def get(self, doc_id):
''' Given a doc id, accesses wp entities and then cross-references entity parses
:param doc_id: the id of the article
'''
entities = WpEntitiesService().nestedGet(doc_id)
coreferences = CoreferenceCountsService().get(doc_id).get(doc_id, {})
exists = lambda x: x is not None
docParaphrases = coreferences.get('paraphrases', {})
coref_mention_keys = map(title_confirmation.preprocess, docParaphrases.keys())
coref_mention_values = map(title_confirmation.preprocess, [item for sublist in docParaphrases.values() for item in sublist])
paraphrases = dict([(title_confirmation.preprocess(item[0]), map(title_confirmation.preprocess, item[1]))\
for item in docParaphrases.items()])
counts ={}
for val in map(title_confirmation.preprocess, entities):
try:
canonical = val
if canonical in coref_mention_keys:
counts[canonical] = len(paraphrases[canonical])
elif canonical != val and val in coref_mention_keys:
counts[canonical] = len(paraphrases[val])
elif canonical in coref_mention_values:
counts[canonical] = len(filter(lambda x: canonical in x[1], paraphrases.items())[0][1])
elif canonical != val and val in coref_mention_values:
counts[canonical] = len(filter(lambda x: val in x[1], paraphrases.items())[0][1])
except:
pass
return {doc_id: counts, 'status': 200}
def get(self, doc_id, entity):
parse = ParsedJsonService().nestedGet(doc_id)
if parse.get('root', {}).get('document', {}).get('sentences', {}) is None:
return {'status': 200, entity: []}
sents = asList(parse.get('root', {}).get('document', {}).get('sentences', {}).get('sentence', []))
corefs = asList(parse.get('root', {}).get('document', {}).get('coreference', {}).get('coreference', []))
sents_processed = [' '.join([token['word'] for token in asList(sent['tokens']['token'])]) for sent in sents]
sents_preprocessed = [' '.join([title_confirmation.preprocess(token['word']) for token in asList(sent['tokens']['token'])]) for sent in sents]
coreferences = []
sentences_to_add = []
for coref in corefs:
dont_add = True
for m in asList(coref['mention']):
try:
mention_string = ' '.join([title_confirmation.preprocess(a['word']) for a in sents[int(m['sentence'])-1]['tokens']['token'][int(m['start'])-1:int(m['end'])-1]])
if title_confirmation.preprocess(entity) == mention_string:
sentences_to_add += [int(m['sentence'])-1 for m in asList(coref['mention'])]
break
except TypeError:
pass # f**k it
sentences_to_add = list(set(sentences_to_add + [i for i in range(0, len(sents)) if entity in sents_preprocessed[i]]))
return {'status': 200, entity: [{'sentiment': sents[i].get('@sentiment', None), 'sentence': sents_processed[i]} for i in range(0, len(sents)) if i in sentences_to_add]}
def get(self, doc_id):
jsonResponse = ParsedJsonService().get(doc_id)
if jsonResponse['status'] != 200:
return jsonResponse
dict = jsonResponse[doc_id]
counter = 0
if not isEmptyDoc(dict):
return {'status':200,
doc_id: [title_confirmation.preprocess(MrgSentence(sentence.get('parse', '')).nodes.getTermHead().getString()) \
for sentence in asList(dict.get('root', {}).get('document', {}).get('sentences', {}).get('sentence', [])) \
]
}
else:
return {'status':400,
'message': "No sentences found"}