Several fixes and improvements

* Add Topic model
* Add PDB post-mortem debugging
* Add logger to plugins (`self.log`)
* Add NLTK resource auto-download
* Force installation of requirements even if adding doesn't work
* Add a method to find files in several possible locations. Now the plugin.open
method will try these locations IF the file is to be opened in read mode.
Otherwise only the SENPY_DATA folder will be used (to avoid writing to the
package folder).

.makefiles/python.mk

@@ -77,7 +77,6 @@ push-latest: $(addprefix push-latest-,$(PYVERSIONS)) ## Push the "latest" tag to
 	docker tag '$(IMAGEWTAG)-python$(PYMAIN)' '$(IMAGEWTAG)'
 	docker tag '$(IMAGEWTAG)-python$(PYMAIN)' '$(IMAGENAME)'
 	docker push '$(IMAGENAME):latest'
-	docker push '$(IMAGEWTAG)'
 
 push-latest-%: build-%  ## Push the latest image for a specific python version
 	docker tag $(IMAGENAME):$(VERSION)-python$* $(IMAGENAME):python$*

Dockerfile.template

@@ -6,8 +6,6 @@ RUN apt-get update && apt-get install -y \
 libblas-dev liblapack-dev liblapacke-dev gfortran \
  && rm -rf /var/lib/apt/lists/*
 
-RUN pip install --no-cache-dir --upgrade numpy scipy scikit-learn
-
 RUN mkdir /cache/ /senpy-plugins /data/
 
 VOLUME /data/

MANIFEST.in

@@ -1,5 +1,6 @@
 include requirements.txt
 include test-requirements.txt
+include extra-requirements.txt
 include README.rst
 include senpy/VERSION
 graft senpy/plugins

requirements.txt

@@ -9,3 +9,6 @@ jsonref
 PyYAML
 rdflib
 rdflib-jsonld
+numpy
+scipy
+scikit-learn

senpy/__main__.py

@@ -130,7 +130,7 @@ def main():
         return
     sp.activate_all()
     if args.only_test:
-        easy_test(sp.plugins())
+        easy_test(sp.plugins(), debug=args.debug)
         return
     print('Senpy version {}'.format(senpy.__version__))
     print('Server running on port %s:%d. Ctrl+C to quit' % (args.host,

senpy/api.py

@@ -147,7 +147,7 @@ def parse_params(indict, *specs):
         for param, options in iteritems(spec):
             for alias in options.get("aliases", []):
                 # Replace each alias with the correct name of the parameter
-                if alias in indict and alias is not param:
+                if alias in indict and alias != param:
                     outdict[param] = indict[alias]
                     del outdict[alias]
                     continue

senpy/blueprints.py

@@ -19,7 +19,7 @@ Blueprints for Senpy
 """
 from flask import (Blueprint, request, current_app, render_template, url_for,
                    jsonify)
-from .models import Error, Response, Help, Plugins, read_schema, Datasets
+from .models import Error, Response, Help, Plugins, read_schema, dump_schema, Datasets
 from . import api
 from .version import __version__
 from functools import wraps
@@ -67,9 +67,9 @@ def index():
 @api_blueprint.route('/schemas/<schema>')
 def schema(schema="definitions"):
     try:
-        return jsonify(read_schema(schema))
+        return dump_schema(read_schema(schema))
-    except Exception:  # Should be FileNotFoundError, but it's missing from py2
+    except Exception as ex:  # Should be FileNotFoundError, but it's missing from py2
-        return Error(message="Schema not found", status=404).flask()
+        return Error(message="Schema not found: {}".format(ex), status=404).flask()
 
 
 def basic_api(f):
@@ -133,6 +133,7 @@ def api_root():
     req = api.parse_call(request.parameters)
     return current_app.senpy.analyse(req)
 
+
 @api_blueprint.route('/evaluate/', methods=['POST', 'GET'])
 @basic_api
 def evaluate():
@@ -145,6 +146,7 @@ def evaluate():
         response = current_app.senpy.evaluate(params)
         return response
 
+
 @api_blueprint.route('/plugins/', methods=['POST', 'GET'])
 @basic_api
 def plugins():
@@ -163,10 +165,10 @@ def plugin(plugin=None):
     return sp.get_plugin(plugin)
 
 
-@api_blueprint.route('/datasets/', methods=['POST','GET'])
+@api_blueprint.route('/datasets/', methods=['POST', 'GET'])
 @basic_api
 def datasets():
     sp = current_app.senpy
     datasets = sp.datasets
-    dic = Datasets(datasets = list(datasets.values()))
+    dic = Datasets(datasets=list(datasets.values()))
     return dic

senpy/models.py

@@ -51,6 +51,10 @@ def read_schema(schema_file, absolute=False):
         return jsonref.load(f, base_uri=schema_uri)
 
 
+def dump_schema(schema):
+    return jsonref.dumps(schema)
+
+
 def load_context(context):
     logging.debug('Loading context: {}'.format(context))
     if not context:
@@ -199,24 +203,27 @@ class BaseModel(with_metaclass(BaseMeta, CustomDict)):
                context_uri=None,
                prefix=None,
                expanded=False):
-        ser = self.serializable()
 
-        result = jsonld.compact(
+        result = self.serializable()
-            ser,
+        if context_uri or with_context:
-            self._context,
+            result['@context'] = context_uri or self._context
-            options={
+
-                'base': prefix,
+        # result = jsonld.compact(result,
-                'expandContext': self._context,
+        #                         self._context,
-                'senpy': prefix
+        #                         options={
-            })
+        #                             'base': prefix,
-        if context_uri:
+        #                             'expandContext': self._context,
-            result['@context'] = context_uri
+        #                             'senpy': prefix
+        #                         })
         if expanded:
             result = jsonld.expand(
                 result, options={'base': prefix,
                                  'expandContext': self._context})
         if not with_context:
+            try:
                 del result['@context']
+            except KeyError:
+                pass
         return result
 
     def validate(self, obj=None):
@@ -319,7 +326,10 @@ def _add_class_from_schema(*args, **kwargs):
 
 
 for i in [
+        'aggregatedEvaluation',
         'analysis',
+        'dataset',
+        'datasets',
         'emotion',
         'emotionConversion',
         'emotionConversionPlugin',
@@ -327,19 +337,17 @@ for i in [
         'emotionModel',
         'emotionPlugin',
         'emotionSet',
+        'evaluation',
         'entity',
         'help',
+        'metric',
         'plugin',
         'plugins',
         'response',
         'results',
         'sentimentPlugin',
         'suggestion',
-        'aggregatedEvaluation',
+        'topic',
-        'evaluation',
-        'metric',
-        'dataset',
-        'datasets',
 
 ]:
     _add_class_from_schema(i)

senpy/plugins/__init__.py

@@ -18,8 +18,7 @@ import subprocess
 import importlib
 import yaml
 import threading
-
+import nltk
-import numpy as np
 
 from .. import models, utils
 from .. import api
@@ -49,11 +48,11 @@ class PluginMeta(models.BaseMeta):
         attrs['name'] = alias
         if 'description' not in attrs:
             doc = attrs.get('__doc__', None)
-            if not doc:
+            if doc:
-                raise Exception(('Please, add a description or '
-                                 'documentation to class {}').format(name))
                 attrs['description'] = doc
-            attrs['name'] = alias
+            else:
+                logger.warn(('Plugin {} does not have a description. '
+                             'Please, add a short summary to help other developers').format(name))
         cls = super(PluginMeta, mcs).__new__(mcs, name, bases, attrs)
 
         if alias in mcs._classes:
@@ -97,6 +96,16 @@ class Plugin(with_metaclass(PluginMeta, models.Plugin)):
         self.is_activated = False
         self._lock = threading.Lock()
         self.data_folder = data_folder or os.getcwd()
+        self._directory = os.path.abspath(os.path.dirname(inspect.getfile(self.__class__)))
+        self._data_paths = ['',
+                            self._directory,
+                            os.path.join(self._directory, 'data'),
+                            self.data_folder]
+        self._log = logging.getLogger('{}.{}'.format(__name__, self.name))
+
+    @property
+    def log(self):
+        return self._log
 
     def validate(self):
         missing = []
@@ -125,9 +134,9 @@ class Plugin(with_metaclass(PluginMeta, models.Plugin)):
         for case in test_cases:
             try:
                 self.test_case(case)
-                logger.debug('Test case passed:\n{}'.format(pprint.pformat(case)))
+                self.log.debug('Test case passed:\n{}'.format(pprint.pformat(case)))
             except Exception as ex:
-                logger.warn('Test case failed:\n{}'.format(pprint.pformat(case)))
+                self.log.warn('Test case failed:\n{}'.format(pprint.pformat(case)))
                 raise
 
     def test_case(self, case):
@@ -150,10 +159,22 @@ class Plugin(with_metaclass(PluginMeta, models.Plugin)):
             raise
         assert not should_fail
 
-    def open(self, fpath, *args, **kwargs):
+    def find_file(self, fname):
+        for p in self._data_paths:
+            alternative = os.path.join(p, fname)
+            if os.path.exists(alternative):
+                return alternative
+        raise IOError('File does not exist: {}'.format(fname))
+
+    def open(self, fpath, mode='r'):
+        if 'w' in mode:
+            # When writing, only use absolute paths or data_folder
             if not os.path.isabs(fpath):
                 fpath = os.path.join(self.data_folder, fpath)
-        return open(fpath, *args, **kwargs)
+        else:
+            fpath = self.find_file(fpath)
+
+        return open(fpath, mode=mode)
 
     def serve(self, debug=True, **kwargs):
         utils.easy(plugin_list=[self, ], plugin_folder=None, debug=debug, **kwargs)
@@ -188,7 +209,7 @@ class Analysis(Plugin):
 
     def analyse_entries(self, entries, parameters):
         for entry in entries:
-            logger.debug('Analysing entry with plugin {}: {}'.format(self, entry))
+            self.log.debug('Analysing entry with plugin {}: {}'.format(self, entry))
             results = self.analyse_entry(entry, parameters)
             if inspect.isgenerator(results):
                 for result in results:
@@ -291,7 +312,7 @@ class Box(AnalysisPlugin):
         return self
 
     def transform(self, X):
-        return np.array([self.predict_one(x) for x in X])
+        return [self.predict_one(x) for x in X]
 
     def predict(self, X):
         return self.transform(X)
@@ -377,7 +398,7 @@ class ShelfMixin(object):
                     with self.open(self.shelf_file, 'rb') as p:
                         self._sh = pickle.load(p)
                 except (IndexError, EOFError, pickle.UnpicklingError):
-                    logger.warning('{} has a corrupted shelf file!'.format(self.id))
+                    self.log.warning('Corrupted shelf file: {}'.format(self.shelf_file))
                     if not self.get('force_shelf', False):
                         raise
         return self._sh
@@ -404,32 +425,31 @@ class ShelfMixin(object):
         self._shelf_file = value
 
     def save(self):
-        logger.debug('saving pickle')
+        self.log.debug('Saving pickle')
         if hasattr(self, '_sh') and self._sh is not None:
             with self.open(self.shelf_file, 'wb') as f:
                 pickle.dump(self._sh, f)
 
 
-def pfilter(plugins, **kwargs):
+def pfilter(plugins, plugin_type=Analysis, **kwargs):
     """ Filter plugins by different criteria """
     if isinstance(plugins, models.Plugins):
         plugins = plugins.plugins
     elif isinstance(plugins, dict):
         plugins = plugins.values()
-    ptype = kwargs.pop('plugin_type', Plugin)
     logger.debug('#' * 100)
-    logger.debug('ptype {}'.format(ptype))
+    logger.debug('plugin_type {}'.format(plugin_type))
-    if ptype:
+    if plugin_type:
-        if isinstance(ptype, PluginMeta):
+        if isinstance(plugin_type, PluginMeta):
-            ptype = ptype.__name__
+            plugin_type = plugin_type.__name__
         try:
-            ptype = ptype[0].upper() + ptype[1:]
+            plugin_type = plugin_type[0].upper() + plugin_type[1:]
-            pclass = globals()[ptype]
+            pclass = globals()[plugin_type]
             logger.debug('Class: {}'.format(pclass))
             candidates = filter(lambda x: isinstance(x, pclass),
                                 plugins)
         except KeyError:
-            raise models.Error('{} is not a valid type'.format(ptype))
+            raise models.Error('{} is not a valid type'.format(plugin_type))
     else:
         candidates = plugins
 
@@ -464,6 +484,7 @@ def _log_subprocess_output(process):
 
 def install_deps(*plugins):
     installed = False
+    nltk_resources = set()
     for info in plugins:
         requirements = info.get('requirements', [])
         if requirements:
@@ -479,6 +500,9 @@ def install_deps(*plugins):
             installed = True
             if exitcode != 0:
                 raise models.Error("Dependencies not properly installed")
+        nltk_resources |= set(info.get('nltk_resources', []))
+
+    installed |= nltk.download(list(nltk_resources))
     return installed
 
 
@@ -575,12 +599,14 @@ def _instances_in_module(module):
 def _from_module_name(module, root, info=None, install=True, **kwargs):
     try:
         module = load_module(module, root)
-    except ImportError:
+    except (ImportError, LookupError):
         if not install or not info:
             raise
         install_deps(info)
         module = load_module(module, root)
     for plugin in _from_loaded_module(module=module, root=root, info=info, **kwargs):
+        if install:
+            install_deps(plugin)
         yield plugin
 
 

senpy/schemas/context.jsonld

@@ -10,8 +10,10 @@
     "wna": "http://www.gsi.dit.upm.es/ontologies/wnaffect/ns#",
     "emoml": "http://www.gsi.dit.upm.es/ontologies/onyx/vocabularies/emotionml/ns#",
     "xsd": "http://www.w3.org/2001/XMLSchema#",
+    "fam": "http://vocab.fusepool.info/fam#",
     "topics": {
-      "@id": "dc:subject"
+      "@id": "nif:topic",
+      "@container": "@set"
     },
     "entities": {
       "@id": "me:hasEntities"

senpy/schemas/evaluation.json

@@ -1,6 +1,6 @@
 {
   "$schema": "http://json-schema.org/draft-04/schema#",
-  "name": "Evalation",
+  "name": "Evaluation",
   "properties": {
     "@id": {
       "type": "string"

senpy/utils.py

@@ -1,6 +1,7 @@
 from . import models, __version__
 from collections import MutableMapping
 import pprint
+import pdb
 
 import logging
 logger = logging.getLogger(__name__)
@@ -32,8 +33,8 @@ def check_template(indict, template):
         if indict != template:
             raise models.Error(('Differences found.\n'
                                 '\tExpected: {}\n'
-                                '\tFound: {}').format(pprint.pformat(indict),
+                                '\tFound: {}').format(pprint.pformat(template),
-                                                      pprint.pformat(template)))
+                                                      pprint.pformat(indict)))
 
 
 def convert_dictionary(original, mappings):
@@ -67,9 +68,10 @@ def easy_load(app=None, plugin_list=None, plugin_folder=None, **kwargs):
     return sp, app
 
 
-def easy_test(plugin_list=None):
+def easy_test(plugin_list=None, debug=True):
     logger.setLevel(logging.DEBUG)
     logging.getLogger().setLevel(logging.INFO)
+    try:
         if not plugin_list:
             import __main__
             logger.info('Loading classes from {}'.format(__main__))
@@ -77,8 +79,12 @@ def easy_test(plugin_list=None):
             plugin_list = plugins.from_module(__main__)
         for plug in plugin_list:
             plug.test()
-        logger.info('The tests for {} passed!'.format(plug.name))
+            plug.log.info('My tests passed!')
             logger.info('All tests passed!')
+    except Exception:
+        if not debug:
+            raise
+        pdb.post_mortem()
 
 
 def easy(host='0.0.0.0', port=5000, debug=True, **kwargs):

tests/test_api.py

@@ -32,7 +32,7 @@ class APITest(TestCase):
         query = {}
         plug_params = {
             'hello': {
-                'aliases': ['hello', 'hiya'],
+                'aliases': ['hiya', 'hello'],
                 'required': True
             }
         }
@@ -48,6 +48,26 @@ class APITest(TestCase):
         assert 'hello' in p
         assert p['hello'] == 'dlrow'
 
+    def test_parameters2(self):
+        in1 = {
+            'meaningcloud-key': 5
+        }
+        in2 = {
+            'apikey': 25
+        }
+        extra_params = {
+            "apikey": {
+                "aliases": [
+                    "apikey",
+                    "meaningcloud-key"
+                ],
+                "required": True
+            }
+        }
+        p1 = parse_params(in1, extra_params)
+        p2 = parse_params(in2, extra_params)
+        assert (p2['apikey'] / p1['apikey']) == 5
+
     def test_default(self):
         spec = {
             'hello': {

tests/test_extensions.py

@@ -47,7 +47,7 @@ class ExtensionsTest(TestCase):
 
     def test_add_delete(self):
         '''Should be able to add and delete new plugins. '''
-        new = plugins.Plugin(name='new', description='new', version=0)
+        new = plugins.Analysis(name='new', description='new', version=0)
         self.senpy.add_plugin(new)
         assert new in self.senpy.plugins()
         self.senpy.delete_plugin(new)

tests/test_schemas.py

@@ -8,6 +8,8 @@ from fnmatch import fnmatch
 
 from jsonschema import RefResolver, Draft4Validator, ValidationError
 
+from senpy.models import read_schema
+
 root_path = path.join(path.dirname(path.realpath(__file__)), '..')
 schema_folder = path.join(root_path, 'senpy', 'schemas')
 examples_path = path.join(root_path, 'docs', 'examples')
@@ -15,7 +17,8 @@ bad_examples_path = path.join(root_path, 'docs', 'bad-examples')
 
 
 class JSONSchemaTests(unittest.TestCase):
-    pass
+    def test_definitions(self):
+        read_schema('definitions.json')
 
 
 def do_create_(jsfile, success):