrayal(agent):
:param agent: the agent in the simulation
:return: the portrayal dictionary
"""
color = max(10, min(agent.wealth*10, 100))
color = max(10, min(agent.wealth * 10, 100))
return {
"Shape": "rect",
"w": 1,
@ -51,18 +54,17 @@ def gridPortrayal(agent):
"Text": agent.unique_id,
"x": agent.pos[0],
"y": agent.pos[1],
"Color": f"rgba(31, 10, 255, 0.{color})"
"Color": f"rgba(31, 10, 255, 0.{color})",
}
grid = MyNetwork(network_portrayal, 500, 500, library="sigma")
grid = MyNetwork(network_portrayal, 500, 500)
chart = ChartModule(
[{"Label": "Gini", "Color": "Black"}], data_collector_name="datacollector"
)
model_params = {
"N": UserSettableParameter(
"slider",
parameters = {
"N": Slider(
"N",
5,
1,
@ -70,9 +72,7 @@ model_params = {
1,
description="Choose how many agents to include in the model",
),
"network_agents": [{"agent_type": SocialMoneyAgent}],
"height": UserSettableParameter(
"slider",
"height": Slider(
"height",
5,
5,
@ -80,8 +80,7 @@ model_params = {
1,
description="Grid height",
),
"width": UserSettableParameter(
"slider",
"width": Slider(
"width",
5,
5,
@ -89,17 +88,24 @@ model_params = {
1,
description="Grid width",
),
"network_params": {
'generator': graph_generator
},
"agent_class": Choice(
"Agent class",
value="MoneyAgent",
choices=["MoneyAgent", "SocialMoneyAgent"],
),
"generator": graph_generator,
}
canvas_element = CanvasGrid(gridPortrayal, model_params["width"].value, model_params["height"].value, 500, 500)
canvas_element = CanvasGrid(
gridPortrayal, parameters["width"].value, parameters["height"].value, 500, 500
)
server = ModularServer(
MoneyEnv, [grid, chart, canvas_element], "Money Model", model_params
MoneyEnv, [grid, chart, canvas_element], "Money Model", parameters
)
server.port = 8521
server.launch(open_browser=False)
if __name__ == '__main__':
server.launch(open_browser=False)

89
examples/mesa/social_wealth.py
View File

@ -1,23 +1,26 @@
'''
"""
This is an example that adds soil agents and environment in a normal
mesa workflow.
'''
"""
from mesa import Agent as MesaAgent
from mesa.space import MultiGrid
# from mesa.time import RandomActivation
from mesa.datacollection import DataCollector
from mesa.batchrunner import BatchRunner
import networkx as nx
from soil import NetworkAgent, Environment
from soil import NetworkAgent, Environment, serialization
def compute_gini(model):
agent_wealths = [agent.wealth for agent in model.agents]
x = sorted(agent_wealths)
N = len(list(model.agents))
B = sum( xi * (N-i) for i,xi in enumerate(x) ) / (N*sum(x))
return (1 + (1/N) - 2*B)
B = sum(xi * (N - i) for i, xi in enumerate(x)) / (N * sum(x))
return 1 + (1 / N) - 2 * B
class MoneyAgent(MesaAgent):
"""
@ -25,15 +28,14 @@ class MoneyAgent(MesaAgent):
It will only share wealth with neighbors based on grid proximity
"""
def __init__(self, unique_id, model):
def __init__(self, unique_id, model, wealth=1, **kwargs):
super().__init__(unique_id=unique_id, model=model)
self.wealth = 1
self.wealth = wealth
def move(self):
possible_steps = self.model.grid.get_neighborhood(
self.pos,
moore=True,
include_center=False)
self.pos, moore=True, include_center=False
)
new_position = self.random.choice(possible_steps)
self.model.grid.move_agent(self, new_position)
@ -45,21 +47,21 @@ class MoneyAgent(MesaAgent):
self.wealth -= 1
def step(self):
self.info("Crying wolf", self.pos)
print("Crying wolf", self.pos)
self.move()
if self.wealth > 0:
self.give_money()
class SocialMoneyAgent(NetworkAgent, MoneyAgent):
class SocialMoneyAgent(MoneyAgent, NetworkAgent):
wealth = 1
def give_money(self):
cellmates = set(self.model.grid.get_cell_list_contents([self.pos]))
friends = set(self.get_neighboring_agents())
friends = set(self.get_neighbors())
self.info("Trying to give money")
self.debug("Cellmates: ", cellmates)
self.debug("Friends: ", friends)
self.info("Cellmates: ", cellmates)
self.info("Friends: ", friends)
nearby_friends = list(cellmates & friends)
@ -69,14 +71,35 @@ class SocialMoneyAgent(NetworkAgent, MoneyAgent):
self.wealth -= 1
def graph_generator(n=5):
G = nx.Graph()
for ix in range(n):
G.add_edge(0, ix)
return G
class MoneyEnv(Environment):
"""A model with some number of agents."""
def __init__(self, N, width, height, *args, network_params, **kwargs):
network_params['n'] = N
super().__init__(*args, network_params=network_params, **kwargs)
def __init__(
self,
width,
height,
N,
generator=graph_generator,
agent_class=SocialMoneyAgent,
topology=None,
**kwargs
):
generator = serialization.deserialize(generator)
agent_class = serialization.deserialize(agent_class, globs=globals())
topology = generator(n=N)
super().__init__(topology=topology, N=N, **kwargs)
self.grid = MultiGrid(width, height, False)
self.populate_network(agent_class=agent_class)
# Create agents
for agent in self.agents:
x = self.random.randrange(self.grid.width)
@ -84,37 +107,31 @@ class MoneyEnv(Environment):
self.grid.place_agent(agent, (x, y))
self.datacollector = DataCollector(
model_reporters={"Gini": compute_gini},
agent_reporters={"Wealth": "wealth"})
model_reporters={"Gini": compute_gini}, agent_reporters={"Wealth": "wealth"}
)
def graph_generator(n=5):
G = nx.Graph()
for ix in range(n):
G.add_edge(0, ix)
return G
if __name__ == "__main__":
if __name__ == '__main__':
G = graph_generator()
fixed_params = {"topology": G,
fixed_params = {
"generator": nx.complete_graph,
"width": 10,
"network_agents": [{"agent_type": SocialMoneyAgent,
'weight': 1}],
"height": 10}
"network_agents": [{"agent_class": SocialMoneyAgent, "weight": 1}],
"height": 10,
}
variable_params = {"N": range(10, 100, 10)}
batch_run = BatchRunner(MoneyEnv,
batch_run = BatchRunner(
MoneyEnv,
variable_parameters=variable_params,
fixed_parameters=fixed_params,
iterations=5,
max_steps=100,
model_reporters={"Gini": compute_gini})
model_reporters={"Gini": compute_gini},
)
batch_run.run_all()
run_data = batch_run.get_model_vars_dataframe()
run_data.head()
print(run_data.Gini)

32
examples/mesa/wealth.py
View File

@ -4,24 +4,26 @@ from mesa.time import RandomActivation
from mesa.datacollection import DataCollector
from mesa.batchrunner import BatchRunner
def compute_gini(model):
agent_wealths = [agent.wealth for agent in model.schedule.agents]
x = sorted(agent_wealths)
N = model.num_agents
B = sum( xi * (N-i) for i,xi in enumerate(x) ) / (N*sum(x))
return (1 + (1/N) - 2*B)
B = sum(xi * (N - i) for i, xi in enumerate(x)) / (N * sum(x))
return 1 + (1 / N) - 2 * B
class MoneyAgent(Agent):
""" An agent with fixed initial wealth."""
"""An agent with fixed initial wealth."""
def __init__(self, unique_id, model):
super().__init__(unique_id, model)
self.wealth = 1
def move(self):
possible_steps = self.model.grid.get_neighborhood(
self.pos,
moore=True,
include_center=False)
self.pos, moore=True, include_center=False
)
new_position = self.random.choice(possible_steps)
self.model.grid.move_agent(self, new_position)
@ -37,8 +39,10 @@ class MoneyAgent(Agent):
if self.wealth > 0:
self.give_money()
class MoneyModel(Model):
"""A model with some number of agents."""
def __init__(self, N, width, height):
self.num_agents = N
self.grid = MultiGrid(width, height, True)
@ -55,29 +59,29 @@ class MoneyModel(Model):
self.grid.place_agent(a, (x, y))
self.datacollector = DataCollector(
model_reporters={"Gini": compute_gini},
agent_reporters={"Wealth": "wealth"})
model_reporters={"Gini": compute_gini}, agent_reporters={"Wealth": "wealth"}
)
def step(self):
self.datacollector.collect(self)
self.schedule.step()
if __name__ == '__main__':
if __name__ == "__main__":
fixed_params = {"width": 10,
"height": 10}
fixed_params = {"width": 10, "height": 10}
variable_params = {"N": range(10, 500, 10)}
batch_run = BatchRunner(MoneyModel,
batch_run = BatchRunner(
MoneyModel,
variable_params,
fixed_params,
iterations=5,
max_steps=100,
model_reporters={"Gini": compute_gini})
model_reporters={"Gini": compute_gini},
)
batch_run.run_all()
run_data = batch_run.get_model_vars_dataframe()
run_data.head()
print(run_data.Gini)

24
examples/newsspread/NewsSpread.ipynb
View File

@ -80,11 +80,11 @@
"max_time: 300\r\n",
"name: Sim_all_dumb\r\n",
"network_agents:\r\n",
"- agent_type: DumbViewer\r\n",
"- agent_class: DumbViewer\r\n",
" state:\r\n",
" has_tv: false\r\n",
" weight: 1\r\n",
"- agent_type: DumbViewer\r\n",
"- agent_class: DumbViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" weight: 1\r\n",
@ -104,19 +104,19 @@
"max_time: 300\r\n",
"name: Sim_half_herd\r\n",
"network_agents:\r\n",
"- agent_type: DumbViewer\r\n",
"- agent_class: DumbViewer\r\n",
" state:\r\n",
" has_tv: false\r\n",
" weight: 1\r\n",
"- agent_type: DumbViewer\r\n",
"- agent_class: DumbViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" weight: 1\r\n",
"- agent_type: HerdViewer\r\n",
"- agent_class: HerdViewer\r\n",
" state:\r\n",
" has_tv: false\r\n",
" weight: 1\r\n",
"- agent_type: HerdViewer\r\n",
"- agent_class: HerdViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" weight: 1\r\n",
@ -136,12 +136,12 @@
"max_time: 300\r\n",
"name: Sim_all_herd\r\n",
"network_agents:\r\n",
"- agent_type: HerdViewer\r\n",
"- agent_class: HerdViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" state_id: neutral\r\n",
" weight: 1\r\n",
"- agent_type: HerdViewer\r\n",
"- agent_class: HerdViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" state_id: neutral\r\n",
@ -163,12 +163,12 @@
"max_time: 300\r\n",
"name: Sim_wise_herd\r\n",
"network_agents:\r\n",
"- agent_type: HerdViewer\r\n",
"- agent_class: HerdViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" state_id: neutral\r\n",
" weight: 1\r\n",
"- agent_type: WiseViewer\r\n",
"- agent_class: WiseViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" weight: 1\r\n",
@ -189,12 +189,12 @@
"max_time: 300\r\n",
"name: Sim_all_wise\r\n",
"network_agents:\r\n",
"- agent_type: WiseViewer\r\n",
"- agent_class: WiseViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" state_id: neutral\r\n",
" weight: 1\r\n",
"- agent_type: WiseViewer\r\n",
"- agent_class: WiseViewer\r\n",
" state:\r\n",
" has_tv: true\r\n",
" weight: 1\r\n",

138
examples/newsspread/NewsSpread.yml
View File

@ -1,138 +0,0 @@
---
default_state: {}
load_module: newsspread
environment_agents: []
environment_params:
prob_neighbor_spread: 0.0
prob_tv_spread: 0.01
interval: 1
max_time: 300
name: Sim_all_dumb
network_agents:
- agent_type: DumbViewer
state:
has_tv: false
weight: 1
- agent_type: DumbViewer
state:
has_tv: true
weight: 1
network_params:
generator: barabasi_albert_graph
n: 500
m: 5
num_trials: 50
---
default_state: {}
load_module: newsspread
environment_agents: []
environment_params:
prob_neighbor_spread: 0.0
prob_tv_spread: 0.01
interval: 1
max_time: 300
name: Sim_half_herd
network_agents:
- agent_type: DumbViewer
state:
has_tv: false
weight: 1
- agent_type: DumbViewer
state:
has_tv: true
weight: 1
- agent_type: HerdViewer
state:
has_tv: false
weight: 1
- agent_type: HerdViewer
state:
has_tv: true
weight: 1
network_params:
generator: barabasi_albert_graph
n: 500
m: 5
num_trials: 50
---
default_state: {}
load_module: newsspread
environment_agents: []
environment_params:
prob_neighbor_spread: 0.0
prob_tv_spread: 0.01
interval: 1
max_time: 300
name: Sim_all_herd
network_agents:
- agent_type: HerdViewer
state:
has_tv: true
state_id: neutral
weight: 1
- agent_type: HerdViewer
state:
has_tv: true
state_id: neutral
weight: 1
network_params:
generator: barabasi_albert_graph
n: 500
m: 5
num_trials: 50
---
default_state: {}
load_module: newsspread
environment_agents: []
environment_params:
prob_neighbor_spread: 0.0
prob_tv_spread: 0.01
prob_neighbor_cure: 0.1
interval: 1
max_time: 300
name: Sim_wise_herd
network_agents:
- agent_type: HerdViewer
state:
has_tv: true
state_id: neutral
weight: 1
- agent_type: WiseViewer
state:
has_tv: true
weight: 1
network_params:
generator: barabasi_albert_graph
n: 500
m: 5
num_trials: 50
---
default_state: {}
load_module: newsspread
environment_agents: []
environment_params:
prob_neighbor_spread: 0.0
prob_tv_spread: 0.01
prob_neighbor_cure: 0.1
interval: 1
max_time: 300
name: Sim_all_wise
network_agents:
- agent_type: WiseViewer
state:
has_tv: true
state_id: neutral
weight: 1
- agent_type: WiseViewer
state:
has_tv: true
weight: 1
network_params:
generator: barabasi_albert_graph
n: 500
m: 5
network_params:
generator: barabasi_albert_graph
n: 500
m: 5
num_trials: 50

86
examples/newsspread/newsspread.py
View File

@ -1,86 +0,0 @@
from soil.agents import FSM, state, default_state, prob
import logging
class DumbViewer(FSM):
'''
A viewer that gets infected via TV (if it has one) and tries to infect
its neighbors once it's infected.
'''
defaults = {
'prob_neighbor_spread': 0.5,
'prob_tv_spread': 0.1,
}
@default_state
@state
def neutral(self):
if self['has_tv']:
if prob(self.env['prob_tv_spread']):
return self.infected
@state
def infected(self):
for neighbor in self.get_neighboring_agents(state_id=self.neutral.id):
if prob(self.env['prob_neighbor_spread']):
neighbor.infect()
def infect(self):
'''
This is not a state. It is a function that other agents can use to try to
infect this agent. DumbViewer always gets infected, but other agents like
HerdViewer might not become infected right away
'''
self.set_state(self.infected)
class HerdViewer(DumbViewer):
'''
A viewer whose probability of infection depends on the state of its neighbors.
'''
def infect(self):
'''Notice again that this is NOT a state. See DumbViewer.infect for reference'''
infected = self.count_neighboring_agents(state_id=self.infected.id)
total = self.count_neighboring_agents()
prob_infect = self.env['prob_neighbor_spread'] * infected/total
self.debug('prob_infect', prob_infect)
if prob(prob_infect):
self.set_state(self.infected)
class WiseViewer(HerdViewer):
'''
A viewer that can change its mind.
'''
defaults = {
'prob_neighbor_spread': 0.5,
'prob_neighbor_cure': 0.25,
'prob_tv_spread': 0.1,
}
@state
def cured(self):
prob_cure = self.env['prob_neighbor_cure']
for neighbor in self.get_neighboring_agents(state_id=self.infected.id):
if prob(prob_cure):
try:
neighbor.cure()
except AttributeError:
self.debug('Viewer {} cannot be cured'.format(neighbor.id))
def cure(self):
self.set_state(self.cured.id)
@state
def infected(self):
cured = max(self.count_neighboring_agents(self.cured.id),
1.0)
infected = max(self.count_neighboring_agents(self.infected.id),
1.0)
prob_cure = self.env['prob_neighbor_cure'] * (cured/infected)
if prob(prob_cure):
return self.cured
return self.set_state(super().infected)

134
examples/newsspread/newsspread_sim.py Normal file
View File

@ -0,0 +1,134 @@
from soil.agents import FSM, NetworkAgent, state, default_state, prob
from soil.parameters import *
import logging
from soil.environment import Environment
class DumbViewer(FSM, NetworkAgent):
"""
A viewer that gets infected via TV (if it has one) and tries to infect
its neighbors once it's infected.
"""
has_been_infected: bool = False
has_tv: bool = False
@default_state
@state
def neutral(self):
if self.has_tv:
if self.prob(self.get("prob_tv_spread")):
return self.infected
if self.has_been_infected:
return self.infected
@state
def infected(self):
for neighbor in self.get_neighbors(state_id=self.neutral.id):
if self.prob(self.get("prob_neighbor_spread")):
neighbor.infect()
def infect(self):
"""
This is not a state. It is a function that other agents can use to try to
infect this agent. DumbViewer always gets infected, but other agents like
HerdViewer might not become infected right away
"""
self.has_been_infected = True
class HerdViewer(DumbViewer):
"""
A viewer whose probability of infection depends on the state of its neighbors.
"""
def infect(self):
"""Notice again that this is NOT a state. See DumbViewer.infect for reference"""
infected = self.count_neighbors(state_id=self.infected.id)
total = self.count_neighbors()
prob_infect = self.get("prob_neighbor_spread") * infected / total
self.debug("prob_infect", prob_infect)
if self.prob(prob_infect):
self.has_been_infected = True
class WiseViewer(HerdViewer):
"""
A viewer that can change its mind.
"""
@state
def cured(self):
prob_cure = self.get("prob_neighbor_cure")
for neighbor in self.get_neighbors(state_id=self.infected.id):
if self.prob(prob_cure):
try:
neighbor.cure()
except AttributeError:
self.debug("Viewer {} cannot be cured".format(neighbor.id))
def cure(self):
self.has_been_cured = True
@state
def infected(self):
if self.has_been_cured:
return self.cured
cured = max(self.count_neighbors(self.cured.id), 1.0)
infected = max(self.count_neighbors(self.infected.id), 1.0)
prob_cure = self.get("prob_neighbor_cure") * (cured / infected)
if self.prob(prob_cure):
return self.cured
class NewsSpread(Environment):
ratio_dumb: probability = 1,
ratio_herd: probability = 0,
ratio_wise: probability = 0,
prob_tv_spread: probability = 0.1,
prob_neighbor_spread: probability = 0.1,
prob_neighbor_cure: probability = 0.05,
def init(self):
self.populate_network([DumbViewer, HerdViewer, WiseViewer],
[self.ratio_dumb, self.ratio_herd, self.ratio_wise])
from itertools import product
from soil import Simulation
# We want to investigate the effect of different agent distributions on the spread of news.
# To do that, we will run different simulations, with a varying ratio of DumbViewers, HerdViewers, and WiseViewers
# Because the effect of these agents might also depend on the network structure, we will run our simulations on two different networks:
# one with a small-world structure and one with a connected structure.
counter = 0
for [r1, r2] in product([0, 0.5, 1.0], repeat=2):
for (generator, netparams) in {
"barabasi_albert_graph": {"m": 5},
"erdos_renyi_graph": {"p": 0.1},
}.items():
print(r1, r2, 1-r1-r2, generator)
# Create new simulation
netparams["n"] = 500
Simulation(
name='newspread_sim',
model=NewsSpread,
parameters=dict(
ratio_dumb=r1,
ratio_herd=r2,
ratio_wise=1-r1-r2,
network_generator=generator,
network_params=netparams,
prob_neighbor_spread=0,
),
iterations=5,
max_steps=300,
dump=False,
).run()
counter += 1
# Run all the necessary instances
print(f"A total of {counter} simulations were run.")

40
examples/programmatic/programmatic.py
View File

@ -1,40 +0,0 @@
'''
Example of a fully programmatic simulation, without definition files.
'''
from soil import Simulation, agents
from networkx import Graph
import logging
def mygenerator():
# Add only a node
G = Graph()
G.add_node(1)
return G
class MyAgent(agents.FSM):
@agents.default_state
@agents.state
def neutral(self):
self.debug('I am running')
if agents.prob(0.2):
self.info('This runs 2/10 times on average')
s = Simulation(name='Programmatic',
network_params={'generator': mygenerator},
num_trials=1,
max_time=100,
agent_type=MyAgent,
dry_run=True)
# By default, logging will only print WARNING logs (and above).
# You need to choose a lower logging level to get INFO/DEBUG traces
logging.basicConfig(level=logging.INFO)
envs = s.run()
# Uncomment this to output the simulation to a YAML file
# s.dump_yaml('simulation.yaml')

53
examples/programmatic/programmatic_sim.py Normal file
View File

@ -0,0 +1,53 @@
"""
Example of a fully programmatic simulation, without definition files.
"""
from soil import Simulation, Environment, agents
from networkx import Graph
import logging
def mygenerator():
# Add only a node
G = Graph()
G.add_node(1)
G.add_node(2)
return G
class MyAgent(agents.NetworkAgent, agents.FSM):
times_run = 0
@agents.default_state
@agents.state
def neutral(self):
self.debug("I am running")
if self.prob(0.2):
self.times_run += 1
self.info("This runs 2/10 times on average")
class ProgrammaticEnv(Environment):
def init(self):
self.create_network(generator=mygenerator)
assert len(self.G)
self.populate_network(agent_class=MyAgent)
self.add_agent_reporter('times_run')
simulation = Simulation(
name="Programmatic",
model=ProgrammaticEnv,
seed='Program',
iterations=1,
max_time=100,
dump=False,
)
if __name__ == "__main__":
# By default, logging will only print WARNING logs (and above).
# You need to choose a lower logging level to get INFO/DEBUG traces
logging.basicConfig(level=logging.INFO)
envs = simulation.run()
for agent in envs[0].agents:
print(agent.times_run)

175
examples/pubcrawl/pubcrawl.py
View File

@ -1,175 +0,0 @@
from soil.agents import FSM, state, default_state
from soil import Environment
from random import random, shuffle
from itertools import islice
import logging
class CityPubs(Environment):
'''Environment with Pubs'''
level = logging.INFO
def __init__(self, *args, number_of_pubs=3, pub_capacity=10, **kwargs):
super(CityPubs, self).__init__(*args, **kwargs)
pubs = {}
for i in range(number_of_pubs):
newpub = {
'name': 'The awesome pub #{}'.format(i),
'open': True,
'capacity': pub_capacity,
'occupancy': 0,
}
pubs[newpub['name']] = newpub
self['pubs'] = pubs
def enter(self, pub_id, *nodes):
'''Agents will try to enter. The pub checks if it is possible'''
try:
pub = self['pubs'][pub_id]
except KeyError:
raise ValueError('Pub {} is not available'.format(pub_id))
if not pub['open'] or (pub['capacity'] < (len(nodes) + pub['occupancy'])):
return False
pub['occupancy'] += len(nodes)
for node in nodes:
node['pub'] = pub_id
return True
def available_pubs(self):
for pub in self['pubs'].values():
if pub['open'] and (pub['occupancy'] < pub['capacity']):
yield pub['name']
def exit(self, pub_id, *node_ids):
'''Agents will notify the pub they want to leave'''
try:
pub = self['pubs'][pub_id]
except KeyError:
raise ValueError('Pub {} is not available'.format(pub_id))
for node_id in node_ids:
node = self.get_agent(node_id)
if pub_id == node['pub']:
del node['pub']
pub['occupancy'] -= 1
class Patron(FSM):
'''Agent that looks for friends to drink with. It will do three things:
1) Look for other patrons to drink with
2) Look for a bar where the agent and other agents in the same group can get in.
3) While in the bar, patrons only drink, until they get drunk and taken home.
'''
level = logging.DEBUG
defaults = {
'pub': None,
'drunk': False,
'pints': 0,
'max_pints': 3,
}
@default_state
@state
def looking_for_friends(self):
'''Look for friends to drink with'''
self.info('I am looking for friends')
available_friends = list(self.get_agents(drunk=False,
pub=None,
state_id=self.looking_for_friends.id))
if not available_friends:
self.info('Life sucks and I\'m alone!')
return self.at_home
befriended = self.try_friends(available_friends)
if befriended:
return self.looking_for_pub
@state
def looking_for_pub(self):
'''Look for a pub that accepts me and my friends'''
if self['pub'] != None:
return self.sober_in_pub
self.debug('I am looking for a pub')
group = list(self.get_neighboring_agents())
for pub in self.env.available_pubs():
self.debug('We\'re trying to get into {}: total: {}'.format(pub, len(group)))
if self.env.enter(pub, self, *group):
self.info('We\'re all {} getting in {}!'.format(len(group), pub))
return self.sober_in_pub
@state
def sober_in_pub(self):
'''Drink up.'''
self.drink()
if self['pints'] > self['max_pints']:
return self.drunk_in_pub
@state
def drunk_in_pub(self):
'''I'm out. Take me home!'''
self.info('I\'m so drunk. Take me home!')
self['drunk'] = True
pass # out drunk
@state
def at_home(self):
'''The end'''
others = self.get_agents(state_id=Patron.at_home.id, limit_neighbors=True)
self.debug('I\'m home. Just like {} of my friends'.format(len(others)))
def drink(self):
self['pints'] += 1
self.debug('Cheers to that')
def kick_out(self):
self.set_state(self.at_home)
def befriend(self, other_agent, force=False):
'''
Try to become friends with another agent. The chances of
success depend on both agents' openness.
'''
if force or self['openness'] > random():
self.env.add_edge(self, other_agent)
self.info('Made some friend {}'.format(other_agent))
return True
return False
def try_friends(self, others):
''' Look for random agents around me and try to befriend them'''
befriended = False
k = int(10*self['openness'])
shuffle(others)
for friend in islice(others, k): # random.choice >= 3.7
if friend == self:
continue
if friend.befriend(self):
self.befriend(friend, force=True)
self.debug('Hooray! new friend: {}'.format(friend.id))
befriended = True
else:
self.debug('{} does not want to be friends'.format(friend.id))
return befriended
class Police(FSM):
'''Simple agent to take drunk people out of pubs.'''
level = logging.INFO
@default_state
@state
def patrol(self):
drunksters = list(self.get_agents(drunk=True,
state_id=Patron.drunk_in_pub.id))
for drunk in drunksters:
self.info('Kicking out the trash: {}'.format(drunk.id))
drunk.kick_out()
else:
self.info('No trash to take out. Too bad.')
if __name__ == '__main__':
from soil import simulation
simulation.run_from_config('pubcrawl.yml',
dry_run=True,
dump=None,
parallel=False)

26
examples/pubcrawl/pubcrawl.yml
View File

@ -1,26 +0,0 @@
---
name: pubcrawl
num_trials: 3
max_time: 10
dump: false
network_params:
# Generate 100 empty nodes. They will be assigned a network agent
generator: empty_graph
n: 30
network_agents:
- agent_type: pubcrawl.Patron
description: Extroverted patron
state:
openness: 1.0
weight: 9
- agent_type: pubcrawl.Patron
description: Introverted patron
state:
openness: 0.1
weight: 1
environment_agents:
- agent_type: pubcrawl.Police
environment_class: pubcrawl.CityPubs
environment_params:
altercations: 0
number_of_pubs: 3

195
examples/pubcrawl/pubcrawl_sim.py Normal file
View File

@ -0,0 +1,195 @@
from soil.agents import FSM, NetworkAgent, state, default_state
from soil import Environment, Simulation, parameters
from itertools import islice
import networkx as nx
import logging
class CityPubs(Environment):
"""Environment with Pubs"""
level = logging.INFO
number_of_pubs: parameters.Integer = 3
ratio_extroverted: parameters.probability = 0.1
pub_capacity: parameters.Integer = 10
def init(self):
self.pubs = {}
for i in range(self.number_of_pubs):
newpub = {
"name": "The awesome pub #{}".format(i),
"open": True,
"capacity": self.pub_capacity,
"occupancy": 0,
}
self.pubs[newpub["name"]] = newpub
self.add_agent(agent_class=Police)
self.populate_network([Patron.w(openness=0.1), Patron.w(openness=1)],
[self.ratio_extroverted, 1-self.ratio_extroverted])
assert all(["agent" in node and isinstance(node["agent"], Patron) for (_, node) in self.G.nodes(data=True)])
def enter(self, pub_id, *nodes):
"""Agents will try to enter. The pub checks if it is possible"""
try:
pub = self["pubs"][pub_id]
except KeyError:
raise ValueError("Pub {} is not available".format(pub_id))
if not pub["open"] or (pub["capacity"] < (len(nodes) + pub["occupancy"])):
return False
pub["occupancy"] += len(nodes)
for node in nodes:
node["pub"] = pub_id
return True
def available_pubs(self):
for pub in self["pubs"].values():
if pub["open"] and (pub["occupancy"] < pub["capacity"]):
yield pub["name"]
def exit(self, pub_id, *node_ids):
"""Agents will notify the pub they want to leave"""
try:
pub = self["pubs"][pub_id]
except KeyError:
raise ValueError("Pub {} is not available".format(pub_id))
for node_id in node_ids:
node = self.get_agent(node_id)
if pub_id == node["pub"]:
del node["pub"]
pub["occupancy"] -= 1
class Patron(FSM, NetworkAgent):
"""Agent that looks for friends to drink with. It will do three things:
1) Look for other patrons to drink with
2) Look for a bar where the agent and other agents in the same group can get in.
3) While in the bar, patrons only drink, until they get drunk and taken home.
"""
level = logging.DEBUG
pub = None
drunk = False
pints = 0
max_pints = 3
kicked_out = False
@default_state
@state
def looking_for_friends(self):
"""Look for friends to drink with"""
self.info("I am looking for friends")
available_friends = list(
self.get_agents(drunk=False, pub=None, state_id=self.looking_for_friends.id)
)
if not available_friends:
self.info("Life sucks and I'm alone!")
return self.at_home
befriended = self.try_friends(available_friends)
if befriended:
return self.looking_for_pub
@state
def looking_for_pub(self):
"""Look for a pub that accepts me and my friends"""
if self["pub"] != None:
return self.sober_in_pub
self.debug("I am looking for a pub")
group = list(self.get_neighbors())
for pub in self.model.available_pubs():
self.debug("We're trying to get into {}: total: {}".format(pub, len(group)))
if self.model.enter(pub, self, *group):
self.info("We're all {} getting in {}!".format(len(group), pub))
return self.sober_in_pub
@state
def sober_in_pub(self):
"""Drink up."""
self.drink()
if self["pints"] > self["max_pints"]:
return self.drunk_in_pub
@state
def drunk_in_pub(self):
"""I'm out. Take me home!"""
self.info("I'm so drunk. Take me home!")
self["drunk"] = True
if self.kicked_out:
return self.at_home
pass # out drun
@state
def at_home(self):
"""The end"""
others = self.get_agents(state_id=Patron.at_home.id, limit_neighbors=True)
self.debug("I'm home. Just like {} of my friends".format(len(others)))
def drink(self):
self["pints"] += 1
self.debug("Cheers to that")
def kick_out(self):
self.kicked_out = True
def befriend(self, other_agent, force=False):
"""
Try to become friends with another agent. The chances of
success depend on both agents' openness.
"""
if force or self["openness"] > self.random.random():
self.add_edge(self, other_agent)
self.info("Made some friend {}".format(other_agent))
return True
return False
def try_friends(self, others):
"""Look for random agents around me and try to befriend them"""
befriended = False
k = int(10 * self["openness"])
self.random.shuffle(others)
for friend in islice(others, k): # random.choice >= 3.7
if friend == self:
continue
if friend.befriend(self):
self.befriend(friend, force=True)
self.debug("Hooray! new friend: {}".format(friend.unique_id))
befriended = True
else:
self.debug("{} does not want to be friends".format(friend.unique_id))
return befriended
class Police(FSM):
"""Simple agent to take drunk people out of pubs."""
level = logging.INFO
@default_state
@state
def patrol(self):
drunksters = list(self.get_agents(drunk=True, state_id=Patron.drunk_in_pub.id))
for drunk in drunksters:
self.info("Kicking out the trash: {}".format(drunk.unique_id))
drunk.kick_out()
else:
self.info("No trash to take out. Too bad.")
sim = Simulation(
model=CityPubs,
name="pubcrawl",
iterations=3,
max_steps=10,
dump=False,
parameters=dict(
network_generator=nx.empty_graph,
network_params={"n": 30},
model=CityPubs,
altercations=0,
number_of_pubs=3,
)
)
if __name__ == "__main__":
sim.run(parallel=False)

14
examples/rabbits/README.md Normal file
View File

@ -0,0 +1,14 @@
There are two similar implementations of this simulation.
- `basic`. Using simple primites
- `improved`. Using more advanced features such as the `time` module to avoid unnecessary computations (i.e., skip steps), and generator functions.
The examples can be run directly in the terminal, and they accept command like arguments.
For example, to enable the CSV exporter and the Summary exporter, while setting `max_time` to `100` and `seed` to `CustomSeed`:
```
python rabbit_agents.py --set max_time=100 --csv -e summary --set 'seed="CustomSeed"'
```
To learn more about how this functionality works, check out the `soil.easy` function.

135
examples/rabbits/rabbit_agents.py
View File

@ -1,135 +0,0 @@
from soil.agents import FSM, state, default_state, BaseAgent, NetworkAgent
from enum import Enum
from random import random, choice
import logging
import math
class Genders(Enum):
male = 'male'
female = 'female'
class RabbitModel(FSM):
defaults = {
'age': 0,
'gender': Genders.male.value,
'mating_prob': 0.001,
'offspring': 0,
}
sexual_maturity = 3 #4*30
life_expectancy = 365 * 3
gestation = 33
pregnancy = -1
max_females = 5
@default_state
@state
def newborn(self):
self.debug(f'I am a newborn at age {self["age"]}')
self['age'] += 1
if self['age'] >= self.sexual_maturity:
self.debug('I am fertile!')
return self.fertile
@state
def fertile(self):
raise Exception("Each subclass should define its fertile state")
@state
def dead(self):
self.info('Agent {} is dying'.format(self.id))
self.die()
class Male(RabbitModel):
@state
def fertile(self):
self['age'] += 1
if self['age'] > self.life_expectancy:
return self.dead
if self['gender'] == Genders.female.value:
return
# Males try to mate
for f in self.get_agents(state_id=Female.fertile.id,
agent_type=Female,
limit_neighbors=False,
limit=self.max_females):
r = random()
if r < self['mating_prob']:
self.impregnate(f)
break # Take a break
def impregnate(self, whom):
whom['pregnancy'] = 0
whom['mate'] = self.id
whom.set_state(whom.pregnant)
self.debug('{} impregnating: {}. {}'.format(self.id, whom.id, whom.state))
class Female(RabbitModel):
@state
def fertile(self):
# Just wait for a Male
pass
@state
def pregnant(self):
self['age'] += 1
if self['age'] > self.life_expectancy:
return self.dead
self['pregnancy'] += 1
self.debug('Pregnancy: {}'.format(self['pregnancy']))
if self['pregnancy'] >= self.gestation:
number_of_babies = int(8+4*random())
self.info('Having {} babies'.format(number_of_babies))
for i in range(number_of_babies):
state = {}
state['gender'] = choice(list(Genders)).value
child = self.env.add_node(self.__class__, state)
self.env.add_edge(self.id, child.id)
self.env.add_edge(self['mate'], child.id)
# self.add_edge()
self.debug('A BABY IS COMING TO LIFE')
self.env['rabbits_alive'] = self.env.get('rabbits_alive', self.topology.number_of_nodes())+1
self.debug('Rabbits alive: {}'.format(self.env['rabbits_alive']))
self['offspring'] += 1
self.env.get_agent(self['mate'])['offspring'] += 1
del self['mate']
self['pregnancy'] = -1
return self.fertile
@state
def dead(self):
super().dead()
if 'pregnancy' in self and self['pregnancy'] > -1:
self.info('A mother has died carrying a baby!!')
class RandomAccident(NetworkAgent):
level = logging.DEBUG
def step(self):
rabbits_total = self.topology.number_of_nodes()
if 'rabbits_alive' not in self.env:
self.env['rabbits_alive'] = 0
rabbits_alive = self.env.get('rabbits_alive', rabbits_total)
prob_death = self.env.get('prob_death', 1e-100)*math.floor(math.log10(max(1, rabbits_alive)))
self.debug('Killing some rabbits with prob={}!'.format(prob_death))
for i in self.env.network_agents:
if i.state['id'] == i.dead.id:
continue
r = random()
if r < prob_death:
self.debug('I killed a rabbit: {}'.format(i.id))
rabbits_alive = self.env['rabbits_alive'] = rabbits_alive -1
self.log('Rabbits alive: {}'.format(self.env['rabbits_alive']))
i.set_state(i.dead)
self.log('Rabbits alive: {}/{}'.format(rabbits_alive, rabbits_total))
if self.count_agents(state_id=RabbitModel.dead.id) == self.topology.number_of_nodes():
self.die()

153
examples/rabbits/rabbit_improved_sim.py Normal file
View File

@ -0,0 +1,153 @@
from soil import FSM, state, default_state, BaseAgent, NetworkAgent, Environment, Simulation
from soil.time import Delta
from enum import Enum
from collections import Counter
import logging
import math
from rabbits_basic_sim import RabbitEnv
class RabbitsImprovedEnv(RabbitEnv):
def init(self):
"""Initialize the environment with the new versions of the agents"""
a1 = self.add_node(Male)
a2 = self.add_node(Female)
a1.add_edge(a2)
self.add_agent(RandomAccident)
class Rabbit(FSM, NetworkAgent):
sexual_maturity = 30
life_expectancy = 300
birth = None
@property
def age(self):
if self.birth is None:
return None
return self.now - self.birth
@default_state
@state
def newborn(self):
self.info("I am a newborn.")
self.birth = self.now
self.offspring = 0
return self.youngling, Delta(self.sexual_maturity - self.age)
@state
def youngling(self):
if self.age >= self.sexual_maturity:
self.info(f"I am fertile! My age is {self.age}")
return self.fertile
@state
def fertile(self):
raise Exception("Each subclass should define its fertile state")
@state
def dead(self):
self.die()
class Male(Rabbit):
max_females = 5
mating_prob = 0.001
@state
def fertile(self):
if self.age > self.life_expectancy:
return self.dead
# Males try to mate
for f in self.model.agents(
agent_class=Female, state_id=Female.fertile.id, limit=self.max_females
):
self.debug("FOUND A FEMALE: ", repr(f), self.mating_prob)
if self.prob(self["mating_prob"]):
f.impregnate(self)
break # Do not try to impregnate other females
class Female(Rabbit):
gestation = 10
conception = None
@state
def fertile(self):
# Just wait for a Male
if self.age > self.life_expectancy:
return self.dead
if self.conception is not None:
return self.pregnant
@property
def pregnancy(self):
if self.conception is None:
return None
return self.now - self.conception
def impregnate(self, male):
self.info(f"impregnated by {repr(male)}")
self.mate = male
self.conception = self.now
self.number_of_babies = int(8 + 4 * self.random.random())
@state
def pregnant(self):
self.debug("I am pregnant")
if self.age > self.life_expectancy:
self.info("Dying before giving birth")
return self.die()
if self.pregnancy >= self.gestation:
self.info("Having {} babies".format(self.number_of_babies))
for i in range(self.number_of_babies):
state = {}
agent_class = self.random.choice([Male, Female])
child = self.model.add_node(agent_class=agent_class, **state)
child.add_edge(self)
if self.mate:
child.add_edge(self.mate)
self.mate.offspring += 1
else:
self.debug("The father has passed away")
self.offspring += 1
self.mate = None
return self.fertile
def die(self):
if self.pregnancy is not None:
self.info("A mother has died carrying a baby!!")
return super().die()
class RandomAccident(BaseAgent):
def step(self):
rabbits_alive = self.model.G.number_of_nodes()
if not rabbits_alive:
return self.die()
prob_death = self.model.get("prob_death", 1e-100) * math.floor(
math.log10(max(1, rabbits_alive))
)
self.debug("Killing some rabbits with prob={}!".format(prob_death))
for i in self.iter_agents(agent_class=Rabbit):
if i.state_id == i.dead.id:
continue
if self.prob(prob_death):
self.info("I killed a rabbit: {}".format(i.id))
rabbits_alive -= 1
i.die()
self.debug("Rabbits alive: {}".format(rabbits_alive))
sim = Simulation(model=RabbitsImprovedEnv, max_time=100, seed="MySeed", iterations=1)
if __name__ == "__main__":
sim.run()

21
examples/rabbits/rabbits.yml
View File

@ -1,21 +0,0 @@
---
load_module: rabbit_agents
name: rabbits_example
max_time: 1000
interval: 1
seed: MySeed
agent_type: rabbit_agents.RabbitModel
environment_agents:
- agent_type: rabbit_agents.RandomAccident
environment_params:
prob_death: 0.001
default_state:
mating_prob: 0.1
topology:
nodes:
- id: 1
agent_type: rabbit_agents.Male
- id: 0
agent_type: rabbit_agents.Female
directed: true
links: []

161
examples/rabbits/rabbits_basic_sim.py Normal file
View File

@ -0,0 +1,161 @@
from soil import FSM, state, default_state, BaseAgent, NetworkAgent, Environment, Simulation, report, parameters as params
from collections import Counter
import logging
import math
class RabbitEnv(Environment):
prob_death: params.probability = 1e-100
def init(self):
a1 = self.add_node(Male)
a2 = self.add_node(Female)
a1.add_edge(a2)
self.add_agent(RandomAccident)
@report
@property
def num_rabbits(self):
return self.count_agents(agent_class=Rabbit)
@report
@property
def num_males(self):
return self.count_agents(agent_class=Male)
@report
@property
def num_females(self):
return self.count_agents(agent_class=Female)
class Rabbit(NetworkAgent, FSM):
sexual_maturity = 30
life_expectancy = 300
@default_state
@state
def newborn(self):
self.info("I am a newborn.")
self.age = 0
self.offspring = 0
return self.youngling
@state
def youngling(self):
self.age += 1
if self.age >= self.sexual_maturity:
self.info(f"I am fertile! My age is {self.age}")
return self.fertile
@state
def fertile(self):
raise Exception("Each subclass should define its fertile state")
@state
def dead(self):
self.die()
class Male(Rabbit):
max_females = 5
mating_prob = 0.001
@state
def fertile(self):
self.age += 1
if self.age > self.life_expectancy:
return self.dead
# Males try to mate
for f in self.model.agents(
agent_class=Female, state_id=Female.fertile.id, limit=self.max_females
):
self.debug("FOUND A FEMALE: ", repr(f), self.mating_prob)
if self.prob(self["mating_prob"]):
f.impregnate(self)
break # Take a break
class Female(Rabbit):
gestation = 10
pregnancy = -1
@state
def fertile(self):
# Just wait for a Male
self.age += 1
if self.age > self.life_expectancy:
return self.dead
if self.pregnancy >= 0:
return self.pregnant
def impregnate(self, male):
self.info(f"impregnated by {repr(male)}")
self.mate = male
self.pregnancy = 0
self.number_of_babies = int(8 + 4 * self.random.random())
@state
def pregnant(self):
self.info("I am pregnant")
self.age += 1
if self.age >= self.life_expectancy:
return self.die()
if self.pregnancy < self.gestation:
self.pregnancy += 1
return
self.info("Having {} babies".format(self.number_of_babies))
for i in range(self.number_of_babies):
state = {}
agent_class = self.random.choice([Male, Female])
child = self.model.add_node(agent_class=agent_class, **state)
child.add_edge(self)
try:
child.add_edge(self.mate)
self.model.agents[self.mate].offspring += 1
except ValueError:
self.debug("The father has passed away")
self.offspring += 1
self.mate = None
self.pregnancy = -1
return self.fertile
def die(self):
if "pregnancy" in self and self["pregnancy"] > -1:
self.info("A mother has died carrying a baby!!")
return super().die()
class RandomAccident(BaseAgent):
def step(self):
rabbits_alive = self.model.G.number_of_nodes()
if not rabbits_alive:
return self.die()
prob_death = self.model.prob_death * math.floor(
math.log10(max(1, rabbits_alive))
)
self.debug("Killing some rabbits with prob={}!".format(prob_death))
for i in self.get_agents(agent_class=Rabbit):
if i.state_id == i.dead.id:
continue
if self.prob(prob_death):
self.info("I killed a rabbit: {}".format(i.id))
rabbits_alive -= 1
i.die()
self.debug("Rabbits alive: {}".format(rabbits_alive))
sim = Simulation(model=RabbitEnv, max_time=100, seed="MySeed", iterations=1)
if __name__ == "__main__":
sim.run()

45
examples/random_delays/random_delays.py
View File

@ -1,45 +0,0 @@
'''
Example of setting a
Example of a fully programmatic simulation, without definition files.
'''
from soil import Simulation, agents
from soil.time import Delta
from random import expovariate
import logging
class MyAgent(agents.FSM):
'''
An agent that first does a ping
'''
defaults = {'pong_counts': 2}
@agents.default_state
@agents.state
def ping(self):
self.info('Ping')
return self.pong, Delta(expovariate(1/16))
@agents.state
def pong(self):
self.info('Pong')
self.pong_counts -= 1
self.info(str(self.pong_counts))
if self.pong_counts < 1:
return self.die()
return None, Delta(expovariate(1/16))
s = Simulation(name='Programmatic',
network_agents=[{'agent_type': MyAgent, 'id': 0}],
topology={'nodes': [{'id': 0}], 'links': []},
num_trials=1,
max_time=100,
agent_type=MyAgent,
dry_run=True)
logging.basicConfig(level=logging.INFO)
envs = s.run()

47
examples/random_delays/random_delays_sim.py Normal file
View File

@ -0,0 +1,47 @@
"""
Example of setting a
Example of a fully programmatic simulation, without definition files.
"""
from soil import Simulation, agents, Environment
from soil.time import Delta
class MyAgent(agents.FSM):
"""
An agent that first does a ping
"""
defaults = {"pong_counts": 2}
@agents.default_state
@agents.state
def ping(self):
self.info("Ping")
return self.pong, Delta(self.random.expovariate(1 / 16))
@agents.state
def pong(self):
self.info("Pong")
self.pong_counts -= 1
self.info(str(self.pong_counts))
if self.pong_counts < 1:
return self.die()
return None, Delta(self.random.expovariate(1 / 16))
class RandomEnv(Environment):
def init(self):
self.add_agent(agent_class=MyAgent)
s = Simulation(
name="Programmatic",
model=RandomEnv,
iterations=1,
max_time=100,
dump=False,
)
envs = s.run()

30
examples/template.yml
View File

@ -1,30 +0,0 @@
---
sampler:
method: "SALib.sample.morris.sample"
N: 10
template:
group: simple
num_trials: 1
interval: 1
max_time: 2
seed: "CompleteSeed!"
dump: false
network_params:
generator: complete_graph
n: 10
network_agents:
- agent_type: CounterModel
weight: "{{ x1 }}"
state:
state_id: 0
- agent_type: AggregatedCounter
weight: "{{ 1 - x1 }}"
environment_params:
name: "{{ x3 }}"
skip_test: true
vars:
bounds:
x1: [0, 1]
x2: [1, 2]
fixed:
x3: ["a", "b", "c"]

208
examples/terrorism/TerroristNetworkModel.py
View File

@ -1,208 +0,0 @@
import random
import networkx as nx
from soil.agents import Geo, NetworkAgent, FSM, state, default_state
from soil import Environment
class TerroristSpreadModel(FSM, Geo):
"""
Settings:
information_spread_intensity
terrorist_additional_influence
min_vulnerability (optional else zero)
max_vulnerability
prob_interaction
"""
def __init__(self, model=None, unique_id=0, state=()):
super().__init__(model=model, unique_id=unique_id, state=state)
self.information_spread_intensity = model.environment_params['information_spread_intensity']
self.terrorist_additional_influence = model.environment_params['terrorist_additional_influence']
self.prob_interaction = model.environment_params['prob_interaction']
if self['id'] == self.civilian.id: # Civilian
self.mean_belief = random.uniform(0.00, 0.5)
elif self['id'] == self.terrorist.id: # Terrorist
self.mean_belief = random.uniform(0.8, 1.00)
elif self['id'] == self.leader.id: # Leader
self.mean_belief = 1.00
else:
raise Exception('Invalid state id: {}'.format(self['id']))
if 'min_vulnerability' in model.environment_params:
self.vulnerability = random.uniform( model.environment_params['min_vulnerability'], model.environment_params['max_vulnerability'] )
else :
self.vulnerability = random.uniform( 0, model.environment_params['max_vulnerability'] )
@state
def civilian(self):
neighbours = list(self.get_neighboring_agents(agent_type=TerroristSpreadModel))
if len(neighbours) > 0:
# Only interact with some of the neighbors
interactions = list(n for n in neighbours if random.random() <= self.prob_interaction)
influence = sum( self.degree(i) for i in interactions )
mean_belief = sum( i.mean_belief * self.degree(i) / influence for i in interactions )
mean_belief = mean_belief * self.information_spread_intensity + self.mean_belief * ( 1 - self.information_spread_intensity )
self.mean_belief = mean_belief * self.vulnerability + self.mean_belief * ( 1 - self.vulnerability )
if self.mean_belief >= 0.8:
return self.terrorist
@state
def leader(self):
self.mean_belief = self.mean_belief ** ( 1 - self.terrorist_additional_influence )
for neighbour in self.get_neighboring_agents(state_id=[self.terrorist.id, self.leader.id]):
if self.betweenness(neighbour) > self.betweenness(self):
return self.terrorist
@state
def terrorist(self):
neighbours = self.get_agents(state_id=[self.terrorist.id, self.leader.id],
agent_type=TerroristSpreadModel,
limit_neighbors=True)
if len(neighbours) > 0:
influence = sum( self.degree(n) for n in neighbours )
mean_belief = sum( n.mean_belief * self.degree(n) / influence for n in neighbours )
mean_belief = mean_belief * self.vulnerability + self.mean_belief * ( 1 - self.vulnerability )
self.mean_belief = self.mean_belief ** ( 1 - self.terrorist_additional_influence )
# Check if there are any leaders in the group
leaders = list(filter(lambda x: x.state.id == self.leader.id, neighbours))
if not leaders:
# Check if this is the potential leader
# Stop once it's found. Otherwise, set self as leader
for neighbour in neighbours:
if self.betweenness(self) < self.betweenness(neighbour):
return
return self.leader
class TrainingAreaModel(FSM, Geo):
"""
Settings:
training_influence
min_vulnerability
Requires TerroristSpreadModel.
"""
def __init__(self, model=None, unique_id=0, state=()):
super().__init__(model=model, unique_id=unique_id, state=state)
self.training_influence = model.environment_params['training_influence']
if 'min_vulnerability' in model.environment_params:
self.min_vulnerability = model.environment_params['min_vulnerability']
else: self.min_vulnerability = 0
@default_state
@state
def terrorist(self):
for neighbour in self.get_neighboring_agents(agent_type=TerroristSpreadModel):
if neighbour.vulnerability > self.min_vulnerability:
neighbour.vulnerability = neighbour.vulnerability ** ( 1 - self.training_influence )
class HavenModel(FSM, Geo):
"""
Settings:
haven_influence
min_vulnerability
max_vulnerability
Requires TerroristSpreadModel.
"""
def __init__(self, model=None, unique_id=0, state=()):
super().__init__(model=model, unique_id=unique_id, state=state)
self.haven_influence = model.environment_params['haven_influence']
if 'min_vulnerability' in model.environment_params:
self.min_vulnerability = model.environment_params['min_vulnerability']
else: self.min_vulnerability = 0
self.max_vulnerability = model.environment_params['max_vulnerability']
def get_occupants(self, **kwargs):
return self.get_neighboring_agents(agent_type=TerroristSpreadModel, **kwargs)
@state
def civilian(self):
civilians = self.get_occupants(state_id=self.civilian.id)
if not civilians:
return self.terrorist
for neighbour in self.get_occupants():
if neighbour.vulnerability > self.min_vulnerability:
neighbour.vulnerability = neighbour.vulnerability * ( 1 - self.haven_influence )
return self.civilian
@state
def terrorist(self):
for neighbour in self.get_occupants():
if neighbour.vulnerability < self.max_vulnerability:
neighbour.vulnerability = neighbour.vulnerability ** ( 1 - self.haven_influence )
return self.terrorist
class TerroristNetworkModel(TerroristSpreadModel):
"""
Settings:
sphere_influence
vision_range
weight_social_distance
weight_link_distance
"""
def __init__(self, model=None, unique_id=0, state=()):
super().__init__(model=model, unique_id=unique_id, state=state)
self.vision_range = model.environment_params['vision_range']
self.sphere_influence = model.environment_params['sphere_influence']
self.weight_social_distance = model.environment_params['weight_social_distance']
self.weight_link_distance = model.environment_params['weight_link_distance']
@state
def terrorist(self):
self.update_relationships()
return super().terrorist()
@state
def leader(self):
self.update_relationships()
return super().leader()
def update_relationships(self):
if self.count_neighboring_agents(state_id=self.civilian.id) == 0:
close_ups = set(self.geo_search(radius=self.vision_range, agent_type=TerroristNetworkModel))
step_neighbours = set(self.ego_search(self.sphere_influence, agent_type=TerroristNetworkModel, center=False))
neighbours = set(agent.id for agent in self.get_neighboring_agents(agent_type=TerroristNetworkModel))
search = (close_ups | step_neighbours) - neighbours
for agent in self.get_agents(search):
social_distance = 1 / self.shortest_path_length(agent.id)
spatial_proximity = ( 1 - self.get_distance(agent.id) )
prob_new_interaction = self.weight_social_distance * social_distance + self.weight_link_distance * spatial_proximity
if agent['id'] == agent.civilian.id and random.random() < prob_new_interaction:
self.add_edge(agent)
break
def get_distance(self, target):
source_x, source_y = nx.get_node_attributes(self.topology, 'pos')[self.id]
target_x, target_y = nx.get_node_attributes(self.topology, 'pos')[target]
dx = abs( source_x - target_x )
dy = abs( source_y - target_y )
return ( dx ** 2 + dy ** 2 ) ** ( 1 / 2 )
def shortest_path_length(self, target):
try:
return nx.shortest_path_length(self.topology, self.id, target)
except nx.NetworkXNoPath:
return float('inf')

63
examples/terrorism/TerroristNetworkModel.yml
View File

@ -1,63 +0,0 @@
name: TerroristNetworkModel_sim
load_module: TerroristNetworkModel
max_time: 150
num_trials: 1
network_params:
generator: random_geometric_graph
radius: 0.2
# generator: geographical_threshold_graph
# theta: 20
n: 100
network_agents:
- agent_type: TerroristNetworkModel
weight: 0.8
state:
id: civilian # Civilians
- agent_type: TerroristNetworkModel
weight: 0.1
state:
id: leader # Leaders
- agent_type: TrainingAreaModel
weight: 0.05
state:
id: terrorist # Terrorism
- agent_type: HavenModel
weight: 0.05
state:
id: civilian # Civilian
environment_params:
# TerroristSpreadModel
information_spread_intensity: 0.7
terrorist_additional_influence: 0.035
max_vulnerability: 0.7
prob_interaction: 0.5
# TrainingAreaModel and HavenModel
training_influence: 0.20
haven_influence: 0.20
# TerroristNetworkModel
vision_range: 0.30
sphere_influence: 2
weight_social_distance: 0.035
weight_link_distance: 0.035
visualization_params:
# Icons downloaded from https://www.iconfinder.com/
shape_property: agent
shapes:
TrainingAreaModel: target
HavenModel: home
TerroristNetworkModel: person
colors:
- attr_id: civilian
color: '#40de40'
- attr_id: terrorist
color: red
- attr_id: leader
color: '#c16a6a'
background_image: 'map_4800x2860.jpg'
background_opacity: '0.9'
background_filter_color: 'blue'
skip_test: true # This simulation takes too long for automated tests.

341
examples/terrorism/TerroristNetworkModel_sim.py Normal file
View File

@ -0,0 +1,341 @@
import networkx as nx
from soil.agents import Geo, NetworkAgent, FSM, custom, state, default_state
from soil import Environment, Simulation
from soil.parameters import *
from soil.utils import int_seed
class TerroristEnvironment(Environment):
n: Integer = 100
radius: Float = 0.2
information_spread_intensity: probability = 0.7
terrorist_additional_influence: probability = 0.03
terrorist_additional_influence: probability = 0.035
max_vulnerability: probability = 0.7
prob_interaction: probability = 0.5
# TrainingAreaModel and HavenModel
training_influence: probability = 0.20
haven_influence: probability = 0.20
# TerroristNetworkModel
vision_range: Float = 0.30
sphere_influence: Integer = 2
weight_social_distance: Float = 0.035
weight_link_distance: Float = 0.035
ratio_civil: probability = 0.8
ratio_leader: probability = 0.1
ratio_training: probability = 0.05
ratio_haven: probability = 0.05
def init(self):
self.create_network(generator=self.generator, n=self.n, radius=self.radius)
self.populate_network([
TerroristNetworkModel.w(state_id='civilian'),
TerroristNetworkModel.w(state_id='leader'),
TrainingAreaModel,
HavenModel
], [self.ratio_civil, self.ratio_leader, self.ratio_training, self.ratio_haven])
def generator(self, *args, **kwargs):
return nx.random_geometric_graph(*args, **kwargs, seed=int_seed(self._seed))
class TerroristSpreadModel(FSM, Geo):
"""
Settings:
information_spread_intensity
terrorist_additional_influence
min_vulnerability (optional else zero)
max_vulnerability
"""
information_spread_intensity = 0.1
terrorist_additional_influence = 0.1
min_vulnerability = 0
max_vulnerability = 1
def init(self):
if self.state_id == self.civilian.id: # Civilian
self.mean_belief = self.model.random.uniform(0.00, 0.5)
elif self.state_id == self.terrorist.id: # Terrorist
self.mean_belief = self.random.uniform(0.8, 1.00)
elif self.state_id == self.leader.id: # Leader
self.mean_belief = 1.00
else:
raise Exception("Invalid state id: {}".format(self["id"]))
self.vulnerability = self.random.uniform(
self.get("min_vulnerability", 0), self.get("max_vulnerability", 1)
)
@default_state
@state
def civilian(self):
neighbours = list(self.get_neighbors(agent_class=TerroristSpreadModel))
if len(neighbours) > 0:
# Only interact with some of the neighbors
interactions = list(
n for n in neighbours if self.random.random() <= self.model.prob_interaction
)
influence = sum(self.degree(i) for i in interactions)
mean_belief = sum(
i.mean_belief * self.degree(i) / influence for i in interactions
)
mean_belief = (
mean_belief * self.information_spread_intensity
+ self.mean_belief * (1 - self.information_spread_intensity)
)
self.mean_belief = mean_belief * self.vulnerability + self.mean_belief * (
1 - self.vulnerability
)
if self.mean_belief >= 0.8:
return self.terrorist
@state
def leader(self):
self.mean_belief = self.mean_belief ** (1 - self.terrorist_additional_influence)
for neighbour in self.get_neighbors(
state_id=[self.terrorist.id, self.leader.id]
):
if self.betweenness(neighbour) > self.betweenness(self):
return self.terrorist
@state
def terrorist(self):
neighbours = self.get_agents(
state_id=[self.terrorist.id, self.leader.id],
agent_class=TerroristSpreadModel,
limit_neighbors=True,
)
if len(neighbours) > 0:
influence = sum(self.degree(n) for n in neighbours)
mean_belief = sum(
n.mean_belief * self.degree(n) / influence for n in neighbours
)
mean_belief = mean_belief * self.vulnerability + self.mean_belief * (
1 - self.vulnerability
)
self.mean_belief = self.mean_belief ** (
1 - self.terrorist_additional_influence
)
# Check if there are any leaders in the group
leaders = list(filter(lambda x: x.state_id == self.leader.id, neighbours))
if not leaders:
# Check if this is the potential leader
# Stop once it's found. Otherwise, set self as leader
for neighbour in neighbours:
if self.betweenness(self) < self.betweenness(neighbour):
return
return self.leader
def ego_search(self, steps=1, center=False, agent=None, **kwargs):
"""Get a list of nodes in the ego network of *node* of radius *steps*"""
node = agent.node_id if agent else self.node_id
G = self.subgraph(**kwargs)
return nx.ego_graph(G, node, center=center, radius=steps).nodes()
def degree(self, agent, force=False):
if (
force
or (not hasattr(self.model, "_degree"))
or getattr(self.model, "_last_step", 0) < self.now
):
self.model._degree = nx.degree_centrality(self.G)
self.model._last_step = self.now
return self.model._degree[agent.node_id]
def betweenness(self, agent, force=False):
if (
force
or (not hasattr(self.model, "_betweenness"))
or getattr(self.model, "_last_step", 0) < self.now
):
self.model._betweenness = nx.betweenness_centrality(self.G)
self.model._last_step = self.now
return self.model._betweenness[agent.node_id]
class TrainingAreaModel(FSM, Geo):
"""
Settings:
training_influence
min_vulnerability
Requires TerroristSpreadModel.
"""
training_influence = 0.1
min_vulnerability = 0
def init(self):
self.mean_believe = 1
self.vulnerability = 0
@default_state
@state
def terrorist(self):
for neighbour in self.get_neighbors(agent_class=TerroristSpreadModel):
if neighbour.vulnerability > self.min_vulnerability:
neighbour.vulnerability = neighbour.vulnerability ** (
1 - self.training_influence
)
class HavenModel(FSM, Geo):
"""
Settings:
haven_influence
min_vulnerability
max_vulnerability
Requires TerroristSpreadModel.
"""
min_vulnerability = 0
haven_influence = 0.1
max_vulnerability = 0.5
def init(self):
self.mean_believe = 0
self.vulnerability = 0
def get_occupants(self, **kwargs):
return self.get_neighbors(agent_class=TerroristSpreadModel,
**kwargs)
@default_state
@state
def civilian(self):
civilians = self.get_occupants(state_id=self.civilian.id)
if not civilians:
return self.terrorist
for neighbour in self.get_occupants():
if neighbour.vulnerability > self.min_vulnerability:
neighbour.vulnerability = neighbour.vulnerability * (
1 - self.haven_influence
)
return self.civilian
@state
def terrorist(self):
for neighbour in self.get_occupants():
if neighbour.vulnerability < self.max_vulnerability:
neighbour.vulnerability = neighbour.vulnerability ** (
1 - self.haven_influence
)
return self.terrorist
class TerroristNetworkModel(TerroristSpreadModel):
"""
Settings:
sphere_influence
vision_range
weight_social_distance
weight_link_distance
"""
sphere_influence: float = 1
vision_range: float = 1
weight_social_distance: float = 0.5
weight_link_distance: float = 0.2
@state
def terrorist(self):
self.update_relationships()
return super().terrorist()
@state
def leader(self):
self.update_relationships()
return super().leader()
def update_relationships(self):
if self.count_neighbors(state_id=self.civilian.id) == 0:
close_ups = set(
self.geo_search(
radius=self.vision_range, agent_class=TerroristNetworkModel
)
)
step_neighbours = set(
self.ego_search(
self.sphere_influence,
agent_class=TerroristNetworkModel,
center=False,
)
)
neighbours = set(
agent.unique_id
for agent in self.get_neighbors(agent_class=TerroristNetworkModel)
)
search = (close_ups | step_neighbours) - neighbours
for agent in self.get_agents(search):
social_distance = 1 / self.shortest_path_length(agent.unique_id)
spatial_proximity = 1 - self.get_distance(agent.unique_id)
prob_new_interaction = (
self.weight_social_distance * social_distance
+ self.weight_link_distance * spatial_proximity
)
if (
agent.state_id == "civilian"
and self.random.random() < prob_new_interaction
):
self.add_edge(agent)
break
def get_distance(self, target):
source_x, source_y = nx.get_node_attributes(self.G, "pos")[self.unique_id]
target_x, target_y = nx.get_node_attributes(self.G, "pos")[target]
dx = abs(source_x - target_x)
dy = abs(source_y - target_y)
return (dx**2 + dy**2) ** (1 / 2)
def shortest_path_length(self, target):
try:
return nx.shortest_path_length(self.G, self.unique_id, target)
except nx.NetworkXNoPath:
return float("inf")
sim = Simulation(
model=TerroristEnvironment,
iterations=1,
name="TerroristNetworkModel_sim",
max_steps=150,
seed="default2",
skip_test=False,
dump=False,
)
# TODO: integrate visualization
# visualization_params:
# # Icons downloaded from https://www.iconfinder.com/
# shape_property: agent
# shapes:
# TrainingAreaModel: target
# HavenModel: home
# TerroristNetworkModel: person
# colors:
# - attr_id: civilian
# color: '#40de40'
# - attr_id: terrorist
# color: red
# - attr_id: leader
# color: '#c16a6a'
# background_image: 'map_4800x2860.jpg'
# background_opacity: '0.9'
# background_filter_color: 'blue'

14
examples/torvalds.yml
View File

@ -1,14 +0,0 @@
---
name: torvalds_example
max_time: 10
interval: 2
agent_type: CounterModel
default_state:
skill_level: 'beginner'
network_params:
path: 'torvalds.edgelist'
states:
Torvalds:
skill_level: 'God'
balkian:
skill_level: 'developer'

25
examples/torvalds_sim.py Normal file
View File

@ -0,0 +1,25 @@
from soil import Environment, Simulation, CounterModel, report
# Get directory path for current file
import os, sys, inspect
currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
class TorvaldsEnv(Environment):
def init(self):
self.create_network(path=os.path.join(currentdir, 'torvalds.edgelist'))
self.populate_network(CounterModel, skill_level='beginner')
self.agent(node_id="Torvalds").skill_level = 'God'
self.agent(node_id="balkian").skill_level = 'developer'
self.add_agent_reporter("times")
@report
def god_developers(self):
return self.count_agents(skill_level='God')
sim = Simulation(name='torvalds_example',
max_steps=10,
interval=2,
model=TorvaldsEnv)

10
examples/tutorial/soil_tutorial.html
View File

@ -12330,11 +12330,11 @@ Notice how node 0 is the only one with a TV.</p>
<span class="n">sim</span> <span class="o">=</span> <span class="n">soil</span><span class="o">.</span><span class="n">Simulation</span><span class="p">(</span><span class="n">topology</span><span class="o">=</span><span class="n">G</span><span class="p">,</span>
<span class="n">num_trials</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
<span class="n">max_time</span><span class="o">=</span><span class="n">MAX_TIME</span><span class="p">,</span>
<span class="n">environment_agents</span><span class="o">=</span><span class="p">[{</span><span class="s1">&#39;agent_type&#39;</span><span class="p">:</span> <span class="n">NewsEnvironmentAgent</span><span class="p">,</span>
<span class="n">environment_agents</span><span class="o">=</span><span class="p">[{</span><span class="s1">&#39;agent_class&#39;</span><span class="p">:</span> <span class="n">NewsEnvironmentAgent</span><span class="p">,</span>
<span class="s1">&#39;state&#39;</span><span class="p">:</span> <span class="p">{</span>
<span class="s1">&#39;event_time&#39;</span><span class="p">:</span> <span class="n">EVENT_TIME</span>
<span class="p">}}],</span>
<span class="n">network_agents</span><span class="o">=</span><span class="p">[{</span><span class="s1">&#39;agent_type&#39;</span><span class="p">:</span> <span class="n">NewsSpread</span><span class="p">,</span>
<span class="n">network_agents</span><span class="o">=</span><span class="p">[{</span><span class="s1">&#39;agent_class&#39;</span><span class="p">:</span> <span class="n">NewsSpread</span><span class="p">,</span>
<span class="s1">&#39;weight&#39;</span><span class="p">:</span> <span class="mi">1</span><span class="p">}],</span>
<span class="n">states</span><span class="o">=</span><span class="p">{</span><span class="mi">0</span><span class="p">:</span> <span class="p">{</span><span class="s1">&#39;has_tv&#39;</span><span class="p">:</span> <span class="kc">True</span><span class="p">}},</span>
<span class="n">default_state</span><span class="o">=</span><span class="p">{</span><span class="s1">&#39;has_tv&#39;</span><span class="p">:</span> <span class="kc">False</span><span class="p">},</span>
@ -12468,14 +12468,14 @@ For this demo, we will use a python dictionary:</p>
<span class="p">},</span>
<span class="s1">&#39;network_agents&#39;</span><span class="p">:</span> <span class="p">[</span>
<span class="p">{</span>
<span class="s1">&#39;agent_type&#39;</span><span class="p">:</span> <span class="n">NewsSpread</span><span class="p">,</span>
<span class="s1">&#39;agent_class&#39;</span><span class="p">:</span> <span class="n">NewsSpread</span><span class="p">,</span>
<span class="s1">&#39;weight&#39;</span><span class="p">:</span> <span class="mi">1</span><span class="p">,</span>
<span class="s1">&#39;state&#39;</span><span class="p">:</span> <span class="p">{</span>
<span class="s1">&#39;has_tv&#39;</span><span class="p">:</span> <span class="kc">False</span>
<span class="p">}</span>
<span class="p">},</span>
<span class="p">{</span>
<span class="s1">&#39;agent_type&#39;</span><span class="p">:</span> <span class="n">NewsSpread</span><span class="p">,</span>
<span class="s1">&#39;agent_class&#39;</span><span class="p">:</span> <span class="n">NewsSpread</span><span class="p">,</span>
<span class="s1">&#39;weight&#39;</span><span class="p">:</span> <span class="mi">2</span><span class="p">,</span>
<span class="s1">&#39;state&#39;</span><span class="p">:</span> <span class="p">{</span>
<span class="s1">&#39;has_tv&#39;</span><span class="p">:</span> <span class="kc">True</span>
@ -12483,7 +12483,7 @@ For this demo, we will use a python dictionary:</p>
<span class="p">}</span>
<span class="p">],</span>
<span class="s1">&#39;environment_agents&#39;</span><span class="p">:[</span>
<span class="p">{</span><span class="s1">&#39;agent_type&#39;</span><span class="p">:</span> <span class="n">NewsEnvironmentAgent</span><span class="p">,</span>
<span class="p">{</span><span class="s1">&#39;agent_class&#39;</span><span class="p">:</span> <span class="n">NewsEnvironmentAgent</span><span class="p">,</span>
<span class="s1">&#39;state&#39;</span><span class="p">:</span> <span class="p">{</span>
<span class="s1">&#39;event_time&#39;</span><span class="p">:</span> <span class="mi">10</span>
<span class="p">}</span>

6076
examples/tutorial/soil_tutorial.ipynb
View File

File diff suppressed because one or more lines are too long

10
requirements.txt
View File

@ -2,8 +2,12 @@ networkx>=2.5
numpy
matplotlib
pyyaml>=5.1
pandas>=0.23
pandas>=1
SALib>=1.3
Jinja2
Mesa>=0.8
tsih>=0.1.9
Mesa>=1.2
pydantic>=1.9
sqlalchemy>=1.4
typing-extensions>=4.4
annotated-types>=0.4
tqdm>=4.64

4
setup.cfg
View File

@ -1,3 +1,7 @@
[metadata]
long_description = file: README.md
long_description_content_type = text/markdown
[aliases]
test=pytest
[tool:pytest]

10
setup.py
View File

@ -44,14 +44,20 @@ setup(
'Operating System :: MacOS :: MacOS X',
'Operating System :: Microsoft :: Windows',
'Operating System :: POSIX',
'Programming Language :: Python :: 3'],
"Programming Language :: Python :: 3 :: Only",
"Programming Language :: Python :: 3.8",
"Programming Language :: Python :: 3.9",
"Programming Language :: Python :: 3.10",
],
install_requires=install_reqs,
extras_require=extras_require,
tests_require=test_reqs,
setup_requires=['pytest-runner', ],
pytest_plugins = ['pytest_profiling'],
include_package_data=True,
python_requires=">=3.8",
entry_points={
'console_scripts':
['soil = soil.__init__:main',
['soil = soil.__main__:main',
'soil-web = soil.web.__init__:main']
})

2
soil/VERSION
View File

@ -1 +1 @@
0.20.8
1.0.0rc2

292
soil/__init__.py
View File

@ -1,8 +1,12 @@
from __future__ import annotations
import importlib
from importlib.resources import path
import sys
import os
import pdb
import logging
import traceback
from contextlib import contextmanager
from .version import __version__
@ -11,89 +15,273 @@ try:
except NameError:
basestring = str
from pathlib import Path
from .analysis import *
from .agents import *
from . import agents
from .simulation import *
from .environment import Environment
from .environment import Environment, EventedEnvironment
from .datacollection import SoilCollector
from . import serialization
from . import analysis
from .utils import logger
from .time import *
from .decorators import *
def main(
cfg="simulation.yml",
exporters=None,
num_processes=1,
output="soil_output",
*,
debug=False,
pdb=False,
**kwargs,
):
sim = None
if isinstance(cfg, Simulation):
sim = cfg
def main():
import argparse
from . import simulation
logger.info('Running SOIL version: {}'.format(__version__))
logger.info("Running SOIL version: {}".format(__version__))
parser = argparse.ArgumentParser(description='Run a SOIL simulation')
parser.add_argument('file', type=str,
parser = argparse.ArgumentParser(description="Run a SOIL simulation")
parser.add_argument(
"file",
type=str,
nargs="?",
default='simulation.yml',
help='Configuration file for the simulation (e.g., YAML or JSON)')
parser.add_argument('--version', action='store_true',
help='Show version info and exit')
parser.add_argument('--module', '-m', type=str,
help='file containing the code of any custom agents.')
parser.add_argument('--dry-run', '--dry', action='store_true',
help='Do not store the results of the simulation.')
parser.add_argument('--pdb', action='store_true',
help='Use a pdb console in case of exception.')
parser.add_argument('--graph', '-g', action='store_true',
help='Dump GEXF graph. Defaults to false.')
parser.add_argument('--csv', action='store_true',
help='Dump history in CSV format. Defaults to false.')
parser.add_argument('--level', type=str,
help='Logging level')
parser.add_argument('--output', '-o', type=str, default="soil_output",
help='folder to write results to. It defaults to the current directory.')
parser.add_argument('--synchronous', action='store_true',
help='Run trials serially and synchronously instead of in parallel. Defaults to false.')
parser.add_argument('-e', '--exporter', action='append',
help='Export environment and/or simulations using this exporter')
default=cfg if sim is None else "",
help="Configuration file for the simulation (e.g., YAML or JSON)",
)
parser.add_argument(
"--version", action="store_true", help="Show version info and exit"
)
parser.add_argument(
"--module",
"-m",
type=str,
help="file containing the code of any custom agents.",
)
parser.add_argument(
"--dry-run",
"--dry",
action="store_true",
help="Do not run the simulation",
)
parser.add_argument(
"--no-dump",
action="store_true",
help="Do not store the results of the simulation to disk, show in terminal instead.",
)
parser.add_argument(
"--pdb", action="store_true", help="Use a pdb console in case of exception."
)
parser.add_argument(
"--debug",
action="store_true",
help="Run a customized version of a pdb console to debug a simulation.",
)
parser.add_argument(
"--graph",
"-g",
action="store_true",
help="Dump each iteration's network topology as a GEXF graph. Defaults to false.",
)
parser.add_argument(
"--csv",
action="store_true",
help="Dump all data collected in CSV format. Defaults to false.",
)
parser.add_argument("--level", type=str, help="Logging level")
parser.add_argument(
"--output",
"-o",
type=str,
default=output or "soil_output",
help="folder to write results to. It defaults to the current directory.",
)
parser.add_argument(
"--num-processes",
default=num_processes,
help="Number of processes to use for parallel execution. Defaults to 1.",
)
parser.add_argument(
"-e",
"--exporter",
action="append",
default=[],
help="Export environment and/or simulations using this exporter",
)
parser.add_argument(
"--max_time",
default="-1",
help="Set maximum time for the simulation to run. ",
)
parser.add_argument(
"--max_steps",
default="-1",
help="Set maximum number of steps for the simulation to run.",
)
parser.add_argument(
"--iterations",
default="",
help="Set maximum number of iterations (runs) for the simulation.",
)
parser.add_argument(
"--seed",
default=None,
help="Manually set a seed for the simulation.",
)
parser.add_argument(
"--only-convert",
"--convert",
action="store_true",
help="Do not run the simulation, only convert the configuration file(s) and output them.",
)
parser.add_argument(
"--set",
metavar="KEY=VALUE",
action="append",
help="Set a number of parameters that will be passed to the simulation."
"(do not put spaces before or after the = sign). "
"If a value contains spaces, you should define "
"it with double quotes: "
'foo="this is a sentence". Note that '
"values are always treated as strings.",
)
args = parser.parse_args()
logging.basicConfig(level=getattr(logging, (args.level or 'INFO').upper()))
level = getattr(logging, (args.level or "INFO").upper())
logger.setLevel(level)
if args.version:
return
exporters = exporters or [
"default",
]
for exp in args.exporter:
if exp not in exporters:
exporters.append(exp)
if args.csv:
exporters.append("csv")
if args.graph:
exporters.append("gexf")
if os.getcwd() not in sys.path:
sys.path.append(os.getcwd())
if args.module:
importlib.import_module(args.module)
if output is None:
output = args.output
logger.info('Loading config file: {}'.format(args.file))
debug = debug or args.debug
if args.pdb:
if args.pdb or debug:
args.synchronous = True
os.environ["SOIL_POSTMORTEM"] = "true"
res = []
try:
exporters = list(args.exporter or ['default', ])
if args.csv:
exporters.append('csv')
if args.graph:
exporters.append('gexf')
exp_params = {}
if args.dry_run:
exp_params['copy_to'] = sys.stdout
if not os.path.exists(args.file):
logger.error('Please, input a valid file')
return
simulation.run_from_config(args.file,
opts = dict(
dry_run=args.dry_run,
dump=not args.no_dump,
debug=debug,
exporters=exporters,
parallel=(not args.synchronous),
outdir=args.output,
exporter_params=exp_params)
except Exception:
if args.pdb:
pdb.post_mortem()
num_processes=args.num_processes,
level=level,
outdir=output,
exporter_params=exp_params,
**kwargs)
if args.seed is not None:
opts["seed"] = args.seed
if args.iterations:
opts["iterations"] =int(args.iterations)
if sim:
logger.info("Loading simulation instance")
for (k, v) in opts.items():
setattr(sim, k, v)
sims = [sim]
else:
logger.info("Loading config file: {}".format(args.file))
if not os.path.exists(args.file):
logger.error("Please, input a valid file")
return
assert opts["debug"] == debug
sims = list(
simulation.iter_from_file(
args.file,
**opts,
)
)
for sim in sims:
assert sim.debug == debug
if args.set:
for s in args.set:
k, v = s.split("=", 1)[:2]
v = eval(v)
tail, *head = k.rsplit(".", 1)[::-1]
target = sim.parameters
if head:
for part in head[0].split("."):
try:
target = getattr(target, part)
except AttributeError:
target = target[part]
try:
setattr(target, tail, v)
except AttributeError:
target[tail] = v
if args.only_convert:
print(sim.to_yaml())
continue
max_time = float(args.max_time) if args.max_time != "-1" else None
max_steps = float(args.max_steps) if args.max_steps != "-1" else None
res.append(sim.run(max_time=max_time, max_steps=max_steps))
except Exception as ex:
if args.pdb:
from .debugging import post_mortem
print(traceback.format_exc())
post_mortem()
else:
raise
if debug:
from .debugging import set_trace
os.environ["SOIL_DEBUG"] = "true"
set_trace()
return res
@contextmanager
def easy(cfg, pdb=False, debug=False, **kwargs):
try:
return main(cfg, debug=debug, pdb=pdb, **kwargs)[0]
except Exception as e:
if os.environ.get("SOIL_POSTMORTEM"):
from .debugging import post_mortem
print(traceback.format_exc())
post_mortem()
raise
if __name__ == '__main__':
if __name__ == "__main__":
main()

11
soil/__main__.py
View File

@ -1,4 +1,9 @@
from . import main
from . import main as init_main
if __name__ == '__main__':
main()
def main():
init_main()
if __name__ == "__main__":
init_main()

8
soil/agents/BassModel.py
View File

@ -1,4 +1,3 @@
import random
from . import FSM, state, default_state
@ -8,6 +7,7 @@ class BassModel(FSM):
innovation_prob
imitation_prob
"""
sentimentCorrelation = 0
def step(self):
@ -16,13 +16,13 @@ class BassModel(FSM):
@default_state
@state
def innovation(self):
if random.random() < self.innovation_prob:
if self.prob(self.innovation_prob):
self.sentimentCorrelation = 1
return self.aware
else:
aware_neighbors = self.get_neighboring_agents(state_id=self.aware.id)
aware_neighbors = self.get_neighbors(state_id=self.aware.id)
num_neighbors_aware = len(aware_neighbors)
if random.random() < (self['imitation_prob']*num_neighbors_aware):
if self.prob((self.imitation_prob * num_neighbors_aware)):
self.sentimentCorrelation = 1
return self.aware

95
soil/agents/BigMarketModel.py
View File

@ -1,95 +0,0 @@
import random
from . import FSM, state, default_state
class BigMarketModel(FSM):
"""
Settings:
Names:
enterprises [Array]
tweet_probability_enterprises [Array]
Users:
tweet_probability_users
tweet_relevant_probability
tweet_probability_about [Array]
sentiment_about [Array]
"""
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.enterprises = self.env.environment_params['enterprises']
self.type = ""
if self.id < len(self.enterprises): # Enterprises
self.set_state(self.enterprise.id)
self.type = "Enterprise"
self.tweet_probability = environment.environment_params['tweet_probability_enterprises'][self.id]
else: # normal users
self.type = "User"
self.set_state(self.user.id)
self.tweet_probability = environment.environment_params['tweet_probability_users']
self.tweet_relevant_probability = environment.environment_params['tweet_relevant_probability']
self.tweet_probability_about = environment.environment_params['tweet_probability_about'] # List
self.sentiment_about = environment.environment_params['sentiment_about'] # List
@state
def enterprise(self):
if random.random() < self.tweet_probability: # Tweets
aware_neighbors = self.get_neighboring_agents(state_id=self.number_of_enterprises) # Nodes neighbour users
for x in aware_neighbors:
if random.uniform(0,10) < 5:
x.sentiment_about[self.id] += 0.1 # Increments for enterprise
else:
x.sentiment_about[self.id] -= 0.1 # Decrements for enterprise
# Establecemos limites
if x.sentiment_about[self.id] > 1:
x.sentiment_about[self.id] = 1
if x.sentiment_about[self.id]< -1:
x.sentiment_about[self.id] = -1
x.attrs['sentiment_enterprise_%s'% self.enterprises[self.id]] = x.sentiment_about[self.id]
@state
def user(self):
if random.random() < self.tweet_probability: # Tweets
if random.random() < self.tweet_relevant_probability: # Tweets something relevant
# Tweet probability per enterprise
for i in range(len(self.enterprises)):
random_num = random.random()
if random_num < self.tweet_probability_about[i]:
# The condition is fulfilled, sentiments are evaluated towards that enterprise
if self.sentiment_about[i] < 0:
# NEGATIVO
self.userTweets("negative",i)
elif self.sentiment_about[i] == 0:
# NEUTRO
pass
else:
# POSITIVO
self.userTweets("positive",i)
for i in range(len(self.enterprises)): # So that it never is set to 0 if there are not changes (logs)
self.attrs['sentiment_enterprise_%s'% self.enterprises[i]] = self.sentiment_about[i]
def userTweets(self, sentiment,enterprise):
aware_neighbors = self.get_neighboring_agents(state_id=self.number_of_enterprises) # Nodes neighbours users
for x in aware_neighbors:
if sentiment == "positive":
x.sentiment_about[enterprise] +=0.003
elif sentiment == "negative":
x.sentiment_about[enterprise] -=0.003
else:
pass
# Establecemos limites
if x.sentiment_about[enterprise] > 1:
x.sentiment_about[enterprise] = 1
if x.sentiment_about[enterprise] < -1:
x.sentiment_about[enterprise] = -1
x.attrs['sentiment_enterprise_%s'% self.enterprises[enterprise]] = x.sentiment_about[enterprise]

42
soil/agents/CounterModel.py
View File

@ -1,19 +1,29 @@
from . import NetworkAgent
from . import BaseAgent, NetworkAgent
class Ticker(BaseAgent):
times = 0
def step(self):
self.times += 1
class CounterModel(NetworkAgent):
"""
Dummy behaviour. It counts the number of nodes in the network and neighbors
in each step and adds it to its state.
"""
times = 0
neighbors = 0
total = 0
def step(self):
# Outside effects
total = len(list(self.get_agents()))
neighbors = len(list(self.get_neighboring_agents()))
self['times'] = self.get('times', 0) + 1
self['neighbors'] = neighbors
self['total'] = total
total = len(list(self.model.schedule._agents))
neighbors = len(list(self.get_neighbors()))
self["times"] = self.get("times", 0) + 1
self["neighbors"] = neighbors
self["total"] = total
class AggregatedCounter(NetworkAgent):
@ -22,17 +32,15 @@ class AggregatedCounter(NetworkAgent):
in each step and adds it to its state.
"""
defaults = {
'times': 0,
'neighbors': 0,
'total': 0
}
times = 0
neighbors = 0
total = 0
def step(self):
# Outside effects
self['times'] += 1
neighbors = len(list(self.get_neighboring_agents()))
self['neighbors'] += neighbors
total = len(list(self.get_agents()))
self['total'] += total
self.debug('Running for step: {}. Total: {}'.format(self.now, total))
self["times"] += 1
neighbors = len(list(self.get_neighbors()))
self["neighbors"] += neighbors
total = len(list(self.model.schedule.agents))
self["total"] += total
self.debug("Running for step: {}. Total: {}".format(self.now, total))

14
soil/agents/Geo.py
View File

@ -1,21 +1,21 @@
from scipy.spatial import cKDTree as KDTree
import networkx as nx
from . import NetworkAgent, as_node
from . import NetworkAgent
class Geo(NetworkAgent):
'''In this type of network, nodes have a "pos" attribute.'''
"""In this type of network, nodes have a "pos" attribute."""
def geo_search(self, radius, node=None, center=False, **kwargs):
'''Get a list of nodes whose coordinates are closer than *radius* to *node*.'''
node = as_node(node if node is not None else self)
def geo_search(self, radius, center=False, **kwargs):
"""Get a list of nodes whose coordinates are closer than *radius* to *node*."""
node = self.node_id
G = self.subgraph(**kwargs)
pos = nx.get_node_attributes(G, 'pos')
pos = nx.get_node_attributes(G, "pos")
if not pos:
return []
nodes, coords = list(zip(*pos.items()))
kdtree = KDTree(coords) # Cannot provide generator.
indices = kdtree.query_ball_point(pos[node], radius)
return [nodes[i] for i in indices if center or (nodes[i] != node)]

53
soil/agents/IndependentCascadeModel.py
View File

@ -1,8 +1,7 @@
import random
from . import BaseAgent
from . import Agent, state, default_state
class IndependentCascadeModel(BaseAgent):
class IndependentCascadeModel(Agent):
"""
Settings:
innovation_prob
@ -10,40 +9,22 @@ class IndependentCascadeModel(BaseAgent):
imitation_prob
"""
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.innovation_prob = self.env.environment_params['innovation_prob']
self.imitation_prob = self.env.environment_params['imitation_prob']
self.state['time_awareness'] = 0
self.state['sentimentCorrelation'] = 0
time_awareness = 0
sentimentCorrelation = 0
def step(self):
self.behaviour()
def behaviour(self):
aware_neighbors_1_time_step = []
# Outside effects
if random.random() < self.innovation_prob:
if self.state['id'] == 0:
self.state['id'] = 1
self.state['sentimentCorrelation'] = 1
self.state['time_awareness'] = self.env.now # To know when they have been infected
else:
pass
@default_state
@state
def outside(self):
if self.prob(self.model.innovation_prob):
self.sentimentCorrelation = 1
self.time_awareness = self.model.now # To know when they have been infected
return self.imitate
return
@state
def imitate(self):
aware_neighbors = self.get_neighbors(state_id=1, time_awareness=self.now-1)
# Imitation effects
if self.state['id'] == 0:
aware_neighbors = self.get_neighboring_agents(state_id=1)
for x in aware_neighbors:
if x.state['time_awareness'] == (self.env.now-1):
aware_neighbors_1_time_step.append(x)
num_neighbors_aware = len(aware_neighbors_1_time_step)
if random.random() < (self.imitation_prob*num_neighbors_aware):
self.state['id'] = 1
self.state['sentimentCorrelation'] = 1
else:
pass
return
if self.prob(self.model.imitation_prob * len(aware_neighbors)):
self.sentimentCorrelation = 1
return self.outside

242
soil/agents/ModelM2.py
View File

@ -1,242 +0,0 @@
import random
import numpy as np
from . import BaseAgent
class SpreadModelM2(BaseAgent):
"""
Settings:
prob_neutral_making_denier
prob_infect
prob_cured_healing_infected
prob_cured_vaccinate_neutral
prob_vaccinated_healing_infected
prob_vaccinated_vaccinate_neutral
prob_generate_anti_rumor
"""
def __init__(self, model=None, unique_id=0, state=()):
super().__init__(model=environment, unique_id=unique_id, state=state)
self.prob_neutral_making_denier = np.random.normal(environment.environment_params['prob_neutral_making_denier'],
environment.environment_params['standard_variance'])
self.prob_infect = np.random.normal(environment.environment_params['prob_infect'],
environment.environment_params['standard_variance'])
self.prob_cured_healing_infected = np.random.normal(environment.environment_params['prob_cured_healing_infected'],
environment.environment_params['standard_variance'])
self.prob_cured_vaccinate_neutral = np.random.normal(environment.environment_params['prob_cured_vaccinate_neutral'],
environment.environment_params['standard_variance'])
self.prob_vaccinated_healing_infected = np.random.normal(environment.environment_params['prob_vaccinated_healing_infected'],
environment.environment_params['standard_variance'])
self.prob_vaccinated_vaccinate_neutral = np.random.normal(environment.environment_params['prob_vaccinated_vaccinate_neutral'],
environment.environment_params['standard_variance'])
self.prob_generate_anti_rumor = np.random.normal(environment.environment_params['prob_generate_anti_rumor'],
environment.environment_params['standard_variance'])
def step(self):
if self.state['id'] == 0: # Neutral
self.neutral_behaviour()
elif self.state['id'] == 1: # Infected
self.infected_behaviour()
elif self.state['id'] == 2: # Cured
self.cured_behaviour()
elif self.state['id'] == 3: # Vaccinated
self.vaccinated_behaviour()
def neutral_behaviour(self):
# Infected
infected_neighbors = self.get_neighboring_agents(state_id=1)
if len(infected_neighbors) > 0:
if random.random() < self.prob_neutral_making_denier:
self.state['id'] = 3 # Vaccinated making denier
def infected_behaviour(self):
# Neutral
neutral_neighbors = self.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors:
if random.random() < self.prob_infect:
neighbor.state['id'] = 1 # Infected
def cured_behaviour(self):
# Vaccinate
neutral_neighbors = self.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors:
if random.random() < self.prob_cured_vaccinate_neutral:
neighbor.state['id'] = 3 # Vaccinated
# Cure
infected_neighbors = self.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors:
if random.random() < self.prob_cured_healing_infected:
neighbor.state['id'] = 2 # Cured
def vaccinated_behaviour(self):
# Cure
infected_neighbors = self.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors:
if random.random() < self.prob_cured_healing_infected:
neighbor.state['id'] = 2 # Cured
# Vaccinate
neutral_neighbors = self.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors:
if random.random() < self.prob_cured_vaccinate_neutral:
neighbor.state['id'] = 3 # Vaccinated
# Generate anti-rumor
infected_neighbors_2 = self.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors_2:
if random.random() < self.prob_generate_anti_rumor:
neighbor.state['id'] = 2 # Cured
class ControlModelM2(BaseAgent):
"""
Settings:
prob_neutral_making_denier
prob_infect
prob_cured_healing_infected
prob_cured_vaccinate_neutral
prob_vaccinated_healing_infected
prob_vaccinated_vaccinate_neutral
prob_generate_anti_rumor
"""
def __init__(self, model=None, unique_id=0, state=()):
super().__init__(model=environment, unique_id=unique_id, state=state)
self.prob_neutral_making_denier = np.random.normal(environment.environment_params['prob_neutral_making_denier'],
environment.environment_params['standard_variance'])
self.prob_infect = np.random.normal(environment.environment_params['prob_infect'],
environment.environment_params['standard_variance'])
self.prob_cured_healing_infected = np.random.normal(environment.environment_params['prob_cured_healing_infected'],
environment.environment_params['standard_variance'])
self.prob_cured_vaccinate_neutral = np.random.normal(environment.environment_params['prob_cured_vaccinate_neutral'],
environment.environment_params['standard_variance'])
self.prob_vaccinated_healing_infected = np.random.normal(environment.environment_params['prob_vaccinated_healing_infected'],
environment.environment_params['standard_variance'])
self.prob_vaccinated_vaccinate_neutral = np.random.normal(environment.environment_params['prob_vaccinated_vaccinate_neutral'],
environment.environment_params['standard_variance'])
self.prob_generate_anti_rumor = np.random.normal(environment.environment_params['prob_generate_anti_rumor'],
environment.environment_params['standard_variance'])
def step(self):
if self.state['id'] == 0: # Neutral
self.neutral_behaviour()
elif self.state['id'] == 1: # Infected
self.infected_behaviour()
elif self.state['id'] == 2: # Cured
self.cured_behaviour()
elif self.state['id'] == 3: # Vaccinated
self.vaccinated_behaviour()
elif self.state['id'] == 4: # Beacon-off
self.beacon_off_behaviour()
elif self.state['id'] == 5: # Beacon-on
self.beacon_on_behaviour()
def neutral_behaviour(self):
self.state['visible'] = False
# Infected
infected_neighbors = self.get_neighboring_agents(state_id=1)
if len(infected_neighbors) > 0:
if random.random() < self.prob_neutral_making_denier:
self.state['id'] = 3 # Vaccinated making denier
def infected_behaviour(self):
# Neutral
neutral_neighbors = self.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors:
if random.random() < self.prob_infect:
neighbor.state['id'] = 1 # Infected
self.state['visible'] = False
def cured_behaviour(self):
self.state['visible'] = True
# Vaccinate
neutral_neighbors = self.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors:
if random.random() < self.prob_cured_vaccinate_neutral:
neighbor.state['id'] = 3 # Vaccinated
# Cure
infected_neighbors = self.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors:
if random.random() < self.prob_cured_healing_infected:
neighbor.state['id'] = 2 # Cured
def vaccinated_behaviour(self):
self.state['visible'] = True
# Cure
infected_neighbors = self.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors:
if random.random() < self.prob_cured_healing_infected:
neighbor.state['id'] = 2 # Cured
# Vaccinate
neutral_neighbors = self.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors:
if random.random() < self.prob_cured_vaccinate_neutral:
neighbor.state['id'] = 3 # Vaccinated
# Generate anti-rumor
infected_neighbors_2 = self.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors_2:
if random.random() < self.prob_generate_anti_rumor:
neighbor.state['id'] = 2 # Cured
def beacon_off_behaviour(self):
self.state['visible'] = False
infected_neighbors = self.get_neighboring_agents(state_id=1)
if len(infected_neighbors) > 0:
self.state['id'] == 5 # Beacon on
def beacon_on_behaviour(self):
self.state['visible'] = False
# Cure (M2 feature added)
infected_neighbors = self.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors:
if random.random() < self.prob_generate_anti_rumor:
neighbor.state['id'] = 2 # Cured
neutral_neighbors_infected = neighbor.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors_infected:
if random.random() < self.prob_generate_anti_rumor:
neighbor.state['id'] = 3 # Vaccinated
infected_neighbors_infected = neighbor.get_neighboring_agents(state_id=1)
for neighbor in infected_neighbors_infected:
if random.random() < self.prob_generate_anti_rumor:
neighbor.state['id'] = 2 # Cured
# Vaccinate
neutral_neighbors = self.get_neighboring_agents(state_id=0)
for neighbor in neutral_neighbors:
if random.random() < self.prob_cured_vaccinate_neutral:
neighbor.state['id'] = 3 # Vaccinated

83
soil/agents/SISaModel.py
View File

@ -1,9 +1,9 @@
import random
import numpy as np
from . import FSM, state
from hashlib import sha512
from . import Agent, state, default_state
class SISaModel(FSM):
class SISaModel(Agent):
"""
Settings:
neutral_discontent_spon_prob
@ -29,65 +29,82 @@ class SISaModel(FSM):
standard_variance
"""
def __init__(self, environment, unique_id=0, state=()):
super().__init__(model=environment, unique_id=unique_id, state=state)
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.neutral_discontent_spon_prob = np.random.normal(self.env['neutral_discontent_spon_prob'],
self.env['standard_variance'])
self.neutral_discontent_infected_prob = np.random.normal(self.env['neutral_discontent_infected_prob'],
self.env['standard_variance'])
self.neutral_content_spon_prob = np.random.normal(self.env['neutral_content_spon_prob'],
self.env['standard_variance'])
self.neutral_content_infected_prob = np.random.normal(self.env['neutral_content_infected_prob'],
self.env['standard_variance'])
seed = self.model._seed
if isinstance(seed, (str, bytes, bytearray)):
if isinstance(seed, str):
seed = seed.encode()
seed = int.from_bytes(seed + sha512(seed).digest(), 'big')
self.discontent_neutral = np.random.normal(self.env['discontent_neutral'],
self.env['standard_variance'])
self.discontent_content = np.random.normal(self.env['discontent_content'],
self.env['variance_d_c'])
random = np.random.default_rng(seed=seed)
self.content_discontent = np.random.normal(self.env['content_discontent'],
self.env['variance_c_d'])
self.content_neutral = np.random.normal(self.env['content_neutral'],
self.env['standard_variance'])
self.neutral_discontent_spon_prob = random.normal(
self.model.neutral_discontent_spon_prob, self.model.standard_variance
)
self.neutral_discontent_infected_prob = random.normal(
self.model.neutral_discontent_infected_prob, self.model.standard_variance
)
self.neutral_content_spon_prob = random.normal(
self.model.neutral_content_spon_prob, self.model.standard_variance
)
self.neutral_content_infected_prob = random.normal(
self.model.neutral_content_infected_prob, self.model.standard_variance
)
self.discontent_neutral = random.normal(
self.model.discontent_neutral, self.model.standard_variance
)
self.discontent_content = random.normal(
self.model.discontent_content, self.model.variance_d_c
)
self.content_discontent = random.normal(
self.model.content_discontent, self.model.variance_c_d
)
self.content_neutral = random.normal(
self.model.discontent_neutral, self.model.standard_variance
)
@default_state
@state
def neutral(self):
# Spontaneous effects
if random.random() < self.neutral_discontent_spon_prob:
if self.prob(self.neutral_discontent_spon_prob):
return self.discontent
if random.random() < self.neutral_content_spon_prob:
if self.prob(self.neutral_content_spon_prob):
return self.content
# Infected
discontent_neighbors = self.count_neighboring_agents(state_id=self.discontent)
if random.random() < discontent_neighbors * self.neutral_discontent_infected_prob:
discontent_neighbors = self.count_neighbors(state_id=self.discontent)
if self.prob(discontent_neighbors * self.neutral_discontent_infected_prob):
return self.discontent
content_neighbors = self.count_neighboring_agents(state_id=self.content.id)
if random.random() < content_neighbors * self.neutral_content_infected_prob:
content_neighbors = self.count_neighbors(state_id=self.content.id)
if self.prob(content_neighbors * self.neutral_content_infected_prob):
return self.content
return self.neutral
@state
def discontent(self):
# Healing
if random.random() < self.discontent_neutral:
if self.prob(self.discontent_neutral):
return self.neutral
# Superinfected
content_neighbors = self.count_neighboring_agents(state_id=self.content.id)
if random.random() < content_neighbors * self.discontent_content:
content_neighbors = self.count_neighbors(state_id=self.content.id)
if self.prob(content_neighbors * self.discontent_content):
return self.content
return self.discontent
@state
def content(self):
# Healing
if random.random() < self.content_neutral:
if self.prob(self.content_neutral):
return self.neutral
# Superinfected
discontent_neighbors = self.count_neighboring_agents(state_id=self.discontent.id)
if random.random() < discontent_neighbors * self.content_discontent:
discontent_neighbors = self.count_neighbors(state_id=self.discontent.id)
if self.prob(discontent_neighbors * self.content_discontent):
self.discontent
return self.content

102
soil/agents/SentimentCorrelationModel.py
View File

@ -1,102 +0,0 @@
import random
from . import BaseAgent
class SentimentCorrelationModel(BaseAgent):
"""
Settings:
outside_effects_prob
anger_prob
joy_prob
sadness_prob
disgust_prob
"""
def __init__(self, environment, unique_id=0, state=()):
super().__init__(model=environment, unique_id=unique_id, state=state)
self.outside_effects_prob = environment.environment_params['outside_effects_prob']
self.anger_prob = environment.environment_params['anger_prob']
self.joy_prob = environment.environment_params['joy_prob']
self.sadness_prob = environment.environment_params['sadness_prob']
self.disgust_prob = environment.environment_params['disgust_prob']
self.state['time_awareness'] = []
for i in range(4): # In this model we have 4 sentiments
self.state['time_awareness'].append(0) # 0-> Anger, 1-> joy, 2->sadness, 3 -> disgust
self.state['sentimentCorrelation'] = 0
def step(self):
self.behaviour()
def behaviour(self):
angry_neighbors_1_time_step = []
joyful_neighbors_1_time_step = []
sad_neighbors_1_time_step = []
disgusted_neighbors_1_time_step = []
angry_neighbors = self.get_neighboring_agents(state_id=1)
for x in angry_neighbors:
if x.state['time_awareness'][0] > (self.env.now-500):
angry_neighbors_1_time_step.append(x)
num_neighbors_angry = len(angry_neighbors_1_time_step)
joyful_neighbors = self.get_neighboring_agents(state_id=2)
for x in joyful_neighbors:
if x.state['time_awareness'][1] > (self.env.now-500):
joyful_neighbors_1_time_step.append(x)
num_neighbors_joyful = len(joyful_neighbors_1_time_step)
sad_neighbors = self.get_neighboring_agents(state_id=3)
for x in sad_neighbors:
if x.state['time_awareness'][2] > (self.env.now-500):
sad_neighbors_1_time_step.append(x)
num_neighbors_sad = len(sad_neighbors_1_time_step)
disgusted_neighbors = self.get_neighboring_agents(state_id=4)
for x in disgusted_neighbors:
if x.state['time_awareness'][3] > (self.env.now-500):
disgusted_neighbors_1_time_step.append(x)
num_neighbors_disgusted = len(disgusted_neighbors_1_time_step)
anger_prob = self.anger_prob+(len(angry_neighbors_1_time_step)*self.anger_prob)
joy_prob = self.joy_prob+(len(joyful_neighbors_1_time_step)*self.joy_prob)
sadness_prob = self.sadness_prob+(len(sad_neighbors_1_time_step)*self.sadness_prob)
disgust_prob = self.disgust_prob+(len(disgusted_neighbors_1_time_step)*self.disgust_prob)
outside_effects_prob = self.outside_effects_prob
num = random.random()
if num<outside_effects_prob:
self.state['id'] = random.randint(1, 4)
self.state['sentimentCorrelation'] = self.state['id'] # It is stored when it has been infected for the dynamic network
self.state['time_awareness'][self.state['id']-1] = self.env.now
self.state['sentiment'] = self.state['id']
if(num<anger_prob):
self.state['id'] = 1
self.state['sentimentCorrelation'] = 1
self.state['time_awareness'][self.state['id']-1] = self.env.now
elif (num<joy_prob+anger_prob and num>anger_prob):
self.state['id'] = 2
self.state['sentimentCorrelation'] = 2
self.state['time_awareness'][self.state['id']-1] = self.env.now
elif (num<sadness_prob+anger_prob+joy_prob and num>joy_prob+anger_prob):
self.state['id'] = 3
self.state['sentimentCorrelation'] = 3
self.state['time_awareness'][self.state['id']-1] = self.env.now
elif (num<disgust_prob+sadness_prob+anger_prob+joy_prob and num>sadness_prob+anger_prob+joy_prob):
self.state['id'] = 4
self.state['sentimentCorrelation'] = 4
self.state['time_awareness'][self.state['id']-1] = self.env.now
self.state['sentiment'] = self.state['id']

921
soil/agents/__init__.py
View File

File diff suppressed because it is too large Load Diff

77
soil/agents/evented.py Normal file
View File

@ -0,0 +1,77 @@
from . import BaseAgent
from ..events import Message, Tell, Ask, TimedOut
from ..time import BaseCond
from functools import partial
from collections import deque
class ReceivedOrTimeout(BaseCond):
def __init__(
self, agent, expiration=None, timeout=None, check=True, ignore=False, **kwargs
):
if expiration is None:
if timeout is not None:
expiration = agent.now + timeout
self.expiration = expiration
self.ignore = ignore
self.check = check
super().__init__(**kwargs)
def expired(self, time):
return self.expiration and self.expiration < time
def ready(self, agent, time):
return len(agent._inbox) or self.expired(time)
def return_value(self, agent):
if not self.ignore and self.expired(agent.now):
raise TimedOut("No messages received")
if self.check:
agent.check_messages()
return None
def schedule_next(self, time, delta, first=False):
if self._delta is not None:
delta = self._delta
return (time + delta, self)
def __repr__(self):
return f"ReceivedOrTimeout(expires={self.expiration})"
class EventedAgent(BaseAgent):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self._inbox = deque()
self._processed = 0
def on_receive(self, *args, **kwargs):
pass
def received(self, *args, **kwargs):
return ReceivedOrTimeout(self, *args, **kwargs)
def tell(self, msg, sender=None):
self._inbox.append(Tell(timestamp=self.now, payload=msg, sender=sender))
def ask(self, msg, timeout=None, **kwargs):
ask = Ask(timestamp=self.now, payload=msg, sender=self)
self._inbox.append(ask)
expiration = float("inf") if timeout is None else self.now + timeout
return ask.replied(expiration=expiration, **kwargs)
def check_messages(self):
changed = False
while self._inbox:
msg = self._inbox.popleft()
self._processed += 1
if msg.expired(self.now):
continue
changed = True
reply = self.on_receive(msg.payload, sender=msg.sender)
if isinstance(msg, Ask):
msg.reply = reply
return changed
Evented = EventedAgent

148
soil/agents/fsm.py Normal file
View File

@ -0,0 +1,148 @@
from . import MetaAgent, BaseAgent
from ..time import Delta
from functools import partial, wraps
import inspect
def state(name=None, default=False):
def decorator(func, name=None):
"""
A state function should return either a state id, or a tuple (state_id, when)
The default value for state_id is the current state id.
The default value for when is the interval defined in the environment.
"""
if inspect.isgeneratorfunction(func):
orig_func = func
@wraps(func)
def func(self):
while True:
if not self._coroutine:
self._coroutine = orig_func(self)
try:
if self._last_except:
n = self._coroutine.throw(self._last_except)
else:
n = self._coroutine.send(self._last_return)
if n:
return None, n
return n
except StopIteration as ex:
self._coroutine = None
next_state = ex.value
if next_state is not None:
self._set_state(next_state)
return next_state
finally:
self._last_return = None
self._last_except = None
func.id = name or func.__name__
func.is_default = default
return func
if callable(name):
return decorator(name)
else:
return partial(decorator, name=name)
def default_state(func):
func.is_default = True
return func
class MetaFSM(MetaAgent):
def __new__(mcls, name, bases, namespace):
states = {}
# Re-use states from inherited classes
default_state = None
for i in bases:
if isinstance(i, MetaFSM):
for state_id, state in i._states.items():
if state.is_default:
default_state = state
states[state_id] = state
# Add new states
for attr, func in namespace.items():
if hasattr(func, "id"):
if func.is_default:
default_state = func
states[func.id] = func
namespace.update(
{
"_default_state": default_state,
"_states": states,
}
)
return super(MetaFSM, mcls).__new__(
mcls=mcls, name=name, bases=bases, namespace=namespace
)
class FSM(BaseAgent, metaclass=MetaFSM):
def __init__(self, init=True, **kwargs):
super().__init__(**kwargs, init=False)
if not hasattr(self, "state_id"):
if not self._default_state:
raise ValueError(
"No default state specified for {}".format(self.unique_id)
)
self.state_id = self._default_state.id
self._coroutine = None
self.default_interval = Delta(self.model.interval)
self._set_state(self.state_id)
if init:
self.init()
@classmethod
def states(cls):
return list(cls._states.keys())
def step(self):
self.debug(f"Agent {self.unique_id} @ state {self.state_id}")
self._check_alive()
next_state = self._states[self.state_id](self)
when = None
try:
next_state, *when = next_state
if not when:
when = None
elif len(when) == 1:
when = when[0]
else:
raise ValueError(
"Too many values returned. Only state (and time) allowed"
)
except TypeError:
pass
if next_state is not None:
self._set_state(next_state)
return when or self.default_interval
def _set_state(self, state, when=None):
if hasattr(state, "id"):
state = state.id
if state not in self._states:
raise ValueError("{} is not a valid state".format(state))
self.state_id = state
if when is not None:
self.model.schedule.add(self, when=when)
return state
def die(self, *args, **kwargs):
return self.dead, super().die(*args, **kwargs)
@state
def dead(self):
return self.die()

100
soil/agents/network_agents.py Normal file
View File

@ -0,0 +1,100 @@
from . import BaseAgent
class NetworkAgent(BaseAgent):
def __init__(self, *args, topology=None, init=True, node_id=None, **kwargs):
super().__init__(*args, init=False, **kwargs)
self.G = topology or self.model.G
assert self.G
if node_id is None:
nodes = self.random.choices(list(self.G.nodes), k=len(self.G))
for n_id in nodes:
if "agent" not in self.G.nodes[n_id] or self.G.nodes[n_id]["agent"] is None:
node_id = n_id
break
else:
node_id = len(self.G)
self.info(f"All nodes ({len(self.G)}) have an agent assigned, adding a new node to the graph for agent {self.unique_id}")
self.G.add_node(node_id)
assert node_id is not None
self.G.nodes[node_id]["agent"] = self
self.node_id = node_id
if init:
self.init()
def count_neighbors(self, state_id=None, **kwargs):
return len(self.get_neighbors(state_id=state_id, **kwargs))
if init:
self.init()
def iter_neighbors(self, **kwargs):
return self.iter_agents(limit_neighbors=True, **kwargs)
def get_neighbors(self, **kwargs):
return list(self.iter_neighbors(**kwargs))
@property
def node(self):
return self.G.nodes[self.node_id]
def iter_agents(self, unique_id=None, *, limit_neighbors=False, **kwargs):
unique_ids = None
if unique_ids is not None:
try:
unique_ids = set(unique_id)
except TypeError:
unique_ids = set([unique_id])
if limit_neighbors:
neighbor_ids = set()
for node_id in self.G.neighbors(self.node_id):
agent = self.G.nodes[node_id].get("agent")
if agent is not None:
neighbor_ids.add(agent.unique_id)
if unique_ids:
unique_ids = unique_ids & neighbor_ids
else:
unique_ids = neighbor_ids
if not unique_ids:
return
unique_ids = list(unique_ids)
yield from super().iter_agents(unique_id=unique_ids, **kwargs)
def subgraph(self, center=True, **kwargs):
include = [self] if center else []
G = self.G.subgraph(
n.node_id for n in list(self.get_agents(**kwargs) + include)
)
return G
def remove_node(self):
self.debug(f"Removing node for {self.unique_id}: {self.node_id}")
self.G.remove_node(self.node_id)
self.node_id = None
def add_edge(self, other, edge_attr_dict=None, *edge_attrs):
if self.node_id not in self.G.nodes(data=False):
raise ValueError(
"{} not in list of existing agents in the network".format(
self.unique_id
)
)
if other.node_id not in self.G.nodes(data=False):
raise ValueError(
"{} not in list of existing agents in the network".format(other)
)
self.G.add_edge(
self.node_id, other.node_id, edge_attr_dict=edge_attr_dict, *edge_attrs
)
def die(self, remove=True):
if not self.alive:
return None
if remove:
self.remove_node()
return super().die()
NetAgent = NetworkAgent

243
soil/analysis.py
View File

@ -1,206 +1,49 @@
import os
import sqlalchemy
import pandas as pd
from collections import namedtuple
import glob
import yaml
from os.path import join
from . import serialization
from tsih import History
def read_data(*args, group=False, **kwargs):
iterable = _read_data(*args, **kwargs)
if group:
return group_trials(iterable)
def plot(env, agent_df=None, model_df=None, steps=False, ignore=["agent_count", ]):
"""Plot the model dataframe and agent dataframe together."""
if agent_df is None:
agent_df = env.agent_df()
if model_df is None:
model_df = env.model_df()
ignore = list(ignore)
if not steps:
ignore.append("step")
else:
return list(iterable)
ignore.append("time")
ax = model_df.drop(ignore, axis='columns').plot();
if not agent_df.empty:
agent_df.unstack().apply(lambda x: x.value_counts(),
axis=1).fillna(0).plot(ax=ax, secondary_y=True);
def _read_data(pattern, *args, from_csv=False, process_args=None, **kwargs):
if not process_args:
process_args = {}
for folder in glob.glob(pattern):
config_file = glob.glob(join(folder, '*.yml'))[0]
config = yaml.load(open(config_file), Loader=yaml.SafeLoader)
df = None
if from_csv:
for trial_data in sorted(glob.glob(join(folder,
'*.environment.csv'))):
df = read_csv(trial_data, **kwargs)
yield config_file, df, config
else:
for trial_data in sorted(glob.glob(join(folder, '*.sqlite'))):
df = read_sql(trial_data, **kwargs)
yield config_file, df, config
Results = namedtuple("Results", ["config", "parameters", "env", "agents"])
#TODO implement reading from CSV and SQLITE
def read_sql(fpath=None, name=None, include_agents=False):
if not (fpath is None) ^ (name is None):
raise ValueError("Specify either a path or a simulation name")
if name:
fpath = os.path.join("soil_output", name, f"{name}.sqlite")
fpath = os.path.abspath(fpath)
# TODO: improve url parsing. This is a hacky way to check we weren't given a URL
if "://" not in fpath:
fpath = f"sqlite:///{fpath}"
engine = sqlalchemy.create_engine(fpath)
with engine.connect() as conn:
env = pd.read_sql_table("env", con=conn,
index_col="step").reset_index().set_index([
"simulation_id", "params_id",
"iteration_id", "step"
])
agents = pd.read_sql_table("agents", con=conn, index_col=["simulation_id", "params_id", "iteration_id", "step", "agent_id"])
config = pd.read_sql_table("configuration", con=conn, index_col="simulation_id")
parameters = pd.read_sql_table("parameters", con=conn, index_col=["iteration_id", "params_id", "simulation_id"])
try:
parameters = parameters.pivot(columns="key", values="value")
except Exception as e:
print(f"warning: coult not pivot parameters: {e}")
def read_sql(db, *args, **kwargs):
h = History(db_path=db, backup=False, readonly=True)
df = h.read_sql(*args, **kwargs)
return df
def read_csv(filename, keys=None, convert_types=False, **kwargs):
'''
Read a CSV in canonical form: ::
<agent_id, t_step, key, value, value_type>
'''
df = pd.read_csv(filename)
if convert_types:
df = convert_types_slow(df)
if keys:
df = df[df['key'].isin(keys)]
df = process_one(df)
return df
def convert_row(row):
row['value'] = serialization.deserialize(row['value_type'], row['value'])
return row
def convert_types_slow(df):
'''
Go over every column in a dataframe and convert it to the type determined by the `get_types`
function.
This is a slow operation.
'''
dtypes = get_types(df)
for k, v in dtypes.items():
t = df[df['key']==k]
t['value'] = t['value'].astype(v)
df = df.apply(convert_row, axis=1)
return df
def split_processed(df):
env = df.loc[:, df.columns.get_level_values(1).isin(['env', 'stats'])]
agents = df.loc[:, ~df.columns.get_level_values(1).isin(['env', 'stats'])]
return env, agents
def split_df(df):
'''
Split a dataframe in two dataframes: one with the history of agents,
and one with the environment history
'''
envmask = (df['agent_id'] == 'env')
n_env = envmask.sum()
if n_env == len(df):
return df, None
elif n_env == 0:
return None, df
agents, env = [x for _, x in df.groupby(envmask)]
return env, agents
def process(df, **kwargs):
'''
Process a dataframe in canonical form ``(t_step, agent_id, key, value, value_type)`` into
two dataframes with a column per key: one with the history of the agents, and one for the
history of the environment.
'''
env, agents = split_df(df)
return process_one(env, **kwargs), process_one(agents, **kwargs)
def get_types(df):
'''
Get the value type for every key stored in a raw history dataframe.
'''
dtypes = df.groupby(by=['key'])['value_type'].unique()
return {k:v[0] for k,v in dtypes.items()}
def process_one(df, *keys, columns=['key', 'agent_id'], values='value',
fill=True, index=['t_step',],
aggfunc='first', **kwargs):
'''
Process a dataframe in canonical form ``(t_step, agent_id, key, value, value_type)`` into
a dataframe with a column per key
'''
if df is None:
return df
if keys:
df = df[df['key'].isin(keys)]
df = df.pivot_table(values=values, index=index, columns=columns,
aggfunc=aggfunc, **kwargs)
if fill:
df = fillna(df)
return df
def get_count(df, *keys):
'''
For every t_step and key, get the value count.
The result is a dataframe with `t_step` as index, an a multiindex column based on `key` and the values found for each `key`.
'''
if keys:
df = df[list(keys)]
df.columns = df.columns.remove_unused_levels()
counts = pd.DataFrame()
for key in df.columns.levels[0]:
g = df[[key]].apply(pd.Series.value_counts, axis=1).fillna(0)
for value, series in g.items():
counts[key, value] = series
counts.columns = pd.MultiIndex.from_tuples(counts.columns)
return counts
def get_majority(df, *keys):
'''
For every t_step and key, get the value of the majority of agents
The result is a dataframe with `t_step` as index, and columns based on `key`.
'''
df = get_count(df, *keys)
return df.stack(level=0).idxmax(axis=1).unstack()
def get_value(df, *keys, aggfunc='sum'):
'''
For every t_step and key, get the value of *numeric columns*, aggregated using a specific function.
'''
if keys:
df = df[list(keys)]
df.columns = df.columns.remove_unused_levels()
df = df.select_dtypes('number')
return df.groupby(level='key', axis=1).agg(aggfunc)
def plot_all(*args, plot_args={}, **kwargs):
'''
Read all the trial data and plot the result of applying a function on them.
'''
dfs = do_all(*args, **kwargs)
ps = []
for line in dfs:
f, df, config = line
if len(df) < 1:
continue
df.plot(title=config['name'], **plot_args)
ps.append(df)
return ps
def do_all(pattern, func, *keys, include_env=False, **kwargs):
for config_file, df, config in read_data(pattern, keys=keys):
if len(df) < 1:
continue
p = func(df, *keys, **kwargs)
yield config_file, p, config
def group_trials(trials, aggfunc=['mean', 'min', 'max', 'std']):
trials = list(trials)
trials = list(map(lambda x: x[1] if isinstance(x, tuple) else x, trials))
return pd.concat(trials).groupby(level=0).agg(aggfunc).reorder_levels([2, 0,1] ,axis=1)
def fillna(df):
new_df = df.ffill(axis=0)
return new_df
return Results(config, parameters, env, agents)

2
soil/config.py Normal file
View File

@ -0,0 +1,2 @@
def load_config(cfg):
return cfg

37
soil/datacollection.py
View File

@ -1,26 +1,19 @@
from mesa import DataCollector as MDC
class SoilDataCollector(MDC):
class SoilCollector(MDC):
def __init__(self, model_reporters=None, agent_reporters=None, tables=None, **kwargs):
model_reporters = model_reporters or {}
agent_reporters = agent_reporters or {}
tables = tables or {}
if 'agent_count' not in model_reporters:
model_reporters['agent_count'] = lambda m: m.schedule.get_agent_count()
if 'time' not in model_reporters:
model_reporters['time'] = lambda m: m.now
# if 'state_id' not in agent_reporters:
# agent_reporters['state_id'] = lambda agent: getattr(agent, 'state_id', None)
def __init__(self, environment, *args, **kwargs):
super().__init__(*args, **kwargs)
# Populate model and env reporters so they have a key per
# So they can be shown in the web interface
self.environment = environment
@property
def model_vars(self):
pass
@model_vars.setter
def model_vars(self, value):
pass
@property
def agent_reporters(self):
self.model._history._
pass
super().__init__(model_reporters=model_reporters,
agent_reporters=agent_reporters,
tables=tables,
**kwargs)

243
soil/debugging.py Normal file
View File

@ -0,0 +1,243 @@
from __future__ import annotations
import pdb
import sys
import os
from textwrap import indent
from functools import wraps
from .agents import FSM, MetaFSM
from mesa import Model, Agent
def wrapcmd(func):
@wraps(func)
def wrapper(self, arg: str, temporary=False):
sys.settrace(self.trace_dispatch)
lastself = self
known = globals()
known.update(self.curframe.f_globals)
known.update(self.curframe.f_locals)
known["attrs"] = arg.strip().split()
this = known.get("self", None)
if isinstance(this, Model):
known["model"] = this
elif isinstance(this, Agent):
known["agent"] = this
known["model"] = this.model
known["self"] = lastself
return exec(func.__code__, known, known)
return wrapper
class Debug(pdb.Pdb):
def __init__(self, *args, skip_soil=False, **kwargs):
skip = kwargs.get("skip", [])
if skip_soil:
skip.append("soil")
skip.append("contextlib")
skip.append("soil.*")
skip.append("mesa.*")
super(Debug, self).__init__(*args, skip=skip, **kwargs)
self.prompt = "[soil-pdb] "
@staticmethod
def _soil_agents(model, attrs=None, pretty=True, **kwargs):
for agent in model.agents(**kwargs):
d = agent
print(" - " + indent(agent.to_str(keys=attrs, pretty=pretty), " "))
@wrapcmd
def do_soil_agents():
return Debug._soil_agents(model, attrs=attrs or None)
do_sa = do_soil_agents
@wrapcmd
def do_soil_list():
return Debug._soil_agents(model, attrs=["state_id"], pretty=False)
do_sl = do_soil_list
def do_continue_state(self, arg):
"""Continue until next time this state is reached"""
self.do_break_state(arg, temporary=True)
return self.do_continue("")
do_cs = do_continue_state
@wrapcmd
def do_soil_agent():
if not agent:
print("No agent available")
return
keys = None
if attrs:
keys = []
for k in attrs:
for key in agent.keys():
if key.startswith(k):
keys.append(key)
print(agent.to_str(pretty=True, keys=keys))
do_aa = do_soil_agent
def do_break_step(self, arg: str):
"""
Break before the next step.
"""
try:
known = globals()
known.update(self.curframe.f_globals)
known.update(self.curframe.f_locals)
func = getattr(known["model"], "step")
except AttributeError as ex:
self.error(f"The model does not have a step function: {ex}")
return
if hasattr(func, "__func__"):
func = func.__func__
code = func.__code__
# use co_name to identify the bkpt (function names
# could be aliased, but co_name is invariant)
funcname = code.co_name
lineno = code.co_firstlineno
filename = code.co_filename
# Check for reasonable breakpoint
line = self.checkline(filename, lineno)
if not line:
raise ValueError("no line found")
# now set the break point
existing = self.get_breaks(filename, line)
if existing:
self.message("Breakpoint already exists at %s:%d" % (filename, line))
return
cond = f"self.schedule.steps > {model.schedule.steps}"
err = self.set_break(filename, line, True, cond, funcname)
if err:
self.error(err)
else:
bp = self.get_breaks(filename, line)[-1]
self.message("Breakpoint %d at %s:%d" % (bp.number, bp.file, bp.line))
return self.do_continue("")
do_bstep = do_break_step
def do_break_state(self, arg: str, instances=None, temporary=False):
"""
Break before a specified state is stepped into.
"""
klass = None
state = arg
if not state:
self.error("Specify at least a state name")
return
state, *tokens = state.lstrip().split()
if tokens:
instances = list(eval(token) for token in tokens)
colon = state.find(":")
if colon > 0:
klass = state[:colon].rstrip()
state = state[colon + 1 :].strip()
print(klass, state, tokens)
klass = eval(klass, self.curframe.f_globals, self.curframe_locals)
if klass:
klasses = [klass]
else:
klasses = [
k
for k in self.curframe.f_globals.values()
if isinstance(k, type) and issubclass(k, FSM)
]
if not klasses:
self.error("No agent classes found")
for klass in klasses:
try:
func = getattr(klass, state)
except AttributeError:
self.error(f"State {state} not found in class {klass}")
continue
if hasattr(func, "__func__"):
func = func.__func__
code = func.__code__
# use co_name to identify the bkpt (function names
# could be aliased, but co_name is invariant)
funcname = code.co_name
lineno = code.co_firstlineno
filename = code.co_filename
# Check for reasonable breakpoint
line = self.checkline(filename, lineno)
if not line:
raise ValueError("no line found")
# now set the break point
cond = None
if instances:
cond = f"self.unique_id in { repr(instances) }"
existing = self.get_breaks(filename, line)
if existing:
self.message("Breakpoint already exists at %s:%d" % (filename, line))
continue
err = self.set_break(filename, line, temporary, cond, funcname)
if err:
self.error(err)
else:
bp = self.get_breaks(filename, line)[-1]
self.message("Breakpoint %d at %s:%d" % (bp.number, bp.file, bp.line))
do_bs = do_break_state
def do_break_state_self(self, arg: str, temporary=False):
"""
Break before a specified state is stepped into, for the current agent
"""
agent = self.curframe.f_locals.get("self")
if not agent:
self.error("No current agent.")
self.error("Try this again when the debugger is stopped inside an agent")
return
arg = f"{agent.__class__.__name__}:{ arg } {agent.unique_id}"
return self.do_break_state(arg)
do_bss = do_break_state_self
debugger = None
def set_trace(frame=None, **kwargs):
global debugger
if debugger is None:
debugger = Debug(**kwargs)
frame = frame or sys._getframe().f_back
debugger.set_trace(frame)
def post_mortem(traceback=None, **kwargs):
global debugger
if debugger is None:
debugger = Debug(**kwargs)
t = sys.exc_info()[2]
debugger.reset()
debugger.interaction(None, t)

6
soil/decorators.py Normal file
View File

@ -0,0 +1,6 @@
def report(f: property):
if isinstance(f, property):
setattr(f.fget, "add_to_report", True)
else:
setattr(f, "add_to_report", True)
return f

660
soil/environment.py
View File

@ -1,208 +1,179 @@
from __future__ import annotations
import os
import sqlite3
import csv
import math
import random
import yaml
import tempfile
import logging
import pandas as pd
import inspect
from typing import Any, Callable, Dict, Optional, Union, List, Type
from collections import namedtuple
from time import time as current_time
from copy import deepcopy
from networkx.readwrite import json_graph
import networkx as nx
from tsih import History, Record, Key, NoHistory
from mesa import Model, Agent
from mesa import Model
from . import agents as agentmod, datacollection, serialization, utils, time, network, events
from . import serialization, agents, analysis, utils, time
# These properties will be copied when pickling/unpickling the environment
_CONFIG_PROPS = [ 'name',
'states',
'default_state',
'interval',
]
# TODO: maybe add metaclass to read attributes of a model
class Environment(Model):
class BaseEnvironment(Model):
"""
The environment is key in a simulation. It contains the network topology,
a reference to network and environment agents, as well as the environment
params, which are used as shared state between agents.
The environment is key in a simulation. It controls how agents interact,
and what information is available to them.
This is an opinionated version of `mesa.Model` class, which adds many
convenience methods and abstractions.
The environment parameters and the state of every agent can be accessed
both by using the environment as a dictionary or with the environment's
both by using the environment as a dictionary and with the environment's
:meth:`soil.environment.Environment.get` method.
"""
def __init__(self, name=None,
network_agents=None,
environment_agents=None,
states=None,
default_state=None,
interval=1,
network_params=None,
seed=None,
topology=None,
schedule=None,
initial_time=0,
environment_params=None,
history=True,
dir_path=None,
**kwargs):
collector_class = datacollection.SoilCollector
def __new__(cls,
*args: Any,
seed="default",
dir_path=None,
collector_class: type = None,
agent_reporters: Optional[Any] = None,
model_reporters: Optional[Any] = None,
tables: Optional[Any] = None,
**kwargs: Any) -> Any:
"""Create a new model with a default seed value"""
self = super().__new__(cls, *args, seed=seed, **kwargs)
self.dir_path = dir_path or os.getcwd()
collector_class = collector_class or cls.collector_class
collector_class = serialization.deserialize(collector_class)
self.datacollector = collector_class(
model_reporters=model_reporters,
agent_reporters=agent_reporters,
tables=tables,
)
for k in dir(cls):
v = getattr(cls, k)
if isinstance(v, property):
v = v.fget
if getattr(v, "add_to_report", False):
self.add_model_reporter(k, v)
return self
def __init__(
self,
*,
id="unnamed_env",
seed="default",
dir_path=None,
schedule_class=time.TimedActivation,
interval=1,
logger = None,
agents: Optional[Dict] = None,
collector_class: type = datacollection.SoilCollector,
agent_reporters: Optional[Any] = None,
model_reporters: Optional[Any] = None,
tables: Optional[Any] = None,
init: bool = True,
**env_params,
):
super().__init__()
self.schedule = schedule
if schedule is None:
self.schedule = time.TimedActivation()
self.name = name or 'UnnamedEnvironment'
seed = seed or current_time()
random.seed(seed)
if isinstance(states, list):
states = dict(enumerate(states))
self.states = deepcopy(states) if states else {}
self.default_state = deepcopy(default_state) or {}
self.current_id = -1
if topology is None:
network_params = network_params or {}
topology = serialization.load_network(network_params,
dir_path=dir_path)
if not topology:
topology = nx.Graph()
self.G = nx.Graph(topology)
self.id = id
if logger:
self.logger = logger
else:
self.logger = utils.logger.getChild(self.id)
self.environment_params = environment_params or {}
self.environment_params.update(kwargs)
if schedule_class is None:
schedule_class = time.TimedActivation
else:
schedule_class = serialization.deserialize(schedule_class)
self._env_agents = {}
self.interval = interval
if history:
history = History
else:
history = NoHistory
self._history = history(name=self.name,
backup=True)
self['SEED'] = seed
self.schedule = schedule_class(self)
if network_agents:
distro = agents.calculate_distribution(network_agents)
self.network_agents = agents._convert_agent_types(distro)
else:
self.network_agents = []
for (k, v) in env_params.items():
self[k] = v
environment_agents = environment_agents or []
if environment_agents:
distro = agents.calculate_distribution(environment_agents)
environment_agents = agents._convert_agent_types(distro)
self.environment_agents = environment_agents
if agents:
self.add_agents(**agents)
if init:
self.init()
self.datacollector.collect(self)
self.logger = utils.logger.getChild(self.name)
def init(self):
pass
@property
def agents(self):
return agentmod.AgentView(self.schedule._agents)
def agent(self, *args, **kwargs):
return agentmod.AgentView(self.schedule._agents).one(*args, **kwargs)
def count_agents(self, *args, **kwargs):
return sum(1 for i in self.agents(*args, **kwargs))
def agent_df(self, steps=False):
df = self.datacollector.get_agent_vars_dataframe()
if steps:
df.index.rename(["step", "agent_id"], inplace=True)
return df
model_df = self.datacollector.get_model_vars_dataframe()
df.index = df.index.set_levels(model_df.time, level=0).rename(["time", "agent_id"])
return df
def model_df(self, steps=False):
df = self.datacollector.get_model_vars_dataframe()
if steps:
return df
df.index.rename("step", inplace=True)
return df.reset_index().set_index("time")
@property
def now(self):
if self.schedule:
return self.schedule.time
raise Exception('The environment has not been scheduled, so it has no sense of time')
@property
def agents(self):
yield from self.environment_agents
yield from self.network_agents
@property
def environment_agents(self):
for ref in self._env_agents.values():
yield ref
@environment_agents.setter
def environment_agents(self, environment_agents):
self._environment_agents = environment_agents
for (ix, agent) in enumerate(self._environment_agents):
self.init_agent(len(self.G) + ix, agent_definitions=environment_agents, with_node=False)
@property
def network_agents(self):
for i in self.G.nodes():
node = self.G.nodes[i]
if 'agent' in node:
yield node['agent']
@network_agents.setter
def network_agents(self, network_agents):
self._network_agents = network_agents
for ix in self.G.nodes():
self.init_agent(ix, agent_definitions=network_agents)
def init_agent(self, agent_id, agent_definitions, with_node=True):
init = False
state = {}
if with_node:
node = self.G.nodes[agent_id]
state = dict(node)
state.update(self.states.get(agent_id, {}))
agent_type = None
if 'agent_type' in state:
agent_type = state['agent_type']
elif with_node and 'agent_type' in node:
agent_type = node['agent_type']
elif 'agent_type' in self.default_state:
agent_type = self.default_state['agent_type']
if agent_type:
agent_type = agents.deserialize_type(agent_type)
elif agent_definitions:
agent_type, state = agents._agent_from_definition(agent_definitions, unique_id=agent_id)
else:
serialization.logger.debug('Skipping agent {}'.format(agent_id))
return
return self.set_agent(agent_id, agent_type, state, with_node=with_node)
def set_agent(self, agent_id, agent_type, state=None, with_node=True):
defstate = deepcopy(self.default_state) or {}
defstate.update(self.states.get(agent_id, {}))
if with_node:
node = self.G.nodes[agent_id]
defstate.update(node.get('state', {}))
if state:
defstate.update(state)
a = None
if agent_type:
state = defstate
a = agent_type(model=self,
unique_id=agent_id
raise Exception(
"The environment has not been scheduled, so it has no sense of time"
)
def init_agents(self):
pass
for (k, v) in state.items():
setattr(a, k, v)
def add_agent(self, agent_class, unique_id=None, **agent):
if unique_id is None:
unique_id = self.next_id()
agent["unique_id"] = unique_id
agent = dict(**agent)
unique_id = agent.pop("unique_id", None)
if unique_id is None:
unique_id = self.next_id()
a = serialization.deserialize(agent_class)(unique_id=unique_id, model=self, **agent)
if with_node:
node['agent'] = a
self.schedule.add(a)
return a
def add_node(self, agent_type, state=None):
agent_id = int(len(self.G.nodes()))
self.G.add_node(agent_id)
a = self.set_agent(agent_id, agent_type, state)
a['visible'] = True
return a
def add_agents(self, agent_classes: List[type], k, weights: Optional[List[float]] = None, **kwargs):
if isinstance(agent_classes, type):
agent_classes = [agent_classes]
if weights is None:
weights = [1] * len(agent_classes)
def add_edge(self, agent1, agent2, start=None, **attrs):
if hasattr(agent1, 'id'):
agent1 = agent1.id
if hasattr(agent2, 'id'):
agent2 = agent2.id
start = start or self.now
return self.G.add_edge(agent1, agent2, **attrs)
for cls in self.random.choices(agent_classes, weights=weights, k=k):
self.add_agent(agent_class=cls, **kwargs)
def log(self, message, *args, level=logging.INFO, **kwargs):
if not self.logger.isEnabledFor(level):
@ -212,185 +183,248 @@ class Environment(Model):
for k, v in kwargs:
message += " {k}={v} ".format(k, v)
extra = {}
extra['now'] = self.now
extra['unique_id'] = self.name
extra["now"] = self.now
extra["id"] = self.id
return self.logger.log(level, message, extra=extra)
def step(self):
"""
Advance one step in the simulation, and update the data collection and scheduler appropriately
"""
super().step()
self.schedule.step()
self.datacollector.collect(self)
def run(self, until, *args, **kwargs):
self._save_state()
if self.logger.isEnabledFor(logging.DEBUG):
msg = "Model data:\n"
max_width = max(len(k) for k in self.datacollector.model_vars.keys())
for (k, v) in self.datacollector.model_vars.items():
msg += f"\t{k:<{max_width}}: {v[-1]:>6}\n"
self.logger.debug(f"--- Steps: {self.schedule.steps:^5} - Time: {self.now:^5} --- " + msg)
while self.schedule.next_time < until:
self.step()
utils.logger.debug(f'Simulation step {self.schedule.time}/{until}. Next: {self.schedule.next_time}')
self.schedule.time = until
self._history.flush_cache()
def add_model_reporter(self, name, func=None):
if not func:
func = lambda env: getattr(env, name)
self.datacollector._new_model_reporter(name, func)
def _save_state(self, now=None):
serialization.logger.debug('Saving state @{}'.format(self.now))
self._history.save_records(self.state_to_tuples(now=now))
def add_agent_reporter(self, name, agent_type=None):
if agent_type:
reporter = lambda a: getattr(a, name) if isinstance(a, agent_type) else None
else:
reporter = lambda a: getattr(a, name, None)
self.datacollector._new_agent_reporter(name, reporter)
@classmethod
def run(cls, *,
iterations=1,
num_processes=1, **kwargs):
from .simulation import Simulation
return Simulation(name=cls.__name__,
model=cls, iterations=iterations,
num_processes=num_processes, **kwargs).run()
def __getitem__(self, key):
if isinstance(key, tuple):
self._history.flush_cache()
return self._history[key]
try:
return getattr(self, key)
except AttributeError:
raise KeyError(f"key {key} not found in environment")
return self.environment_params[key]
def __setitem__(self, key, value):
if isinstance(key, tuple):
k = Key(*key)
self._history.save_record(*k,
value=value)
return
self.environment_params[key] = value
self._history.save_record(dict_id='env',
t_step=self.now,
key=key,
value=value)
def __delitem__(self, key):
return delattr(self, key)
def __contains__(self, key):
return key in self.environment_params
return hasattr(self, key)
def __setitem__(self, key, value):
setattr(self, key, value)
def __str__(self):
return str(dict(self))
def __len__(self):
return sum(1 for n in self.keys())
def __iter__(self):
return iter(self.agents())
def get(self, key, default=None):
'''
Get the value of an environment attribute in a
given point in the simulation (history).
If key is an attribute name, this method returns
the current value.
To get values at other times, use a
:meth: `soil.history.Key` tuple.
'''
return self[key] if key in self else default
def get_agent(self, agent_id):
return self.G.nodes[agent_id]['agent']
def keys(self):
return (k for k in self.__dict__ if k[0] != "_")
def get_agents(self, nodes=None):
if nodes is None:
return self.agents
return (self.G.nodes[i]['agent'] for i in nodes)
class NetworkEnvironment(BaseEnvironment):
"""
The NetworkEnvironment is an environment that includes one or more networkx.Graph intances
and methods to associate agents to nodes and vice versa.
"""
def dump_csv(self, f):
with utils.open_or_reuse(f, 'w') as f:
cr = csv.writer(f)
cr.writerow(('agent_id', 't_step', 'key', 'value'))
for i in self.history_to_tuples():
cr.writerow(i)
def dump_gexf(self, f):
G = self.history_to_graph()
# Workaround for geometric models
# See soil/soil#4
for node in G.nodes():
if 'pos' in G.nodes[node]:
G.nodes[node]['viz'] = {"position": {"x": G.nodes[node]['pos'][0], "y": G.nodes[node]['pos'][1], "z": 0.0}}
del (G.nodes[node]['pos'])
nx.write_gexf(G, f, version="1.2draft")
def dump(self, *args, formats=None, **kwargs):
if not formats:
return
functions = {
'csv': self.dump_csv,
'gexf': self.dump_gexf
}
for f in formats:
if f in functions:
functions[f](*args, **kwargs)
def __init__(self,
*args,
topology: Optional[Union[nx.Graph, str]] = None,
agent_class: Optional[Type[agentmod.Agent]] = None,
network_generator: Optional[Callable] = None,
network_params: Optional[Dict] = {},
init=True,
**kwargs):
self.topology = topology
self.network_generator = network_generator
self.network_params = network_params
if topology or network_params or network_generator:
self.create_network(topology, generator=network_generator, **network_params)
else:
raise ValueError('Unknown format: {}'.format(f))
self.G = nx.Graph()
super().__init__(*args, **kwargs, init=False)
def dump_sqlite(self, f):
return self._history.dump(f)
self.agent_class = agent_class
if agent_class:
self.agent_class = serialization.deserialize(agent_class)
if self.agent_class:
self.populate_network(self.agent_class)
self._check_agent_nodes()
if init:
self.init()
self.datacollector.collect(self)
def state_to_tuples(self, now=None):
if now is None:
now = self.now
for k, v in self.environment_params.items():
yield Record(dict_id='env',
t_step=now,
key=k,
value=v)
for agent in self.agents:
for k, v in agent.state.items():
yield Record(dict_id=agent.id,
t_step=now,
key=k,
value=v)
def add_agent(self, agent_class, *args, node_id=None, topology=None, **kwargs):
if node_id is None and topology is None:
return super().add_agent(agent_class, *args, **kwargs)
try:
a = super().add_agent(agent_class, *args, node_id=node_id, **kwargs)
except TypeError:
self.logger.warning(f"Agent constructor for {agent_class} does not have a node_id attribute. Might be a bug.")
a = super().add_agent(agent_class, *args, **kwargs)
self.G.nodes[node_id]["agent"] = a
return a
def history_to_tuples(self):
return self._history.to_tuples()
def add_agents(self, *args, k=None, **kwargs):
if not k and not self.G:
raise ValueError("Cannot add agents to an empty network")
super().add_agents(*args, k=k or len(self.G), **kwargs)
def history_to_graph(self):
G = nx.Graph(self.G)
for agent in self.network_agents:
attributes = {'agent': str(agent.__class__)}
lastattributes = {}
spells = []
lastvisible = False
laststep = None
history = self[agent.id, None, None]
if not history:
continue
for t_step, attribute, value in sorted(list(history)):
if attribute == 'visible':
nowvisible = value
if nowvisible and not lastvisible:
laststep = t_step
if not nowvisible and lastvisible:
spells.append((laststep, t_step))
lastvisible = nowvisible
continue
key = 'attr_' + attribute
if key not in attributes:
attributes[key] = list()
if key not in lastattributes:
lastattributes[key] = (value, t_step)
elif lastattributes[key][0] != value:
last_value, laststep = lastattributes[key]
commit_value = (last_value, laststep, t_step)
if key not in attributes:
attributes[key] = list()
attributes[key].append(commit_value)
lastattributes[key] = (value, t_step)
for k, v in lastattributes.items():
attributes[k].append((v[0], v[1], None))
if lastvisible:
spells.append((laststep, None))
if spells:
G.add_node(agent.id, spells=spells, **attributes)
def create_network(self, topology=None, generator=None, path=None, **network_params):
if topology is not None:
topology = network.from_topology(topology, dir_path=self.dir_path)
elif path is not None:
topology = network.from_topology(path, dir_path=self.dir_path)
elif generator is not None:
topology = network.from_params(generator=generator, dir_path=self.dir_path, **network_params)
else:
G.add_node(agent.id, **attributes)
raise ValueError("topology must be a networkx.Graph or a string, or network_generator must be provided")
self.G = topology
return G
def init_agents(self, *args, **kwargs):
"""Initialize the agents from a"""
super().init_agents(*args, **kwargs)
def __getstate__(self):
state = {}
for prop in _CONFIG_PROPS:
state[prop] = self.__dict__[prop]
state['G'] = json_graph.node_link_data(self.G)
state['environment_agents'] = self._env_agents
state['history'] = self._history
state['schedule'] = self.schedule
return state
@property
def network_agents(self):
"""Return agents still alive and assigned to a node in the network."""
for (id, data) in self.G.nodes(data=True):
if "agent" in data:
agent = data["agent"]
if getattr(agent, "alive", True):
yield agent
def __setstate__(self, state):
for prop in _CONFIG_PROPS:
self.__dict__[prop] = state[prop]
self._env_agents = state['environment_agents']
self.G = json_graph.node_link_graph(state['G'])
self._history = state['history']
# self._env = None
self.schedule = state['schedule']
self._queue = []
def add_node(self, agent_class, unique_id=None, node_id=None, **kwargs):
if unique_id is None:
unique_id = self.next_id()
if node_id is None:
node_id = network.find_unassigned(
G=self.G, shuffle=True, random=self.random
)
if node_id is None:
node_id = f"node_for_{unique_id}"
if node_id not in self.G.nodes:
self.G.add_node(node_id)
assert "agent" not in self.G.nodes[node_id]
a = self.add_agent(
unique_id=unique_id,
agent_class=agent_class,
topology=self.G,
node_id=node_id,
**kwargs,
)
a["visible"] = True
return a
def _check_agent_nodes(self):
"""
Detect nodes that have agents assigned to them.
"""
for (id, data) in self.G.nodes(data=True):
if "agent_id" in data:
agent = self.agents(data["agent_id"])
self.G.nodes[id]["agent"] = agent
assert not getattr(agent, "node_id", None) or agent.node_id == id
agent.node_id = id
for agent in self.agents():
if hasattr(agent, "node_id"):
node_id = agent["node_id"]
if node_id not in self.G.nodes:
raise ValueError(f"Agent {agent} is assigned to node {agent.node_id} which is not in the network")
node = self.G.nodes[node_id]
if node.get("agent") is not None and node["agent"] != agent:
raise ValueError(f"Node {node_id} already has a different agent assigned to it")
self.G.nodes[node_id]["agent"] = agent
def add_agents(self, agent_classes: List[type], k=None, weights: Optional[List[float]] = None, **kwargs):
if k is None:
k = len(self.G)
if not k:
raise ValueError("Cannot add agents to an empty network")
super().add_agents(agent_classes, k=k, weights=weights, **kwargs)
def agent_for_node_id(self, node_id):
return self.G.nodes[node_id].get("agent")
def populate_network(self, agent_class: List[Model], weights: List[float] = None, **agent_params):
if isinstance(agent_class, type):
agent_class = [agent_class]
else:
agent_class = list(agent_class)
if not weights:
weights = [1] * len(agent_class)
assert len(self.G)
classes = self.random.choices(agent_class, weights, k=len(self.G))
toadd = []
for (cls, (node_id, node)) in zip(classes, self.G.nodes(data=True)):
if "agent" in node:
continue
node["agent"] = None # Reserve
toadd.append(dict(node_id=node_id, topology=self.G, agent_class=cls, **agent_params))
for d in toadd:
a = self.add_agent(**d)
self.G.nodes[d["node_id"]]["agent"] = a
assert all("agent" in node for (_, node) in self.G.nodes(data=True))
assert len(list(self.network_agents))
SoilEnvironment = Environment
class EventedEnvironment(BaseEnvironment):
def broadcast(self, msg, sender=None, expiration=None, ttl=None, **kwargs):
for agent in self.agents(**kwargs):
if agent == sender:
continue
self.logger.debug(f"Telling {repr(agent)}: {msg} ttl={ttl}")
try:
inbox = agent._inbox
except AttributeError:
self.logger.info(
f"Agent {agent.unique_id} cannot receive events because it does not have an inbox"
)
continue
# Allow for AttributeError exceptions in this part of the code
inbox.append(
events.Tell(
payload=msg,
sender=sender,
expiration=expiration if ttl is None else self.now + ttl,
)
)
class Environment(NetworkEnvironment, EventedEnvironment):
"""Default environment class, has both network and event capabilities"""

56
soil/events.py Normal file
View File

@ -0,0 +1,56 @@
from .time import BaseCond
from dataclasses import dataclass, field
from typing import Any
from uuid import uuid4
class Event:
pass
@dataclass
class Message:
payload: Any
sender: Any = None
expiration: float = None
timestamp: float = None
id: int = field(default_factory=uuid4)
def expired(self, when):
return self.expiration is not None and self.expiration < when
class Reply(Message):
source: Message
class ReplyCond(BaseCond):
def __init__(self, ask, *args, **kwargs):
self._ask = ask
super().__init__(*args, **kwargs)
def ready(self, agent, time):
return self._ask.reply is not None or self._ask.expired(time)
def return_value(self, agent):
if self._ask.expired(agent.now):
raise TimedOut()
return self._ask.reply
def __repr__(self):
return f"ReplyCond({self._ask.id})"
class Ask(Message):
reply: Message = None
def replied(self, expiration=None):
return ReplyCond(self)
class Tell(Message):
pass
class TimedOut(Exception):
pass

288
soil/exporters.py
View File

@ -1,17 +1,21 @@
import os
import csv as csvlib
import time
import sys
from time import time as current_time
from io import BytesIO
from sqlalchemy import create_engine
from textwrap import dedent, indent
import matplotlib.pyplot as plt
import networkx as nx
import pandas as pd
from .serialization import deserialize
from .utils import open_or_reuse, logger, timer
from .serialization import deserialize, serialize
from .utils import try_backup, open_or_reuse, logger, timer
from . import utils
from . import utils, network
class DryRunner(BytesIO):
@ -22,51 +26,59 @@ class DryRunner(BytesIO):
def write(self, txt):
if self.__copy_to:
self.__copy_to.write('{}:::{}'.format(self.__fname, txt))
self.__copy_to.write("{}:::{}".format(self.__fname, txt))
try:
super().write(txt)
except TypeError:
super().write(bytes(txt, 'utf-8'))
super().write(bytes(txt, "utf-8"))
def close(self):
content = '(binary data not shown)'
content = "(binary data not shown)"
try:
content = self.getvalue().decode()
except UnicodeDecodeError:
pass
logger.info('**Not** written to {} (dry run mode):\n\n{}\n\n'.format(self.__fname, content))
logger.info(
"**Not** written to {} (no_dump mode):\n\n{}\n\n".format(
self.__fname, content
)
)
super().close()
class Exporter:
'''
"""
Interface for all exporters. It is not necessary, but it is useful
if you don't plan to implement all the methods.
'''
"""
def __init__(self, simulation, outdir=None, dry_run=None, copy_to=None):
def __init__(self, simulation, outdir=None, dump=True, copy_to=None):
self.simulation = simulation
outdir = outdir or os.path.join(os.getcwd(), 'soil_output')
self.outdir = os.path.join(outdir,
simulation.group or '',
simulation.name)
self.dry_run = dry_run
outdir = outdir or os.path.join(os.getcwd(), "soil_output")
self.outdir = os.path.join(outdir, simulation.group or "", simulation.name)
self.dump = dump
if copy_to is None and not dump:
copy_to = sys.stdout
self.copy_to = copy_to
def start(self):
'''Method to call when the simulation starts'''
def sim_start(self):
"""Method to call when the simulation starts"""
pass
def end(self, stats):
'''Method to call when the simulation ends'''
def sim_end(self):
"""Method to call when the simulation ends"""
pass
def trial(self, env, stats):
'''Method to call when a trial ends'''
def iteration_start(self, env):
"""Method to call when a iteration start"""
pass
def output(self, f, mode='w', **kwargs):
if self.dry_run:
def iteration_end(self, env, params, params_id):
"""Method to call when a iteration ends"""
pass
def output(self, f, mode="w", **kwargs):
if not self.dump:
f = DryRunner(f, copy_to=self.copy_to)
else:
try:
@ -74,85 +86,197 @@ class Exporter:
f = os.path.join(self.outdir, f)
except TypeError:
pass
return open_or_reuse(f, mode=mode, **kwargs)
return open_or_reuse(f, mode=mode, backup=self.simulation.backup, **kwargs)
def get_dfs(self, env, **kwargs):
yield from get_dc_dfs(env.datacollector,
simulation_id=self.simulation.id,
iteration_id=env.id,
**kwargs)
class default(Exporter):
'''Default exporter. Writes sqlite results, as well as the simulation YAML'''
def get_dc_dfs(dc, **kwargs):
dfs = {}
dfe = dc.get_model_vars_dataframe()
dfe.index.rename("step", inplace=True)
dfs["env"] = dfe
try:
dfa = dc.get_agent_vars_dataframe()
dfa.index.rename(["step", "agent_id"], inplace=True)
dfs["agents"] = dfa
except UserWarning:
pass
for table_name in dc.tables:
dfs[table_name] = dc.get_table_dataframe(table_name)
for (name, df) in dfs.items():
for (k, v) in kwargs.items():
df[k] = v
df.set_index(["simulation_id", "iteration_id"], append=True, inplace=True)
def start(self):
if not self.dry_run:
logger.info('Dumping results to %s', self.outdir)
self.simulation.dump_yaml(outdir=self.outdir)
else:
logger.info('NOT dumping results')
def trial(self, env, stats):
if not self.dry_run:
with timer('Dumping simulation {} trial {}'.format(self.simulation.name,
env.name)):
with self.output('{}.sqlite'.format(env.name), mode='wb') as f:
env.dump_sqlite(f)
def end(self, stats):
with timer('Dumping simulation {}\'s stats'.format(self.simulation.name)):
with self.output('{}.sqlite'.format(self.simulation.name), mode='wb') as f:
self.simulation.dump_sqlite(f)
yield from dfs.items()
class SQLite(Exporter):
"""Writes sqlite results"""
sim_started = False
class csv(Exporter):
'''Export the state of each environment (and its agents) in a separate CSV file'''
def trial(self, env, stats):
with timer('[CSV] Dumping simulation {} trial {} @ dir {}'.format(self.simulation.name,
env.name,
self.outdir)):
with self.output('{}.csv'.format(env.name)) as f:
env.dump_csv(f)
def sim_start(self):
if not self.dump:
logger.debug("NOT dumping results")
return
self.dbpath = os.path.join(self.outdir, f"{self.simulation.name}.sqlite")
logger.info("Dumping results to %s", self.dbpath)
if self.simulation.backup:
try_backup(self.dbpath, remove=True)
with self.output('{}.stats.csv'.format(env.name)) as f:
statwriter = csvlib.writer(f, delimiter='\t', quotechar='"', quoting=csvlib.QUOTE_ALL)
if self.simulation.overwrite:
if os.path.exists(self.dbpath):
os.remove(self.dbpath)
for stat in stats:
statwriter.writerow(stat)
self.engine = create_engine(f"sqlite:///{self.dbpath}", echo=False)
sim_dict = {k: serialize(v)[0] for (k,v) in self.simulation.to_dict().items()}
sim_dict["simulation_id"] = self.simulation.id
df = pd.DataFrame([sim_dict])
df.to_sql("configuration", con=self.engine, if_exists="append")
class gexf(Exporter):
def trial(self, env, stats):
if self.dry_run:
logger.info('Not dumping GEXF in dry_run mode')
def iteration_end(self, env, params, params_id, *args, **kwargs):
if not self.dump:
logger.info("Running in NO DUMP mode. Results will NOT be saved to a DB.")
return
with timer('[GEXF] Dumping simulation {} trial {}'.format(self.simulation.name,
env.name)):
with self.output('{}.gexf'.format(env.name), mode='wb') as f:
env.dump_gexf(f)
with timer(
"Dumping simulation {} iteration {}".format(self.simulation.name, env.id)
):
pd.DataFrame([{"simulation_id": self.simulation.id,
"params_id": params_id,
"iteration_id": env.id,
"key": k,
"value": serialize(v)[0]} for (k,v) in params.items()]).to_sql("parameters", con=self.engine, if_exists="append")
for (t, df) in self.get_dfs(env, params_id=params_id):
df.to_sql(t, con=self.engine, if_exists="append")
class csv(Exporter):
"""Export the state of each environment (and its agents) a CSV file for the simulation"""
def sim_start(self):
super().sim_start()
def iteration_end(self, env, params, params_id, *args, **kwargs):
with timer(
"[CSV] Dumping simulation {} iteration {} @ dir {}".format(
self.simulation.name, env.id, self.outdir
)
):
for (df_name, df) in self.get_dfs(env, params_id=params_id):
with self.output("{}.{}.csv".format(env.id, df_name), mode="a") as f:
df.to_csv(f)
# TODO: reimplement GEXF exporting without history
class gexf(Exporter):
def iteration_end(self, env, *args, **kwargs):
if not self.dump:
logger.info("Not dumping GEXF (NO_DUMP mode)")
return
with timer(
"[GEXF] Dumping simulation {} iteration {}".format(self.simulation.name, env.id)
):
with self.output("{}.gexf".format(env.id), mode="wb") as f:
network.dump_gexf(env.history_to_graph(), f)
self.dump_gexf(env, f)
class dummy(Exporter):
def sim_start(self):
with self.output("dummy", "w") as f:
f.write("simulation started @ {}\n".format(current_time()))
def start(self):
with self.output('dummy', 'w') as f:
f.write('simulation started @ {}\n'.format(time.time()))
def iteration_start(self, env):
with self.output("dummy", "w") as f:
f.write("iteration started@ {}\n".format(current_time()))
def trial(self, env, stats):
with self.output('dummy', 'w') as f:
for i in env.history_to_tuples():
f.write(','.join(map(str, i)))
f.write('\n')
def sim(self, stats):
with self.output('dummy', 'a') as f:
f.write('simulation ended @ {}\n'.format(time.time()))
def iteration_end(self, env, *args, **kwargs):
with self.output("dummy", "w") as f:
f.write("iteration ended@ {}\n".format(current_time()))
def sim_end(self):
with self.output("dummy", "a") as f:
f.write("simulation ended @ {}\n".format(current_time()))
class graphdrawing(Exporter):
def trial(self, env, stats):
def iteration_end(self, env, *args, **kwargs):
# Outside effects
f = plt.figure()
nx.draw(env.G, node_size=10, width=0.2, pos=nx.spring_layout(env.G, scale=100), ax=f.add_subplot(111))
with open('graph-{}.png'.format(env.name)) as f:
nx.draw(
env.G,
node_size=10,
width=0.2,
pos=nx.spring_layout(env.G, scale=100),
ax=f.add_subplot(111),
)
with open("graph-{}.png".format(env.id)) as f:
f.savefig(f)
class summary(Exporter):
"""Print a summary of each iteration to sys.stdout"""
def iteration_end(self, env, *args, **kwargs):
msg = ""
for (t, df) in self.get_dfs(env):
if not len(df):
continue
tabs = "\t" * 2
description = indent(str(df.describe()), tabs)
last_line = indent(str(df.iloc[-1:]), tabs)
# value_counts = indent(str(df.value_counts()), tabs)
value_counts = indent(str(df.apply(lambda x: x.value_counts()).T.stack()), tabs)
msg += dedent("""
Dataframe {t}:
Last line: :
{last_line}
Description:
{description}
Value counts:
{value_counts}
""").format(**locals())
logger.info(msg)
class YAML(Exporter):
"""Writes the configuration of the simulation to a YAML file"""
def sim_start(self):
if not self.dump:
logger.debug("NOT dumping results")
return
with self.output(self.simulation.id + ".dumped.yml") as f:
logger.info(f"Dumping simulation configuration to {self.outdir}")
f.write(self.simulation.to_yaml())
class default(Exporter):
"""Default exporter. Writes sqlite results, as well as the simulation YAML"""
def __init__(self, *args, exporter_cls=[], **kwargs):
exporter_cls = exporter_cls or [YAML, SQLite]
self.inner = [cls(*args, **kwargs) for cls in exporter_cls]
def sim_start(self, *args, **kwargs):
for exporter in self.inner:
exporter.sim_start(*args, **kwargs)
def sim_end(self, *args, **kwargs):
for exporter in self.inner:
exporter.sim_end(*args, **kwargs)
def iteration_end(self, *args, **kwargs):
for exporter in self.inner:
exporter.iteration_end(*args, **kwargs)

83
soil/network.py Normal file
View File

@ -0,0 +1,83 @@
from __future__ import annotations
from typing import Dict
import os
import sys
import random
import networkx as nx
from . import config, serialization, basestring
def from_topology(topology, dir_path: str = None):
if topology is None:
return nx.Graph()
if isinstance(topology, nx.Graph):
return topology
# If it's a dict, assume it's a node-link graph
if isinstance(topology, dict):
try:
return nx.json_graph.node_link_graph(topology)
except Exception as ex:
raise ValueError("Unknown topology format")
# Otherwise, treat like a path
path = topology
if dir_path and not os.path.isabs(path):
path = os.path.join(dir_path, path)
extension = os.path.splitext(path)[1][1:]
kwargs = {}
if extension == "gexf":
kwargs["version"] = "1.2draft"
kwargs["node_type"] = int
try:
method = getattr(nx.readwrite, "read_" + extension)
except AttributeError:
raise AttributeError("Unknown format")
return method(path, **kwargs)
def from_params(generator, dir_path: str = None, **params):
if dir_path not in sys.path:
sys.path.append(dir_path)
method = serialization.deserializer(
generator,
known_modules=[
"networkx.generators",
],
)
return method(**params)
def find_unassigned(G, shuffle=False, random=random):
"""
Link an agent to a node in a topology.
If node_id is None, a node without an agent_id will be found.
"""
candidates = list(G.nodes(data=True))
if shuffle:
random.shuffle(candidates)
for next_id, data in candidates:
if "agent" not in data:
return next_id
return None
def dump_gexf(G, f):
for node in G.nodes():
if "pos" in G.nodes[node]:
G.nodes[node]["viz"] = {
"position": {
"x": G.nodes[node]["pos"][0],
"y": G.nodes[node]["pos"][1],
"z": 0.0,
}
}
del G.nodes[node]["pos"]
nx.write_gexf(G, f, version="1.2draft")

32
soil/parameters.py Normal file
View File

@ -0,0 +1,32 @@
from __future__ import annotations
from typing_extensions import Annotated
import annotated_types
from typing import *
from dataclasses import dataclass
class Parameter:
pass
def floatrange(
*,
gt: Optional[float] = None,
ge: Optional[float] = None,
lt: Optional[float] = None,
le: Optional[float] = None,
multiple_of: Optional[float] = None,
) -> type[float]:
return Annotated[
float,
annotated_types.Interval(gt=gt, ge=ge, lt=lt, le=le),
annotated_types.MultipleOf(multiple_of) if multiple_of is not None else None,
]
function = Annotated[Callable, Parameter]
Integer = Annotated[int, Parameter]
Float = Annotated[float, Parameter]
probability = floatrange(ge=0, le=1)

226
soil/serialization.py
View File

@ -2,58 +2,28 @@ import os
import logging
import ast
import sys
import re
import importlib
import importlib.machinery, importlib.util
from glob import glob
from itertools import product, chain
import yaml
import networkx as nx
from . import config
from jinja2 import Template
logger = logging.getLogger('soil')
def load_network(network_params, dir_path=None):
G = nx.Graph()
if 'path' in network_params:
path = network_params['path']
if dir_path and not os.path.isabs(path):
path = os.path.join(dir_path, path)
extension = os.path.splitext(path)[1][1:]
kwargs = {}
if extension == 'gexf':
kwargs['version'] = '1.2draft'
kwargs['node_type'] = int
try:
method = getattr(nx.readwrite, 'read_' + extension)
except AttributeError:
raise AttributeError('Unknown format')
G = method(path, **kwargs)
elif 'generator' in network_params:
net_args = network_params.copy()
net_gen = net_args.pop('generator')
if dir_path not in sys.path:
sys.path.append(dir_path)
method = deserializer(net_gen,
known_modules=['networkx.generators',])
G = method(**net_args)
return G
logger = logging.getLogger("soil")
def load_file(infile):
folder = os.path.dirname(infile)
if folder not in sys.path:
sys.path.append(folder)
with open(infile, 'r') as f:
with open(infile, "r") as f:
return list(chain.from_iterable(map(expand_template, load_string(f))))
@ -62,14 +32,15 @@ def load_string(string):
def expand_template(config):
if 'template' not in config:
if "template" not in config:
yield config
return
if 'vars' not in config:
raise ValueError(('You must provide a definition of variables'
' for the template.'))
if "vars" not in config:
raise ValueError(
("You must provide a definition of variables" " for the template.")
)
template = config['template']
template = config["template"]
if not isinstance(template, str):
template = yaml.dump(template)
@ -81,9 +52,9 @@ def expand_template(config):
blank_str = template.render({k: 0 for k in params[0].keys()})
blank = list(load_string(blank_str))
if len(blank) > 1:
raise ValueError('Templates must not return more than one configuration')
if 'name' in blank[0]:
raise ValueError('Templates cannot be named, use group instead')
raise ValueError("Templates must not return more than one configuration")
if "name" in blank[0]:
raise ValueError("Templates cannot be named, use group instead")
for ps in params:
string = template.render(ps)
@ -92,131 +63,200 @@ def expand_template(config):
def params_for_template(config):
sampler_config = config.get('sampler', {'N': 100})
sampler = sampler_config.pop('method', 'SALib.sample.morris.sample')
sampler_config = config.get("sampler", {"N": 100})
sampler = sampler_config.pop("method", "SALib.sample.morris.sample")
sampler = deserializer(sampler)
bounds = config['vars']['bounds']
bounds = config["vars"]["bounds"]
problem = {
'num_vars': len(bounds),
'names': list(bounds.keys()),
'bounds': list(v for v in bounds.values())
"num_vars": len(bounds),
"names": list(bounds.keys()),
"bounds": list(v for v in bounds.values()),
}
samples = sampler(problem, **sampler_config)
lists = config['vars'].get('lists', {})
lists = config["vars"].get("lists", {})
names = list(lists.keys())
values = list(lists.values())
combs = list(product(*values))
allnames = names + problem['names']
allvalues = [(list(i[0])+list(i[1])) for i in product(combs, samples)]
allnames = names + problem["names"]
allvalues = [(list(i[0]) + list(i[1])) for i in product(combs, samples)]
params = list(map(lambda x: dict(zip(allnames, x)), allvalues))
return params
def load_files(*patterns, **kwargs):
for pattern in patterns:
for i in glob(pattern, **kwargs):
for config in load_file(i):
for i in glob(pattern, **kwargs, recursive=True):
for cfg in load_file(i):
path = os.path.abspath(i)
if 'dir_path' not in config:
config['dir_path'] = os.path.dirname(path)
yield config, path
yield cfg, path
def load_config(config):
if isinstance(config, dict):
yield config, os.getcwd()
def load_config(cfg):
if isinstance(cfg, dict):
yield config.load_config(cfg), os.getcwd()
else:
yield from load_files(config)
yield from load_files(cfg)
builtins = importlib.import_module('builtins')
builtins = importlib.import_module("builtins")
def name(value, known_modules=[]):
'''Return a name that can be imported, to serialize/deserialize an object'''
KNOWN_MODULES = {
'soil': None,
}
MODULE_FILES = {}
def add_source_file(file):
"""Add a file to the list of known modules"""
file = os.path.abspath(file)
if file in MODULE_FILES:
logger.warning(f"File {file} already added as module {MODULE_FILES[file]}. Reloading")
remove_source_file(file)
modname = f"imported_module_{len(MODULE_FILES)}"
loader = importlib.machinery.SourceFileLoader(modname, file)
spec = importlib.util.spec_from_loader(loader.name, loader)
my_module = importlib.util.module_from_spec(spec)
loader.exec_module(my_module)
MODULE_FILES[file] = modname
KNOWN_MODULES[modname] = my_module
def remove_source_file(file):
"""Remove a file from the list of known modules"""
file = os.path.abspath(file)
modname = None
try:
modname = MODULE_FILES.pop(file)
KNOWN_MODULES.pop(modname)
except KeyError as ex:
raise ValueError(f"File {file} had not been added as a module: {ex}")
def get_module(modname):
"""Get a module from the list of known modules"""
if modname not in KNOWN_MODULES or KNOWN_MODULES[modname] is None:
module = importlib.import_module(modname)
KNOWN_MODULES[modname] = module
return KNOWN_MODULES[modname]
def name(value, known_modules=KNOWN_MODULES):
"""Return a name that can be imported, to serialize/deserialize an object"""
if value is None:
return 'None'
return "None"
if not isinstance(value, type): # Get the class name first
value = type(value)
tname = value.__name__
if hasattr(builtins, tname):
return tname
modname = value.__module__
if modname == '__main__':
if modname == "__main__":
return tname
if known_modules and modname in known_modules:
return tname
for kmod in known_modules:
if not kmod:
continue
module = importlib.import_module(kmod)
module = get_module(kmod)
if hasattr(module, tname):
return tname
return '{}.{}'.format(modname, tname)
return "{}.{}".format(modname, tname)
def serializer(type_):
if type_ != 'str' and hasattr(builtins, type_):
if type_ != "str" and hasattr(builtins, type_):
return repr
return lambda x: x
def serialize(v, known_modules=[]):
'''Get a text representation of an object.'''
def serialize(v, known_modules=KNOWN_MODULES):
"""Get a text representation of an object."""
tname = name(v, known_modules=known_modules)
func = serializer(tname)
return func(v), tname
def deserializer(type_, known_modules=[]):
def serialize_dict(d, known_modules=KNOWN_MODULES):
try:
d = dict(d)
except (ValueError, TypeError) as ex:
return serialize(d)[0]
for (k, v) in reversed(list(d.items())):
if isinstance(v, dict):
d[k] = serialize_dict(v, known_modules=known_modules)
elif isinstance(v, list):
for ix in range(len(v)):
v[ix] = serialize_dict(v[ix], known_modules=known_modules)
elif isinstance(v, type):
d[k] = serialize(v, known_modules=known_modules)[1]
return d
IS_CLASS = re.compile(r"<class '(.*)'>")
def deserializer(type_, known_modules=KNOWN_MODULES):
if type(type_) != str: # Already deserialized
return type_
if type_ == 'str':
return lambda x='': x
if type_ == 'None':
if type_ == "str":
return lambda x="": x
if type_ == "None":
return lambda x=None: None
if hasattr(builtins, type_): # Check if it's a builtin type
cls = getattr(builtins, type_)
return lambda x=None: ast.literal_eval(x) if x is not None else cls()
match = IS_CLASS.match(type_)
if match:
modname, tname = match.group(1).rsplit(".", 1)
module = get_module(modname)
cls = getattr(module, tname)
return getattr(cls, "deserialize", cls)
# Otherwise, see if we can find the module and the class
modules = known_modules or []
options = []
for mod in modules:
for mod in known_modules:
if mod:
options.append((mod, type_))
if '.' in type_: # Fully qualified module
if "." in type_: # Fully qualified module
module, type_ = type_.rsplit(".", 1)
options.append ((module, type_))
options.append((module, type_))
errors = []
for modname, tname in options:
try:
module = importlib.import_module(modname)
module = get_module(modname)
cls = getattr(module, tname)
return getattr(cls, 'deserialize', cls)
return getattr(cls, "deserialize", cls)
except (ImportError, AttributeError) as ex:
errors.append((modname, tname, ex))
raise Exception('Could not find type {}. Tried: {}'.format(type_, errors))
raise ValueError('Could not find type "{}". Tried: {}'.format(type_, errors))
def deserialize(type_, value=None, **kwargs):
'''Get an object from a text representation'''
def deserialize(type_, value=None, globs=None, **kwargs):
"""Get an object from a text representation"""
if not isinstance(type_, str):
return type_
if globs and type_ in globs:
des = globs[type_]
else:
try:
des = deserializer(type_, **kwargs)
except ValueError as ex:
try:
des = eval(type_)
except Exception:
raise ex
if value is None:
return des
return des(value)
def deserialize_all(names, *args, known_modules=['soil'], **kwargs):
'''Return the set of exporters for a simulation, given the exporter names'''
exporters = []
def deserialize_all(names, *args, known_modules=KNOWN_MODULES, **kwargs):
"""Return the list of deserialized objects"""
objects = []
for name in names:
mod = deserialize(name, known_modules=known_modules)
exporters.append(mod(*args, **kwargs))
return exporters
objects.append(mod(*args, **kwargs))
return objects

638
soil/simulation.py
View File

@ -1,355 +1,395 @@
import os
import importlib
from time import time as current_time, strftime
import sys
import yaml
import traceback
import hashlib
import inspect
import logging
import networkx as nx
from time import strftime
from networkx.readwrite import json_graph
from multiprocessing import Pool
from tqdm.auto import tqdm
from textwrap import dedent
from dataclasses import dataclass, field, asdict, replace
from typing import Any, Dict, Union, Optional, List
from functools import partial
from tsih import History
from contextlib import contextmanager
from itertools import product
import json
import pickle
from . import serialization, utils, basestring, agents
from . import serialization, exporters, utils, basestring, agents
from .environment import Environment
from .utils import logger
from .exporters import default
from .stats import defaultStats
from .utils import logger, run_and_return_exceptions
from .debugging import set_trace
_AVOID_RUNNING = False
_QUEUED = []
@contextmanager
def do_not_run():
global _AVOID_RUNNING
_AVOID_RUNNING = True
try:
logger.debug("NOT RUNNING")
yield
finally:
logger.debug("RUNNING AGAIN")
_AVOID_RUNNING = False
#TODO: change documentation for simulation
def _iter_queued():
while _QUEUED:
(cls, params) = _QUEUED.pop(0)
yield replace(cls, parameters=params)
# TODO: change documentation for simulation
# TODO: rename iterations to iterations
# TODO: make parameters a dict of iterable/any
@dataclass
class Simulation:
"""
Similar to nsim.NetworkSimulation with three main differences:
1) agent type can be specified by name or by class.
2) instead of just one type, a network agents distribution can be used.
The distribution specifies the weight (or probability) of each
agent type in the topology. This is an example distribution: ::
[
{'agent_type': 'agent_type_1',
'weight': 0.2,
'state': {
'id': 0
}
},
{'agent_type': 'agent_type_2',
'weight': 0.8,
'state': {
'id': 1
}
}
]
In this example, 20% of the nodes will be marked as type
'agent_type_1'.
3) if no initial state is given, each node's state will be set
to `{'id': 0}`.
Parameters
---------
name : str, optional
name of the Simulation
group : str, optional
a group name can be used to link simulations
topology : networkx.Graph instance, optional
network_params : dict
parameters used to create a topology with networkx, if no topology is given
network_agents : dict
definition of agents to populate the topology with
agent_type : NetworkAgent subclass, optional
Default type of NetworkAgent to use for nodes not specified in network_agents
states : list, optional
List of initial states corresponding to the nodes in the topology. Basic form is a list of integers
whose value indicates the state
dir_path: str, optional
Directory path to load simulation assets (files, modules...)
seed : str, optional
Seed to use for the random generator
num_trials : int, optional
Number of independent simulation runs
max_time : int, optional
Time how long the simulation should run
environment_params : dict, optional
Dictionary of globally-shared environmental parameters
environment_agents: dict, optional
Similar to network_agents. Distribution of Agents that control the environment
environment_class: soil.environment.Environment subclass, optional
Class for the environment. It defailts to soil.environment.Environment
load_module : str, module name, deprecated
If specified, soil will load the content of this module under 'soil.agents.custom'
A simulation is a collection of agents and a model. It is responsible for running the model and agents, and collecting data from them.
Args:
version: The version of the simulation. This is used to determine how to load the simulation.
name: The name of the simulation.
description: A description of the simulation.
group: The group that the simulation belongs to.
model: The model to use for the simulation. This can be a string or a class.
parameters: The parameters to pass to the model.
matrix: A matrix of values for each parameter.
seed: The seed to use for the simulation.
dir_path: The directory path to use for the simulation.
max_time: The maximum time to run the simulation.
max_steps: The maximum number of steps to run the simulation.
interval: The interval to use for the simulation.
iterations: The number of iterations (times) to run the simulation.
num_processes: The number of processes to use for the simulation. If greater than one, simulations will be performed in parallel. This may make debugging and error handling difficult.
tables: The tables to use in the simulation datacollector
agent_reporters: The agent reporters to use in the datacollector
model_reporters: The model reporters to use in the datacollector
dry_run: Whether or not to run the simulation. If True, the simulation will not be run.
backup: Whether or not to backup the simulation. If True, the simulation files will be backed up to a different directory.
overwrite: Whether or not to replace existing simulation data.
source_file: Python file to use to find additional classes.
"""
def __init__(self, name=None, group=None, topology=None, network_params=None,
network_agents=None, agent_type=None, states=None,
default_state=None, interval=1, num_trials=1,
max_time=100, load_module=None, seed=None,
dir_path=None, environment_agents=None,
environment_params=None, environment_class=None,
**kwargs):
version: str = "2"
source_file: Optional[str] = None
name: Optional[str] = None
description: Optional[str] = ""
group: str = None
backup: bool = False
overwrite: bool = False
dry_run: bool = False
dump: bool = False
model: Union[str, type] = "soil.Environment"
parameters: dict = field(default_factory=dict)
matrix: dict = field(default_factory=dict)
seed: str = "default"
dir_path: str = field(default_factory=lambda: os.getcwd())
max_time: float = None
max_steps: int = None
interval: int = 1
iterations: int = 1
num_processes: Optional[int] = 1
exporters: Optional[List[str]] = field(default_factory=lambda: [exporters.default])
model_reporters: Optional[Dict[str, Any]] = field(default_factory=dict)
agent_reporters: Optional[Dict[str, Any]] = field(default_factory=dict)
tables: Optional[Dict[str, Any]] = field(default_factory=dict)
outdir: str = field(default_factory=lambda: os.path.join(os.getcwd(), "soil_output"))
# outdir: Optional[str] = None
exporter_params: Optional[Dict[str, Any]] = field(default_factory=dict)
level: int = logging.INFO
skip_test: Optional[bool] = False
debug: Optional[bool] = False
self.load_module = load_module
self.network_params = network_params
self.name = name or 'Unnamed'
self.seed = str(seed or name)
self._id = '{}_{}'.format(self.name, strftime("%Y-%m-%d_%H.%M.%S"))
self.group = group or ''
self.num_trials = num_trials
self.max_time = max_time
self.default_state = default_state or {}
self.dir_path = dir_path or os.getcwd()
self.interval = interval
sys.path += list(x for x in [os.getcwd(), self.dir_path] if x not in sys.path)
if topology is None:
topology = serialization.load_network(network_params,
dir_path=self.dir_path)
elif isinstance(topology, basestring) or isinstance(topology, dict):
topology = json_graph.node_link_graph(topology)
self.topology = nx.Graph(topology)
self.environment_params = environment_params or {}
self.environment_class = serialization.deserialize(environment_class,
known_modules=['soil.environment', ]) or Environment
environment_agents = environment_agents or []
self.environment_agents = agents._convert_agent_types(environment_agents,
known_modules=[self.load_module])
distro = agents.calculate_distribution(network_agents,
agent_type)
self.network_agents = agents._convert_agent_types(distro,
known_modules=[self.load_module])
self.states = agents._validate_states(states,
self.topology)
self._history = History(name=self.name,
backup=False)
def run_simulation(self, *args, **kwargs):
return self.run(*args, **kwargs)
def run(self, *args, **kwargs):
'''Run the simulation and return the list of resulting environments'''
return list(self.run_gen(*args, **kwargs))
def _run_sync_or_async(self, parallel=False, **kwargs):
if parallel and not os.environ.get('SENPY_DEBUG', None):
p = Pool()
func = partial(self.run_trial_exceptions, **kwargs)
for i in p.imap_unordered(func, range(self.num_trials)):
if isinstance(i, Exception):
logger.error('Trial failed:\n\t%s', i.message)
continue
yield i
def __post_init__(self):
if self.name is None:
if isinstance(self.model, str):
self.name = self.model
else:
for i in range(self.num_trials):
yield self.run_trial(trial_id=i,
**kwargs)
self.name = self.model.__name__
self.logger = logger.getChild(self.name)
self.logger.setLevel(self.level)
def run_gen(self, parallel=False, dry_run=False,
exporters=[default, ], stats=[], outdir=None, exporter_params={},
stats_params={}, log_level=None,
**kwargs):
'''Run the simulation and yield the resulting environments.'''
if log_level:
logger.setLevel(log_level)
logger.info('Using exporters: %s', exporters or [])
logger.info('Output directory: %s', outdir)
exporters = serialization.deserialize_all(exporters,
if self.source_file:
source_file = self.source_file
if not os.path.isabs(source_file):
source_file = os.path.abspath(os.path.join(self.dir_path, source_file))
serialization.add_source_file(source_file)
self.source_file = source_file
if isinstance(self.model, str):
self.model = serialization.deserialize(self.model)
def deserialize_reporters(reporters):
for (k, v) in reporters.items():
if isinstance(v, str) and v.startswith("py:"):
reporters[k] = serialization.deserialize(v.split(":", 1)[1])
return reporters
self.agent_reporters = deserialize_reporters(self.agent_reporters)
self.model_reporters = deserialize_reporters(self.model_reporters)
self.tables = deserialize_reporters(self.tables)
if self.source_file:
serialization.remove_source_file(self.source_file)
self.id = f"{self.name}_{current_time()}"
def run(self, **kwargs):
"""Run the simulation and return the list of resulting environments"""
if kwargs:
return replace(self, **kwargs).run()
self.logger.debug(
dedent(
"""
Simulation:
---
"""
)
+ self.to_yaml()
)
param_combinations = self._collect_params(**kwargs)
if _AVOID_RUNNING:
_QUEUED.extend((self, param) for param in param_combinations)
return []
self.logger.debug("Using exporters: %s", self.exporters or [])
exporters = serialization.deserialize_all(
self.exporters,
simulation=self,
known_modules=['soil.exporters',],
dry_run=dry_run,
outdir=outdir,
**exporter_params)
stats = serialization.deserialize_all(simulation=self,
names=stats,
known_modules=['soil.stats',],
**stats_params)
known_modules=[
"soil.exporters",
],
dump=self.dump and not self.dry_run,
outdir=self.outdir,
**self.exporter_params,
)
with utils.timer('simulation {}'.format(self.name)):
for stat in stats:
stat.start()
results = []
for exporter in exporters:
exporter.sim_start()
for params in tqdm(param_combinations, desc=self.name, unit="configuration"):
for (k, v) in params.items():
tqdm.write(f"{k} = {v}")
sha = hashlib.sha256()
sha.update(repr(sorted(params.items())).encode())
params_id = sha.hexdigest()[:7]
for env in self._run_iters_for_params(params):
for exporter in exporters:
exporter.iteration_end(env, params, params_id)
results.append(env)
for exporter in exporters:
exporter.start()
for env in self._run_sync_or_async(parallel=parallel,
log_level=log_level,
**kwargs):
exporter.sim_end()
collected = list(stat.trial(env) for stat in stats)
return results
saved = self.save_stats(collected, t_step=env.now, trial_id=env.name)
def _collect_params(self):
for exporter in exporters:
exporter.trial(env, saved)
parameters = []
if self.parameters:
parameters.append(self.parameters)
if self.matrix:
assert isinstance(self.matrix, dict)
for values in product(*(self.matrix.values())):
parameters.append(dict(zip(self.matrix.keys(), values)))
if not parameters:
parameters = [{}]
if self.dump:
self.logger.info("Output directory: %s", self.outdir)
return parameters
def _run_iters_for_params(
self,
params
):
"""Run the simulation and yield the resulting environments."""
try:
if self.source_file:
serialization.add_source_file(self.source_file)
with utils.timer(f"running for config {params}"):
if self.dry_run:
def func(*args, **kwargs):
return None
else:
func = self._run_model
for env in tqdm(utils.run_parallel(
func=func,
iterable=range(self.iterations),
**params,
), total=self.iterations, leave=False):
if env is None and self.dry_run:
continue
yield env
finally:
if self.source_file:
serialization.remove_source_file(self.source_file)
def _get_env(self, iteration_id, params):
"""Create an environment for a iteration of the simulation"""
collected = list(stat.end() for stat in stats)
saved = self.save_stats(collected)
iteration_id = str(iteration_id)
for exporter in exporters:
exporter.end(saved)
agent_reporters = self.agent_reporters
agent_reporters.update(params.pop("agent_reporters", {}))
model_reporters = self.model_reporters
model_reporters.update(params.pop("model_reporters", {}))
return self.model(
id=iteration_id,
seed=f"{self.seed}_iteration_{iteration_id}",
dir_path=self.dir_path,
interval=self.interval,
logger=self.logger.getChild(iteration_id),
agent_reporters=agent_reporters,
model_reporters=model_reporters,
tables=self.tables,
**params,
)
def save_stats(self, collection, **kwargs):
stats = dict(kwargs)
for stat in collection:
stats.update(stat)
self._history.save_stats(utils.flatten_dict(stats))
return stats
def get_stats(self, **kwargs):
return self._history.get_stats(**kwargs)
def log_stats(self, stats):
logger.info('Stats: \n{}'.format(yaml.dump(stats, default_flow_style=False)))
def get_env(self, trial_id=0, **kwargs):
'''Create an environment for a trial of the simulation'''
opts = self.environment_params.copy()
opts.update({
'name': '{}_trial_{}'.format(self.name, trial_id),
'topology': self.topology.copy(),
'network_params': self.network_params,
'seed': '{}_trial_{}'.format(self.seed, trial_id),
'initial_time': 0,
'interval': self.interval,
'network_agents': self.network_agents,
'initial_time': 0,
'states': self.states,
'dir_path': self.dir_path,
'default_state': self.default_state,
'environment_agents': self.environment_agents,
})
opts.update(kwargs)
env = self.environment_class(**opts)
return env
def run_trial(self, trial_id=0, until=None, log_level=logging.INFO, **opts):
def _run_model(self, iteration_id, **params):
"""
Run a single trial of the simulation
Run a single iteration of the simulation
"""
if log_level:
logger.setLevel(log_level)
# Set-up trial environment and graph
until = until or self.max_time
env = self.get_env(trial_id=trial_id, **opts)
# Set up agents on nodes
with utils.timer('Simulation {} trial {}'.format(self.name, trial_id)):
env.run(until)
return env
# Set-up iteration environment and graph
model = self._get_env(iteration_id, params)
with utils.timer("Simulation {} iteration {}".format(self.name, iteration_id)):
def run_trial_exceptions(self, *args, **kwargs):
'''
A wrapper for run_trial that catches exceptions and returns them.
It is meant for async simulations
'''
try:
return self.run_trial(*args, **kwargs)
except Exception as ex:
if ex.__cause__ is not None:
ex = ex.__cause__
ex.message = ''.join(traceback.format_exception(type(ex), ex, ex.__traceback__)[:])
return ex
max_time = self.max_time
max_steps = self.max_steps
if (max_time is not None) and (max_steps is not None):
is_done = lambda model: (not model.running) or (model.schedule.time >= max_time) or (model.schedule.steps >= max_steps)
elif max_time is not None:
is_done = lambda model: (not model.running) or (model.schedule.time >= max_time)
elif max_steps is not None:
is_done = lambda model: (not model.running) or (model.schedule.steps >= max_steps)
else:
is_done = lambda model: not model.running
if not model.schedule.agents:
raise Exception("No agents in model. This is probably a bug. Make sure that the model has agents scheduled after its initialization.")
newline = "\n"
self.logger.debug(
dedent(
f"""
Model stats:
Agent count: { model.schedule.get_agent_count() }):
Topology size: { len(model.G) if hasattr(model, "G") else 0 }
"""
)
)
if self.debug:
set_trace()
while not is_done(model):
self.logger.debug(
f'Simulation time {model.schedule.time}/{max_time}.'
)
model.step()
return model
def to_dict(self):
return self.__getstate__()
d = asdict(self)
return serialization.serialize_dict(d)
def to_yaml(self):
return yaml.dump(self.to_dict())
def dump_yaml(self, f=None, outdir=None):
if not f and not outdir:
raise ValueError('specify a file or an output directory')
if not f:
f = os.path.join(outdir, '{}.dumped.yml'.format(self.name))
with utils.open_or_reuse(f, 'w') as f:
f.write(self.to_yaml())
def dump_pickle(self, f=None, outdir=None):
if not outdir and not f:
raise ValueError('specify a file or an output directory')
if not f:
f = os.path.join(outdir,
'{}.simulation.pickle'.format(self.name))
with utils.open_or_reuse(f, 'wb') as f:
pickle.dump(self, f)
def dump_sqlite(self, f):
return self._history.dump(f)
def __getstate__(self):
state={}
for k, v in self.__dict__.items():
if k[0] != '_':
state[k] = v
state['topology'] = json_graph.node_link_data(self.topology)
state['network_agents'] = agents.serialize_definition(self.network_agents,
known_modules = [])
state['environment_agents'] = agents.serialize_definition(self.environment_agents,
known_modules = [])
state['environment_class'] = serialization.serialize(self.environment_class,
known_modules=['soil.environment'])[1] # func, name
if state['load_module'] is None:
del state['load_module']
return state
def __setstate__(self, state):
self.__dict__ = state
self.load_module = getattr(self, 'load_module', None)
if self.dir_path not in sys.path:
sys.path += [self.dir_path, os.getcwd()]
self.topology = json_graph.node_link_graph(state['topology'])
self.network_agents = agents.calculate_distribution(agents._convert_agent_types(self.network_agents))
self.environment_agents = agents._convert_agent_types(self.environment_agents,
known_modules=[self.load_module])
self.environment_class = serialization.deserialize(self.environment_class,
known_modules=[self.load_module, 'soil.environment', ]) # func, name
def iter_from_file(*files, **kwargs):
for f in files:
try:
yield from iter_from_py(f, **kwargs)
except ValueError as ex:
yield from iter_from_config(f, **kwargs)
def all_from_config(config):
def from_file(*args, **kwargs):
return list(iter_from_file(*args, **kwargs))
def iter_from_config(*cfgs, **kwargs):
for config in cfgs:
configs = list(serialization.load_config(config))
for config, _ in configs:
sim = Simulation(**config)
yield sim
for config, path in configs:
d = dict(config)
d.update(kwargs)
if "dir_path" not in d:
d["dir_path"] = os.path.dirname(path)
yield Simulation(**d)
def from_config(conf_or_path):
config = list(serialization.load_config(conf_or_path))
if len(config) > 1:
raise AttributeError('Provide only one configuration')
config = config[0][0]
sim = Simulation(**config)
return sim
lst = list(iter_from_config(conf_or_path))
if len(lst) > 1:
raise AttributeError("Provide only one configuration")
return lst[0]
def run_from_config(*configs, **kwargs):
for config_def in configs:
# logger.info("Found {} config(s)".format(len(ls)))
for config, path in serialization.load_config(config_def):
name = config.get('name', 'unnamed')
logger.info("Using config(s): {name}".format(name=name))
def iter_from_py(pyfile, module_name='imported_file', **kwargs):
"""Try to load every Simulation instance in a given Python file"""
import importlib
added = False
sims = []
assert not _AVOID_RUNNING
with do_not_run():
assert _AVOID_RUNNING
spec = importlib.util.spec_from_file_location(module_name, pyfile)
folder = os.path.dirname(pyfile)
if folder not in sys.path:
added = True
sys.path.append(folder)
if not spec:
raise ValueError(f"{pyfile} does not seem to be a Python module")
module = importlib.util.module_from_spec(spec)
sys.modules[module_name] = module
spec.loader.exec_module(module)
for (_name, sim) in inspect.getmembers(module, lambda x: isinstance(x, Simulation)):
sims.append(sim)
for sim in _iter_queued():
sims.append(sim)
if not sims:
for (_name, sim) in inspect.getmembers(module, lambda x: inspect.isclass(x) and issubclass(x, Simulation)):
sims.append(sim(**kwargs))
del sys.modules[module_name]
assert not _AVOID_RUNNING
if not sims:
raise AttributeError(f"No valid configurations found in {pyfile}")
if added:
sys.path.remove(folder)
for sim in sims:
yield replace(sim, **kwargs)
dir_path = config.pop('dir_path', os.path.dirname(path))
sim = Simulation(dir_path=dir_path,
**config)
def from_py(pyfile):
return next(iter_from_py(pyfile))
def run_from_file(*files, **kwargs):
for sim in iter_from_file(*files):
logger.info(f"Using config(s): {sim.name}")
sim.run_simulation(**kwargs)
def run(env, iterations=1, num_processes=1, dump=False, name="test", **kwargs):
return Simulation(model=env, iterations=iterations, name=name, dump=dump, num_processes=num_processes, **kwargs).run()

106
soil/stats.py
View File

@ -1,106 +0,0 @@
import pandas as pd
from collections import Counter
class Stats:
'''
Interface for all stats. It is not necessary, but it is useful
if you don't plan to implement all the methods.
'''
def __init__(self, simulation):
self.simulation = simulation
def start(self):
'''Method to call when the simulation starts'''
pass
def end(self):
'''Method to call when the simulation ends'''
return {}
def trial(self, env):
'''Method to call when a trial ends'''
return {}
class distribution(Stats):
'''
Calculate the distribution of agent states at the end of each trial,
the mean value, and its deviation.
'''
def start(self):
self.means = []
self.counts = []
def trial(self, env):
df = env[None, None, None].df()
df = df.drop('SEED', axis=1)
ix = df.index[-1]
attrs = df.columns.get_level_values(0)
vc = {}
stats = {
'mean': {},
'count': {},
}
for a in attrs:
t = df.loc[(ix, a)]
try:
stats['mean'][a] = t.mean()
self.means.append(('mean', a, t.mean()))
except TypeError:
pass
for name, count in t.value_counts().items():
if a not in stats['count']:
stats['count'][a] = {}
stats['count'][a][name] = count
self.counts.append(('count', a, name, count))
return stats
def end(self):
dfm = pd.DataFrame(self.means, columns=['metric', 'key', 'value'])
dfc = pd.DataFrame(self.counts, columns=['metric', 'key', 'value', 'count'])
count = {}
mean = {}
if self.means:
res = dfm.drop('metric', axis=1).groupby(by=['key']).agg(['mean', 'std', 'count', 'median', 'max', 'min'])
mean = res['value'].to_dict()
if self.counts:
res = dfc.drop('metric', axis=1).groupby(by=['key', 'value']).agg(['mean', 'std', 'count', 'median', 'max', 'min'])
for k,v in res['count'].to_dict().items():
if k not in count:
count[k] = {}
for tup, times in v.items():
subkey, subcount = tup
if subkey not in count[k]:
count[k][subkey] = {}
count[k][subkey][subcount] = times
return {'count': count, 'mean': mean}
class defaultStats(Stats):
def trial(self, env):
c = Counter()
c.update(a.__class__.__name__ for a in env.network_agents)
c2 = Counter()
c2.update(a['id'] for a in env.network_agents)
return {
'network ': {
'n_nodes': env.G.number_of_nodes(),
'n_edges': env.G.number_of_edges(),
},
'agents': {
'model_count': dict(c),
'state_count': dict(c2),
}
}

185
soil/time.py
View File

@ -1,12 +1,22 @@
from mesa.time import BaseScheduler
from queue import Empty
from heapq import heappush, heappop
from heapq import heappush, heappop, heapreplace
import math
from inspect import getsource
from numbers import Number
from textwrap import dedent
from .utils import logger
from mesa import Agent
from mesa import Agent as MesaAgent
INFINITY = float('inf')
INFINITY = float("inf")
class DeadAgent(Exception):
pass
class When:
def __init__(self, time):
@ -17,6 +27,10 @@ class When:
def abs(self, time):
return self._time
def schedule_next(self, time, delta, first=False):
return (self._time, None)
NEVER = When(INFINITY)
@ -24,11 +38,53 @@ class Delta(When):
def __init__(self, delta):
self._delta = delta
def __eq__(self, other):
return self._delta == other._delta
def abs(self, time):
return time + self._delta
return self._time + self._delta
def __eq__(self, other):
if isinstance(other, Delta):
return self._delta == other._delta
return False
def schedule_next(self, time, delta, first=False):
return (time + self._delta, None)
def __repr__(self):
return str(f"Delta({self._delta})")
class BaseCond:
def __init__(self, msg=None, delta=None, eager=False):
self._msg = msg
self._delta = delta
self.eager = eager
def schedule_next(self, time, delta, first=False):
if first and self.eager:
return (time, self)
if self._delta:
delta = self._delta
return (time + delta, self)
def return_value(self, agent):
return None
def __repr__(self):
return self._msg or self.__class__.__name__
class Cond(BaseCond):
def __init__(self, func, *args, **kwargs):
self._func = func
super().__init__(*args, **kwargs)
def ready(self, agent, time):
return self._func(agent)
def __repr__(self):
if self._msg:
return self._msg
return str(f'Cond("{dedent(getsource(self._func)).strip()}")')
class TimedActivation(BaseScheduler):
@ -36,45 +92,122 @@ class TimedActivation(BaseScheduler):
In each activation, each agent will update its 'next_time'.
"""
def __init__(self, *args, **kwargs):
super().__init__(self)
def __init__(self, *args, shuffle=True, **kwargs):
super().__init__(*args, **kwargs)
self._next = {}
self._queue = []
self.next_time = 0
self._shuffle = shuffle
# self.step_interval = getattr(self.model, "interval", 1)
self.step_interval = self.model.interval
self.logger = getattr(self.model, "logger", logger).getChild(f"time_{ self.model }")
self.next_time = self.time
def add(self, agent: Agent):
if agent.unique_id not in self._agents:
heappush(self._queue, (self.time, agent.unique_id))
def add(self, agent: MesaAgent, when=None):
if when is None:
when = self.time
elif isinstance(when, When):
when = when.abs()
self._schedule(agent, None, when)
super().add(agent)
def _schedule(self, agent, condition=None, when=None, replace=False):
if condition:
if not when:
when, condition = condition.schedule_next(
when or self.time, self.step_interval
)
else:
if when is None:
when = self.time + self.step_interval
condition = None
if self._shuffle:
key = (when, self.model.random.random(), condition)
else:
key = (when, agent.unique_id, condition)
self._next[agent.unique_id] = key
if replace:
heapreplace(self._queue, (key, agent))
else:
heappush(self._queue, (key, agent))
def step(self) -> None:
"""
Executes agents in order, one at a time. After each step,
an agent will signal when it wants to be scheduled next.
"""
if self.next_time == INFINITY:
self.logger.debug(f"Simulation step {self.time}")
if not self.model.running or self.time == INFINITY:
return
self.time = self.next_time
when = self.time
self.logger.debug(f"Queue length: %s", len(self._queue))
while self._queue and self._queue[0][0] == self.time:
(when, agent_id) = heappop(self._queue)
logger.debug(f'Stepping agent {agent_id}')
while self._queue:
((when, _id, cond), agent) = self._queue[0]
if when > self.time:
break
returned = self._agents[agent_id].step()
when = (returned or Delta(1)).abs(self.time)
if when < self.time:
raise Exception("Cannot schedule an agent for a time in the past ({} < {})".format(when, self.time))
if cond:
if not cond.ready(agent, self.time):
self._schedule(agent, cond, replace=True)
continue
try:
agent._last_return = cond.return_value(agent)
except Exception as ex:
agent._last_except = ex
else:
agent._last_return = None
agent._last_except = None
heappush(self._queue, (when, agent_id))
self.logger.debug("Stepping agent %s", agent)
self._next.pop(agent.unique_id, None)
try:
returned = agent.step()
except DeadAgent:
agent.alive = False
heappop(self._queue)
continue
# Check status for MESA agents
if not getattr(agent, "alive", True):
heappop(self._queue)
continue
if returned:
next_check = returned.schedule_next(
self.time, self.step_interval, first=True
)
self._schedule(agent, when=next_check[0], condition=next_check[1], replace=True)
else:
next_check = (self.time + self.step_interval, None)
self._schedule(agent, replace=True)
self.steps += 1
if not self._queue:
self.model.running = False
self.time = INFINITY
self.next_time = INFINITY
return
self.next_time = self._queue[0][0]
next_time = self._queue[0][0][0]
if next_time < self.time:
raise Exception(
f"An agent has been scheduled for a time in the past, there is probably an error ({when} < {self.time})"
)
self.logger.debug("Updating time step: %s -> %s ", self.time, next_time)
self.time = next_time
class ShuffledTimedActivation(TimedActivation):
def __init__(self, *args, **kwargs):
super().__init__(*args, shuffle=True, **kwargs)
class OrderedTimedActivation(TimedActivation):
def __init__(self, *args, **kwargs):
super().__init__(*args, shuffle=False, **kwargs)

127
soil/utils.py
View File

@ -1,71 +1,106 @@
import logging
import time
from time import time as current_time, strftime, gmtime, localtime
import os
import traceback
from shutil import copyfile
from functools import partial
from shutil import copyfile, move
from multiprocessing import Pool, cpu_count
from contextlib import contextmanager
logger = logging.getLogger('soil')
# logging.basicConfig()
# logger.setLevel(logging.INFO)
logger = logging.getLogger("soil")
logger.setLevel(logging.WARNING)
timeformat = "%H:%M:%S"
if os.environ.get("SOIL_VERBOSE", ""):
logformat = "[%(levelname)-5.5s][%(asctime)s][%(name)s]: %(message)s"
else:
logformat = "[%(levelname)-5.5s][%(asctime)s] %(message)s"
logFormatter = logging.Formatter(logformat, timeformat)
consoleHandler = logging.StreamHandler()
consoleHandler.setFormatter(logFormatter)
logging.basicConfig(
level=logging.INFO,
handlers=[
consoleHandler,
],
)
@contextmanager
def timer(name='task', pre="", function=logger.info, to_object=None):
start = time.time()
function('{}Starting {} at {}.'.format(pre, name,
time.strftime("%X", time.gmtime(start))))
def timer(name="task", pre="", function=logger.info, to_object=None):
start = current_time()
function("{}Starting {} at {}.".format(pre, name, strftime("%X", gmtime(start))))
yield start
end = time.time()
function('{}Finished {} at {} in {} seconds'.format(pre, name,
time.strftime("%X", time.gmtime(end)),
str(end-start)))
end = current_time()
function(
"{}Finished {} at {} in {} seconds".format(
pre, name, strftime("%X", gmtime(end)), str(end - start)
)
)
if to_object:
to_object.start = start
to_object.end = end
def safe_open(path, mode='r', backup=True, **kwargs):
def try_backup(path, remove=False):
if not os.path.exists(path):
return None
outdir = os.path.dirname(path)
if outdir and not os.path.exists(outdir):
os.makedirs(outdir)
if backup and 'w' in mode and os.path.exists(path):
creation = os.path.getctime(path)
stamp = time.strftime('%Y-%m-%d_%H.%M.%S', time.localtime(creation))
stamp = strftime("%Y-%m-%d_%H.%M.%S", localtime(creation))
backup_dir = os.path.join(outdir, 'backup')
backup_dir = os.path.join(outdir, "backup")
if not os.path.exists(backup_dir):
os.makedirs(backup_dir)
newpath = os.path.join(backup_dir, '{}@{}'.format(os.path.basename(path),
stamp))
newpath = os.path.join(backup_dir, "{}@{}".format(os.path.basename(path), stamp))
if remove:
move(path, newpath)
else:
copyfile(path, newpath)
return newpath
def safe_open(path, mode="r", backup=True, **kwargs):
outdir = os.path.dirname(path)
if outdir and not os.path.exists(outdir):
os.makedirs(outdir)
if backup and "w" in mode:
try_backup(path)
return open(path, mode=mode, **kwargs)
@contextmanager
def open_or_reuse(f, *args, **kwargs):
try:
return safe_open(f, *args, **kwargs)
except (AttributeError, TypeError):
return f
with safe_open(f, *args, **kwargs) as f:
yield f
except (AttributeError, TypeError) as ex:
yield f
def flatten_dict(d):
if not isinstance(d, dict):
return d
return dict(_flatten_dict(d))
def _flatten_dict(d, prefix=''):
def _flatten_dict(d, prefix=""):
if not isinstance(d, dict):
# print('END:', prefix, d)
yield prefix, d
return
if prefix:
prefix = prefix + '.'
prefix = prefix + "."
for k, v in d.items():
# print(k, v)
res = list(_flatten_dict(v, prefix='{}{}'.format(prefix, k)))
res = list(_flatten_dict(v, prefix="{}{}".format(prefix, k)))
# print('RES:', res)
yield from res
@ -77,7 +112,7 @@ def unflatten_dict(d):
if not isinstance(k, str):
target[k] = v
continue
tokens = k.split('.')
tokens = k.split(".")
if len(tokens) < 2:
target[k] = v
continue
@ -87,3 +122,39 @@ def unflatten_dict(d):
target = target[token]
target[tokens[-1]] = v
return out
def run_and_return_exceptions(func, *args, **kwargs):
"""
A wrapper for a function that catches exceptions and returns them.
It is meant for async simulations.
"""
try:
return func(*args, **kwargs)
except Exception as ex:
if ex.__cause__ is not None:
ex = ex.__cause__
ex.message = "".join(
traceback.format_exception(type(ex), ex, ex.__traceback__)[:]
)
return ex
def run_parallel(func, iterable, num_processes=1, **kwargs):
if num_processes > 1 and not os.environ.get("SOIL_DEBUG", None):
if num_processes < 1:
num_processes = cpu_count() - num_processes
p = Pool(processes=num_processes)
wrapped_func = partial(run_and_return_exceptions, func, **kwargs)
for i in p.imap_unordered(wrapped_func, iterable):
if isinstance(i, Exception):
logger.error("Trial failed:\n\t%s", i.message)
continue
yield i
else:
for i in iterable:
yield func(i, **kwargs)
def int_seed(seed: str):
return int.from_bytes(seed.encode(), "little")

9
soil/version.py
View File

@ -4,7 +4,7 @@ import logging
logger = logging.getLogger(__name__)
ROOT = os.path.dirname(__file__)
DEFAULT_FILE = os.path.join(ROOT, 'VERSION')
DEFAULT_FILE = os.path.join(ROOT, "VERSION")
def read_version(versionfile=DEFAULT_FILE):
@ -12,9 +12,10 @@ def read_version(versionfile=DEFAULT_FILE):
with open(versionfile) as f:
return f.read().strip()
except IOError: # pragma: no cover
logger.error(('Running an unknown version of {}.'
'Be careful!.').format(__name__))
return '0.0'
logger.error(
("Running an unknown version of {}." "Be careful!.").format(__name__)
)
return "0.0"
__version__ = read_version()

5
soil/visualization.py
View File

@ -1,5 +0,0 @@
from mesa.visualization.UserParam import UserSettableParameter
class UserSettableParameter(UserSettableParameter):
def __str__(self):
return self.value

264
soil/web/__init__.py
View File

@ -20,6 +20,7 @@ from tornado.concurrent import run_on_executor
from concurrent.futures import ThreadPoolExecutor
from ..simulation import Simulation
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
@ -31,140 +32,183 @@ LOGGING_INTERVAL = 0.5
# Workaround to let Soil load the required modules
sys.path.append(ROOT)
class PageHandler(tornado.web.RequestHandler):
""" Handler for the HTML template which holds the visualization. """
"""Handler for the HTML template which holds the visualization."""
def get(self):
self.render('index.html', port=self.application.port,
name=self.application.name)
self.render(
"index.html", port=self.application.port, name=self.application.name
)
class SocketHandler(tornado.websocket.WebSocketHandler):
""" Handler for websocket. """
"""Handler for websocket."""
executor = ThreadPoolExecutor(max_workers=MAX_WORKERS)
def open(self):
if self.application.verbose:
logger.info('Socket opened!')
logger.info("Socket opened!")
def check_origin(self, origin):
return True
def on_message(self, message):
""" Receiving a message from the websocket, parse, and act accordingly. """
"""Receiving a message from the websocket, parse, and act accordingly."""
msg = tornado.escape.json_decode(message)
if msg['type'] == 'config_file':
if msg["type"] == "config_file":
if self.application.verbose:
print(msg['data'])
print(msg["data"])
self.config = list(yaml.load_all(msg['data']))
self.config = list(yaml.load_all(msg["data"]))
if len(self.config) > 1:
error = 'Please, provide only one configuration.'
error = "Please, provide only one configuration."
if self.application.verbose:
logger.error(error)
self.write_message({'type': 'error',
'error': error})
self.write_message({"type": "error", "error": error})
return
self.config = self.config[0]
self.send_log('INFO.' + self.simulation_name,
'Using config: {name}'.format(name=self.config['name']))
self.send_log(
"INFO." + self.simulation_name,
"Using config: {name}".format(name=self.config["name"]),
)
if 'visualization_params' in self.config:
self.write_message({'type': 'visualization_params',
'data': self.config['visualization_params']})
self.name = self.config['name']
if "visualization_params" in self.config:
self.write_message(
{
"type": "visualization_params",
"data": self.config["visualization_params"],
}
)
self.name = self.config["name"]
self.run_simulation()
settings = []
for key in self.config['environment_params']:
if type(self.config['environment_params'][key]) == float or type(self.config['environment_params'][key]) == int:
if self.config['environment_params'][key] <= 1:
setting_type = 'number'
for key in self.config["environment_params"]:
if (
type(self.config["environment_params"][key]) == float
or type(self.config["environment_params"][key]) == int
):
if self.config["environment_params"][key] <= 1:
setting_type = "number"
else:
setting_type = 'great_number'
elif type(self.config['environment_params'][key]) == bool:
setting_type = 'boolean'
setting_type = "great_number"
elif type(self.config["environment_params"][key]) == bool:
setting_type = "boolean"
else:
setting_type = 'undefined'
setting_type = "undefined"
settings.append({
'label': key,
'type': setting_type,
'value': self.config['environment_params'][key]
})
settings.append(
{
"label": key,
"type": setting_type,
"value": self.config["environment_params"][key],
}
)
self.write_message({'type': 'settings',
'data': settings})
self.write_message({"type": "settings", "data": settings})
elif msg['type'] == 'get_trial':
elif msg["type"] == "get_trial":
if self.application.verbose:
logger.info('Trial {} requested!'.format(msg['data']))
self.send_log('INFO.' + __name__, 'Trial {} requested!'.format(msg['data']))
self.write_message({'type': 'get_trial',
'data': self.get_trial(int(msg['data']))})
logger.info("Trial {} requested!".format(msg["data"]))
self.send_log("INFO." + __name__, "Trial {} requested!".format(msg["data"]))
self.write_message(
{"type": "get_trial", "data": self.get_trial(int(msg["data"]))}
)
elif msg['type'] == 'run_simulation':
elif msg["type"] == "run_simulation":
if self.application.verbose:
logger.info('Running new simulation for {name}'.format(name=self.config['name']))
self.send_log('INFO.' + self.simulation_name, 'Running new simulation for {name}'.format(name=self.config['name']))
self.config['environment_params'] = msg['data']
logger.info(
"Running new simulation for {name}".format(name=self.config["name"])
)
self.send_log(
"INFO." + self.simulation_name,
"Running new simulation for {name}".format(name=self.config["name"]),
)
self.config["environment_params"] = msg["data"]
self.run_simulation()
elif msg['type'] == 'download_gexf':
G = self.trials[ int(msg['data']) ].history_to_graph()
elif msg["type"] == "download_gexf":
G = self.trials[int(msg["data"])].history_to_graph()
for node in G.nodes():
if 'pos' in G.nodes[node]:
G.nodes[node]['viz'] = {"position": {"x": G.nodes[node]['pos'][0], "y": G.nodes[node]['pos'][1], "z": 0.0}}
del (G.nodes[node]['pos'])
writer = nx.readwrite.gexf.GEXFWriter(version='1.2draft')
if "pos" in G.nodes[node]:
G.nodes[node]["viz"] = {
"position": {
"x": G.nodes[node]["pos"][0],
"y": G.nodes[node]["pos"][1],
"z": 0.0,
}
}
del G.nodes[node]["pos"]
writer = nx.readwrite.gexf.GEXFWriter(version="1.2draft")
writer.add_graph(G)
self.write_message({'type': 'download_gexf',
'filename': self.config['name'] + '_trial_' + str(msg['data']),
'data': tostring(writer.xml).decode(writer.encoding) })
self.write_message(
{
"type": "download_gexf",
"filename": self.config["name"] + "_trial_" + str(msg["data"]),
"data": tostring(writer.xml).decode(writer.encoding),
}
)
elif msg['type'] == 'download_json':
G = self.trials[ int(msg['data']) ].history_to_graph()
elif msg["type"] == "download_json":
G = self.trials[int(msg["data"])].history_to_graph()
for node in G.nodes():
if 'pos' in G.nodes[node]:
G.nodes[node]['viz'] = {"position": {"x": G.nodes[node]['pos'][0], "y": G.nodes[node]['pos'][1], "z": 0.0}}
del (G.nodes[node]['pos'])
self.write_message({'type': 'download_json',
'filename': self.config['name'] + '_trial_' + str(msg['data']),
'data': nx.node_link_data(G) })
if "pos" in G.nodes[node]:
G.nodes[node]["viz"] = {
"position": {
"x": G.nodes[node]["pos"][0],
"y": G.nodes[node]["pos"][1],
"z": 0.0,
}
}
del G.nodes[node]["pos"]
self.write_message(
{
"type": "download_json",
"filename": self.config["name"] + "_trial_" + str(msg["data"]),
"data": nx.node_link_data(G),
}
)
else:
if self.application.verbose:
logger.info('Unexpected message!')
logger.info("Unexpected message!")
def update_logging(self):
try:
if (not self.log_capture_string.closed and self.log_capture_string.getvalue()):
for i in range(len(self.log_capture_string.getvalue().split('\n')) - 1):
self.send_log('INFO.' + self.simulation_name, self.log_capture_string.getvalue().split('\n')[i])
if (
not self.log_capture_string.closed
and self.log_capture_string.getvalue()
):
for i in range(len(self.log_capture_string.getvalue().split("\n")) - 1):
self.send_log(
"INFO." + self.simulation_name,
self.log_capture_string.getvalue().split("\n")[i],
)
self.log_capture_string.truncate(0)
self.log_capture_string.seek(0)
finally:
if self.capture_logging:
tornado.ioloop.IOLoop.current().call_later(LOGGING_INTERVAL, self.update_logging)
tornado.ioloop.IOLoop.current().call_later(
LOGGING_INTERVAL, self.update_logging
)
def on_close(self):
if self.application.verbose:
logger.info('Socket closed!')
logger.info("Socket closed!")
def send_log(self, logger, logging):
self.write_message({'type': 'log',
'logger': logger,
'logging': logging})
self.write_message({"type": "log", "logger": logger, "logging": logging})
@property
def simulation_name(self):
return self.config.get('name', 'NoSimulationRunning')
return self.config.get("name", "NoSimulationRunning")
@run_on_executor
def nonblocking(self, config):
@ -174,28 +218,31 @@ class SocketHandler(tornado.websocket.WebSocketHandler):
@tornado.gen.coroutine
def run_simulation(self):
# Run simulation and capture logs
logger.info('Running simulation!')
if 'visualization_params' in self.config:
del self.config['visualization_params']
logger.info("Running simulation!")
if "visualization_params" in self.config:
del self.config["visualization_params"]
with self.logging(self.simulation_name):
try:
config = dict(**self.config)
config['outdir'] = os.path.join(self.application.outdir, config['name'])
config['dump'] = self.application.dump
config["outdir"] = os.path.join(self.application.outdir, config["name"])
config["dump"] = self.application.dump
self.trials = yield self.nonblocking(config)
self.write_message({'type': 'trials',
'data': list(trial.name for trial in self.trials) })
self.write_message(
{
"type": "trials",
"data": list(trial.name for trial in self.trials),
}
)
except Exception as ex:
error = 'Something went wrong:\n\t{}'.format(ex)
error = "Something went wrong:\n\t{}".format(ex)
logging.info(error)
self.write_message({'type': 'error',
'error': error})
self.send_log('ERROR.' + self.simulation_name, error)
self.write_message({"type": "error", "error": error})
self.send_log("ERROR." + self.simulation_name, error)
def get_trial(self, trial):
logger.info('Available trials: %s ' % len(self.trials))
logger.info('Ask for : %s' % trial)
logger.info("Available trials: %s " % len(self.trials))
logger.info("Ask for : %s" % trial)
trial = self.trials[trial]
G = trial.history_to_graph()
return nx.node_link_data(G)
@ -218,21 +265,24 @@ class SocketHandler(tornado.websocket.WebSocketHandler):
class ModularServer(tornado.web.Application):
""" Main visualization application. """
"""Main visualization application."""
port = 8001
page_handler = (r'/', PageHandler)
socket_handler = (r'/ws', SocketHandler)
static_handler = (r'/(.*)', tornado.web.StaticFileHandler,
{'path': os.path.join(ROOT, 'static')})
local_handler = (r'/local/(.*)', tornado.web.StaticFileHandler,
{'path': ''})
page_handler = (r"/", PageHandler)
socket_handler = (r"/ws", SocketHandler)
static_handler = (
r"/(.*)",
tornado.web.StaticFileHandler,
{"path": os.path.join(ROOT, "static")},
)
local_handler = (r"/local/(.*)", tornado.web.StaticFileHandler, {"path": ""})
handlers = [page_handler, socket_handler, static_handler, local_handler]
settings = {'debug': True,
'template_path': ROOT + '/templates'}
settings = {"debug": True, "template_path": ROOT + "/templates"}
def __init__(self, dump=False, outdir='output', name='SOIL', verbose=True, *args, **kwargs):
def __init__(
self, dump=False, outdir="output", name="SOIL", verbose=True, *args, **kwargs
):
self.verbose = verbose
self.name = name
@ -243,12 +293,12 @@ class ModularServer(tornado.web.Application):
super().__init__(self.handlers, **self.settings)
def launch(self, port=None):
""" Run the app. """
"""Run the app."""
if port is not None:
self.port = port
url = 'http://127.0.0.1:{PORT}'.format(PORT=self.port)
print('Interface starting at {url}'.format(url=url))
url = "http://127.0.0.1:{PORT}".format(PORT=self.port)
print("Interface starting at {url}".format(url=url))
self.listen(self.port)
# webbrowser.open(url)
tornado.ioloop.IOLoop.instance().start()
@ -263,12 +313,22 @@ def run(*args, **kwargs):
def main():
import argparse
parser = argparse.ArgumentParser(description='Visualization of a Graph Model')
parser = argparse.ArgumentParser(description="Visualization of a Graph Model")
parser.add_argument('--name', '-n', nargs=1, default='SOIL', help='name of the simulation')
parser.add_argument('--dump', '-d', help='dumping results in folder output', action='store_true')
parser.add_argument('--port', '-p', nargs=1, default=8001, help='port for launching the server')
parser.add_argument('--verbose', '-v', help='verbose mode', action='store_true')
parser.add_argument(
"--name", "-n", nargs=1, default="SOIL", help="name of the simulation"
)
parser.add_argument(
"--dump", "-d", help="dumping results in folder output", action="store_true"
)
parser.add_argument(
"--port", "-p", nargs=1, default=8001, help="port for launching the server"
)
parser.add_argument("--verbose", "-v", help="verbose mode", action="store_true")
args = parser.parse_args()
run(name=args.name, port=(args.port[0] if isinstance(args.port, list) else args.port), verbose=args.verbose)
run(
name=args.name,
port=(args.port[0] if isinstance(args.port, list) else args.port),
verbose=args.verbose,
)

4
soil/web/config.yml
View File

@ -6,11 +6,11 @@ network_params:
n: 100
m: 2
network_agents:
- agent_type: ControlModelM2
- agent_class: ControlModelM2
weight: 0.1
state:
id: 1
- agent_type: ControlModelM2
- agent_class: ControlModelM2
weight: 0.9
state:
id: 0

27
soil/web/run.py
View File

@ -4,20 +4,33 @@ from simulator import Simulator
def run(simulator, name="SOIL", port=8001, verbose=False):
server = ModularServer(simulator, name=(name[0] if isinstance(name, list) else name), verbose=verbose)
server = ModularServer(
simulator, name=(name[0] if isinstance(name, list) else name), verbose=verbose
)
server.port = port
server.launch()
if __name__ == "__main__":
parser = argparse.ArgumentParser(description='Visualization of a Graph Model')
parser = argparse.ArgumentParser(description="Visualization of a Graph Model")
parser.add_argument('--name', '-n', nargs=1, default='SOIL', help='name of the simulation')
parser.add_argument('--dump', '-d', help='dumping results in folder output', action='store_true')
parser.add_argument('--port', '-p', nargs=1, default=8001, help='port for launching the server')
parser.add_argument('--verbose', '-v', help='verbose mode', action='store_true')
parser.add_argument(
"--name", "-n", nargs=1, default="SOIL", help="name of the simulation"
)
parser.add_argument(
"--dump", "-d", help="dumping results in folder output", action="store_true"
)
parser.add_argument(
"--port", "-p", nargs=1, default=8001, help="port for launching the server"
)
parser.add_argument("--verbose", "-v", help="verbose mode", action="store_true")
args = parser.parse_args()
soil = Simulator(dump=args.dump)
run(soil, name=args.name, port=(args.port[0] if isinstance(args.port, list) else args.port), verbose=args.verbose)
run(
soil,
name=args.name,
port=(args.port[0] if isinstance(args.port, list) else args.port),
verbose=args.verbose,
)

BIN
soil/web/static/img/background/map.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.1 MiB

BIN
soil/web/static/img/background/map_4800x2860.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 8.0 MiB

Some files were not shown because too many files have changed in this diff Show More