balkian/senpy
1
0
Fork 0
mirror of https://github.com/gsi-upm/senpy synced 2024-12-22 04:58:12 +00:00
Code Issues Releases Wiki Activity

Several fixes and improvements

Browse Source 
* 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).
This commit is contained in:
J. Fernando Sánchez 2018-06-14 15:10:16 +02:00
parent 697e779767
commit 1313853788
6 changed files with 91 additions and 51 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
senpy/__main__.py
View File

@ -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,

38
senpy/models.py
View File

@ -203,24 +203,27 @@ class BaseModel(with_metaclass(BaseMeta, CustomDict)):
context_uri=None,
prefix=None,
expanded=False):
ser = self.serializable()
result = jsonld.compact(
ser,
self._context,
options={
'base': prefix,
'expandContext': self._context,
'senpy': prefix
})
if context_uri:
result['@context'] = context_uri
result = self.serializable()
if context_uri or with_context:
result['@context'] = context_uri or self._context
# result = jsonld.compact(result,
# self._context,
# options={
# 'base': prefix,
# 'expandContext': self._context,
# 'senpy': prefix
# })
if expanded:
result = jsonld.expand(
result, options={'base': prefix,
'expandContext': self._context})
if not with_context:
del result['@context']
try:
del result['@context']
except KeyError:
pass
return result
def validate(self, obj=None):
@ -323,7 +326,10 @@ def _add_class_from_schema(*args, **kwargs):
for i in [
'aggregatedEvaluation',
'analysis',
'dataset',
'datasets',
'emotion',
'emotionConversion',
'emotionConversionPlugin',
@ -331,19 +337,17 @@ for i in [
'emotionModel',
'emotionPlugin',
'emotionSet',
'evaluation',
'entity',
'help',
'metric',
'plugin',
'plugins',
'response',
'results',
'sentimentPlugin',
'suggestion',
'aggregatedEvaluation',
'evaluation',
'metric',
'dataset',
'datasets',
'topic',
]:
_add_class_from_schema(i)

66
senpy/plugins/__init__.py
View File

@ -18,6 +18,7 @@ import subprocess
import importlib
import yaml
import threading
import nltk
from .. import models, utils
from .. import api
@ -95,6 +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 = []
@ -123,9 +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):
@ -148,10 +159,22 @@ class Plugin(with_metaclass(PluginMeta, models.Plugin)):
raise
assert not should_fail
def open(self, fpath, *args, **kwargs):
if not os.path.isabs(fpath):
fpath = os.path.join(self.data_folder, fpath)
return open(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)
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)
@ -186,7 +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:
@ -375,7 +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
@ -402,32 +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))
if ptype:
if isinstance(ptype, PluginMeta):
ptype = ptype.__name__
logger.debug('plugin_type {}'.format(plugin_type))
if plugin_type:
if isinstance(plugin_type, PluginMeta):
plugin_type = plugin_type.__name__
try:
ptype = ptype[0].upper() + ptype[1:]
pclass = globals()[ptype]
plugin_type = plugin_type[0].upper() + plugin_type[1:]
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
@ -462,6 +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:
@ -477,6 +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
@ -573,12 +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

4
senpy/schemas/context.jsonld
View File

@ -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"

30
senpy/utils.py
View File

@ -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),
pprint.pformat(template)))
'\tFound: {}').format(pprint.pformat(template),
pprint.pformat(indict)))
def convert_dictionary(original, mappings):
@ -67,18 +68,23 @@ 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)
if not plugin_list:
import __main__
logger.info('Loading classes from {}'.format(__main__))
from . import plugins
plugin_list = plugins.from_module(__main__)
for plug in plugin_list:
plug.test()
logger.info('The tests for {} passed!'.format(plug.name))
logger.info('All tests passed!')
try:
if not plugin_list:
import __main__
logger.info('Loading classes from {}'.format(__main__))
from . import plugins
plugin_list = plugins.from_module(__main__)
for plug in plugin_list:
plug.test()
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):

2
tests/test_extensions.py
View File

@ -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)