123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778 |
- #!/usr/bin/env python
- #encoding:utf-8
- from deepdive import *
- import random
- from collections import namedtuple
- from commonutil import *
- Label = namedtuple('Label', 'entity_id, label, type')
- @tsv_extractor
- @returns(lambda
- entity_id = "text",
- label = "int",
- rule_id = "text",
- :[])
- # heuristic rules for finding positive/negative examples of transaction relationship mentions
- def supervise(
- entity_id="text", entity_begin="int", entity_end="int",
- doc_id="text", sentence_index="int", sentence_text="text",
- tokens="text[]", pos_tags="text[]", ner_tags="text[]",
- ):
- # Constants
- label = Label(entity_id=entity_id,label=None,type=None)
- Bidding = frozenset(["招标","报名","投标"])
- TIME_PREFIX = frozenset(["开始"])
- NOT_TIME_PREFIX = frozenset(["结束","截止"])
- TIME = frozenset(["时间"])
- SpecificTime = frozenset(["年","月"])
- MAX_DIST = 10
- TYPE_MENTION = frozenset(["org","company"])
- # Common data objects
- if entity_begin>MAX_DIST:
- begin = entity_begin-MAX_DIST
- else:
- begin = 0
- if len(tokens)-entity_end>MAX_DIST:
- end = entity_end+MAX_DIST
- else:
- end = -1
- front_tokens = tokens[begin:entity_begin]
- end_tokens = tokens[entity_end:end]
- front_ner = ner_tags[begin:entity_begin]
- end_ner = ner_tags[end:entity_end]
-
- log(sentence_text)
-
- # Rule: Candidates that are too far apart
- '''
- if len(intermediate_lemmas) > MAX_DIST:
- yield transaction._replace(label=-1, type='neg:far_apart')
- # Rule: Candidates that have a third company in between
- if 'company' in intermediate_ner_tags:
- yield transaction._replace(label=-1, type='neg:third_company_between')
- # Rule: Sentences that contain wife/husband in between
- # (<P1>)([ A-Za-z]+)(wife|husband)([ A-Za-z]+)(<P2>)
- #if len(TRANSLATION.intersection(intermediate_lemmas)) > 0 and len(STOCK.intersection(intermediate_lemmas)) > 0:
- # yield transaction._replace(label=1, type='A购买股权B')
- '''
- if len(Bidding.intersection(front_tokens))>0:
- if len(TIME.intersection(front_tokens))>0:
- if len(TIME_PREFIX.intersection(front_tokens))>=0 and len(NOT_TIME_PREFIX.intersection(front_tokens))==0:
- if find_index(list(SpecificTime),"".join(tokens[entity_begin:entity_end+1]))>=0:
- yield label._replace(label=1,type="投标开始时间:time")
- else:
- yield label._replace(label=-1,type="pos:非具体时间")
- else:
- yield label._replace(label=-1,type="pos:截止时间")
-
- else:
- yield label._replace(label=-1,type="pos:no match")
- else:
- yield label._replace(label=-1,type="pos:no match")
|