def process_2_1(sentence, tagged_corpus):
# adding another part for the first time
# if possible, integrate this into the main loop
if sentence["pdt"] is None: # try to re-cast this
possibles = []
for word_index in range(len(tagged_corpus)):
word = tagged_corpus[word_index]
if word[1] == "NNS":
word_processed = (NLP.inflect.singular_noun(word[0]), "NN")
obj = EC.Container(word=word,
word_processed=word_processed,
trd_p=True,
index=word_index)
else:
word_processed = word
obj = EC.Container(word=word,
word_processed=word_processed,
index=word_index)
if word_processed[0] in NLP.wordtags and NLP.wordtags[
word_processed[0]].index("VERB") < 3:
possibles.append(obj)
for possibility in possibles:
tagged_alternative = copy.deepcopy(tagged_corpus)
tag = "VB"
if hasattr(possibility, "trd_p"):
tag = "VBZ"
tagged_alternative[possibility.index] = (possibility.word[0], tag)
if possibility.index == 0:
continue
elif not utils.deep_in(tagged_alternative[:possibility.index],
"n",
key=NLP.Converter.penn_to_wn):
print(tagged_alternative)
continue
new_sentence = extract_2(tagged_alternative)
if utils.verify(new_sentence):
yield new_sentence
def extract_7(tagged_corpus, special=False):
global TIER
TIER = 3.7
sentence = EC.SentenceResult()
mark = False
for word in tagged_corpus:
if word[0].lower() == "by":
mark = True
if word[0].lower() == "to":
if sentence["sbj"] and sentence["pdt"] and not sentence["obj"]:
sentence["obj_toi"] = True
elif not sentence["sbj"]:
sentence["sbj_toi"] = True
if word[0].lower() in ["a", "the", "an"]:
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_adt"] or sentence["obj_att"]:
print("determiner too late error")
return {"sbj": None, "pdt": None}
sentence["obj_dt"] = word
elif not sentence["sbj_toi"]:
if sentence["sbj_adt"] or sentence["sbj_att"]:
print("determiner too late error (subject)")
return {
"sbj": None,
"pdt": None
} # so that the error will be noted and passed over
sentence["sbj_dt"] = word
if NLP.Converter.penn_to_wn(word[1]) == "r":
if sentence["pdt"]:
if not sentence["obj_dtr"] and not mark and not sentence[
"obj_adt"] and not sentence["obj_att"] and not sentence[
"obj"]: # the same thing
sentence["pdt_ptc"].append(word)
elif not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"])
- 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["obj_att"].append([[word], None])
else:
sentence["obj_att"].append([[word], None])
else:
if sentence["obj"]:
sentence["obj_tpa"].append(word)
else:
if not sentence["pss_vce"] and sentence["sbj"]:
sentence["pdt_ptc"].append(word)
if not sentence["sbj_toi"]:
if not sentence["pdt"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][
len(sentence["sbj_att"]) - 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["sbj_att"].append([[word], None])
else:
sentence["sbj_att"].append([[word], None])
else:
if sentence["sbj"]:
sentence["sbj_tpa"].append(word)
if NLP.Converter.penn_to_wn(
word[1]) == "a" or word[1] == "PRP$" or word[1] == "CD":
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["obj_att"].append([[], word])
else:
sentence["obj_att"].append([[], word])
elif not sentence["sbj_toi"] and not sentence["pdt"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["sbj_att"].append([[], word])
else:
sentence["sbj_att"].append([[], word])
if sentence["obj_cmp"] and (NLP.Converter.penn_to_wn(word[1]) == "a" or
NLP.Converter.penn_to_wn(word[1]) == "r"):
sentence["obj"] = word
res = utils.deep_in(sentence["obj_att"], word)
if res:
sentence["obj_att"].remove(res)
if NLP.Converter.penn_to_wn(word[1]) == "n" or \
word[1] == "PRP" or word[0].lower() == "it" or \
word[1] == "VBG" or \
NLP.Determiner.is_gerund(word[0]):
if NLP.Converter.penn_to_wn(word[1]) == "n":
if not sentence["pdt"] and not sentence["sbj_toi"]:
if sentence["sbj"]:
sentence["sbj_adt"].append(sentence["sbj"])
sentence["sbj"] = word
else:
sentence["sbj"] = word
elif not sentence["obj_toi"]:
if sentence["obj"]:
if not sentence["obj_cmp"]:
sentence["obj_adt"].append(sentence["obj"])
sentence["obj"] = word
else:
sentence["obj"] = word
else:
if sentence["pdt"] and not sentence["obj_toi"]:
sentence["obj"] = word
if word[1] == "PRP" and word[0].lower() in [
"he", "she", "they", "I"
]:
sentence["opn_dis"] = True
sentence["err"] = True
elif not sentence["sbj_toi"]:
if word[1] == "PRP" and word[0].lower() in [
"him", "her", "them", "me"
]:
if not special:
sentence["spn_dis"] = True
sentence["err"] = True
else:
sentence["sbj"] = (NLP.Converter.to_sbj(word[0]),
word[1])
else:
sentence["sbj"] = word
if NLP.Converter.penn_to_wn(word[1]) == "v" and \
word[1] != "VBG":
if sentence["obj_cmp"] and word[1] == "VBN":
sentence["obj_cmp"] = False
sentence["pss_vce"] = True
sentence["pdt"] = word
if sentence["sbj_toi"] and not sentence["sbj"]:
sentence["sbj"] = word
elif not sentence["pdt"]:
sentence["pdt"] = word
if not sentence["sbj"]:
print("subjectless predicate error")
print("tgd", tagged_corpus)
print("stc", sentence)
return {"sbj": None, "pdt": None}
if sentence["sbj"][1] in ["NNS", "NNPS"] and word[1] == "VBZ":
sentence["svb_dis"] = True
sentence["err"] = True
if (sentence["sbj"][1] in ["NN", "NNP"]
or sentence["sbj"][0].lower()
in ["he", "she", "it"]) and word[1] == "VB":
sentence["svb_dis"] = True
sentence["err"] = True
if word[0].lower() in [
"am", "is", "are", "was", "were", "be", "been"
]:
sentence["obj_cmp"] = True
elif sentence["obj_toi"] and not sentence["obj"]:
sentence["obj"] = word
if word[1] == "RP":
if not sentence["obj_cmp"]:
sentence["pdt_ptc"].append(word)
# placing apposition detection outside the main loop will also be clearer and more efficient, and do it before the swap
if sentence["sbj"][
1] == "NNP": # appositions only apply to proper nouns currently, because most are named entities
for adt in sentence["sbj_adt"][::-1]:
if adt[1] == "NNP":
sentence["sbj_adt"].remove(adt)
sentence["sbj"] = (adt[0] + " " + sentence["sbj"][0], "NNP")
else:
break
if sentence["obj"]:
if sentence["obj"][
1] == "NNP": # appositions only apply to proper nouns currently, because most are named entities
for adt in sentence["obj_adt"][::-1]:
if adt[1] == "NNP":
sentence["obj_adt"].remove(adt)
sentence["obj"] = (adt[0] + " " + sentence["obj"][0],
"NNP")
else:
break
# remove those adjective phrases in attributes that have no body, just adverbs
for char in sentence["sbj_att"]:
if char[1] is None:
sentence["sbj_att"].remove(char)
for char in sentence["obj_att"]:
if char[1] is None:
sentence["obj_att"].remove(char)
# will be more readable to leave this switch outside the main loop
if sentence["pss_vce"]:
sentence.swap()
if not sentence["sbj"]:
sentence["sbj"] = EC.Pointer("&any")
return sentence
def extract_3(tagged_corpus):
global TIER
TIER = 3.3
sentence = EC.SentenceResult()
mark = False
for word in tagged_corpus:
if word[0].lower() == "by":
mark = True
if word[0].lower() == "to":
if sentence["sbj"] and sentence["pdt"] and not sentence["obj"]:
sentence["obj_toi"] = True
elif not sentence["sbj"]:
sentence["sbj_toi"] = True
if word[0].lower() in ["a", "the", "an"]:
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_adt"] or sentence["obj_att"]:
return {"sbj": None, "pdt": None}
sentence["obj_dt"] = word
elif not sentence["sbj_toi"]:
if sentence["sbj_adt"] or sentence["sbj_att"]:
return {
"sbj": None,
"pdt": None
} # so that the error will be noted and passed over
sentence["sbj_dt"] = word
if NLP.Converter.penn_to_wn(word[1]) == "r":
if sentence["pdt"]:
if not sentence["obj_dtr"] and not mark: # the same thing
sentence["pdt_ptc"].append(word)
elif not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"])
- 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["obj_att"].append([[word], None])
else:
sentence["obj_att"].append([[word], None])
else:
if sentence["obj"]:
sentence["obj_tpa"].append(word)
else:
if not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"])
- 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["sbj_att"].append([[word], None])
else:
sentence["sbj_att"].append([[word], None])
else:
if sentence["sbj"]:
sentence["sbj_tpa"].append(word)
if NLP.Converter.penn_to_wn(
word[1]) == "a" or word[1] == "PRP$" or word[1] == "CD":
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["obj_att"].append([[], word])
else:
sentence["obj_att"].append([[], word])
elif not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["sbj_att"].append([[], word])
else:
sentence["sbj_att"].append([[], word])
if sentence["obj_cmp"] and NLP.Converter.penn_to_wn(word[1]) == "a":
sentence["obj"] = word
res = utils.deep_in(sentence["obj_att"], word)
if res:
sentence["obj_att"].remove(res)
if NLP.Converter.penn_to_wn(word[1]) == "n" or \
word[1] == "PRP" or word[0].lower() == "it" or \
word[1] == "VBG" or \
NLP.Determiner.is_gerund(word[0]):
if NLP.Converter.penn_to_wn(word[1]) == "n":
if not sentence["pdt"] and not sentence["sbj_toi"]:
if sentence["sbj"]:
sentence["sbj_adt"].append(sentence["sbj"])
sentence["sbj"] = word
else:
sentence["sbj"] = word
elif not sentence["obj_toi"]:
if sentence["obj"]:
if not sentence["obj_cmp"]:
sentence["obj_adt"].append(sentence["obj"])
sentence["obj"] = word
else:
sentence["obj"] = word
else:
if sentence["pdt"] and not sentence["obj_toi"]:
sentence["obj"] = word
if word[1] == "PRP" and word[0].lower() in [
"he", "she", "they", "I"
]:
sentence["opn_dis"] = True
sentence["err"] = True
elif not sentence["sbj_toi"]:
if word[1] == "PRP" and word[0].lower() in [
"him", "her", "them", "me"
]:
sentence["spn_dis"] = True
sentence["err"] = True
sentence["sbj"] = word
if NLP.Converter.penn_to_wn(word[1]) == "v" and \
word[1] != "VBG":
if sentence["obj_cmp"] and word[1] == "VBN":
sentence["obj_cmp"] = False
sentence["pss_vce"] = True
sentence["pdt"] = word
if sentence["sbj_toi"] and not sentence["sbj"]:
sentence["sbj"] = word
elif not sentence["pdt"]:
sentence["pdt"] = word
if not sentence["sbj"]:
return {"sbj": None, "pdt": None}
if sentence["sbj"][1] in ["NNS", "NNPS"] and word[1] == "VBZ":
sentence["svb_dis"] = True
sentence["err"] = True
if (sentence["sbj"][1] in ["NN", "NNP"]
or sentence["sbj"][0].lower()
in ["he", "she", "it"]) and word[1] == "VB":
sentence["svb_dis"] = True
sentence["err"] = True
if word[0].lower() in [
"am", "is", "are", "was", "were", "be", "been"
]:
sentence["obj_cmp"] = True
elif sentence["obj_toi"] and not sentence["obj"]:
sentence["obj"] = word
if word[1] == "RP":
if not sentence["obj_cmp"]:
sentence["pdt_ptc"].append(word)
# will be more readable to leave this switch outside the main loop
if sentence["pss_vce"]:
sentence.swap()
if not sentence["sbj"]:
sentence["sbj"] = EC.Pointer("&any")
return sentence
def extract_1(tagged_corpus):
global TIER
TIER = 3.1
sentence = {
"sbj_toi": False,
"sbj_dt": None,
"sbj_att": [],
"sbj_adt": [],
"sbj": None,
"sbj_tpa": [],
"pdt": None,
"obj_toi": False,
"obj_cmp": False,
"obj_dt": None,
"obj_att": [],
"obj_adt": [],
"obj": None,
"obj_tpa": [],
"pss_vce": False, # passive voice
"svb_dis": False, # subject-verb disagreement
"spn_dis": False, # subject pronoun nominative disagreement
"opn_dis": False,
"err": False
}
for word in tagged_corpus:
if word[0].lower() == "to":
if sentence["sbj"] and sentence["pdt"] and not sentence["obj"]:
sentence["obj_toi"] = True
elif not sentence["sbj"]:
sentence["sbj_toi"] = True
if word[0].lower() in ["a", "the"]:
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_adt"] or sentence["obj_att"]:
return {"sbj": None, "pdt": None}
sentence["obj_dt"] = word
elif not sentence["sbj_toi"]:
if sentence["sbj_adt"] or sentence["sbj_att"]:
return {
"sbj": None,
"pdt": None
} # so that the error will be noted and passed over
sentence["sbj_dt"] = word
if NLP.Converter.penn_to_wn(word[1]) == "r":
if sentence["pdt"]:
if not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"])
- 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["obj_att"].append([[word], None])
else:
sentence["obj_att"].append([[word], None])
else:
if sentence["obj"]:
sentence["obj_tpa"].append(word)
else:
if not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"])
- 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["sbj_att"].append([[word], None])
else:
sentence["sbj_att"].append([[word], None])
else:
if sentence["sbj"]:
sentence["sbj_tpa"].append(word)
if NLP.Converter.penn_to_wn(
word[1]) == "a" or word[1] == "PRP$" or word[1] == "CD":
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["obj_att"].append([[], word])
else:
sentence["obj_att"].append([[], word])
elif not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["sbj_att"].append([[], word])
else:
sentence["sbj_att"].append([[], word])
if sentence["obj_cmp"] and NLP.Converter.penn_to_wn(word[1]) == "a":
sentence["obj"] = word
res = utils.deep_in(sentence["obj_att"], word)
if res:
sentence["obj_att"].remove(res)
if NLP.Converter.penn_to_wn(word[1]) == "n" or \
word[1] == "PRP" or word[0].lower() == "it" or \
word[1] == "VBG" or \
NLP.Determiner.is_gerund(word[0]):
if NLP.Converter.penn_to_wn(word[1]) == "n":
if not sentence["pdt"] and not sentence["sbj_toi"]:
if sentence["sbj"]:
sentence["sbj_adt"].append(sentence["sbj"])
sentence["sbj"] = word
else:
sentence["sbj"] = word
elif not sentence["obj_toi"]:
if sentence["obj"]:
if not sentence["obj_cmp"]:
sentence["obj_adt"].append(sentence["obj"])
sentence["obj"] = word
else:
sentence["obj"] = word
else:
if sentence["pdt"] and not sentence["obj_toi"]:
sentence["obj"] = word
if word[1] == "PRP" and word[0].lower() in [
"he", "she", "they"
]:
sentence["opn_dis"] = True
sentence["err"] = True
elif not sentence["sbj_toi"]:
if word[1] == "PRP" and word[0].lower() in [
"him", "her", "them"
]:
sentence["spn_dis"] = True
sentence["err"] = True
sentence["sbj"] = word
if NLP.Converter.penn_to_wn(word[1]) == "v" and \
word[1] != "VBG":
if sentence["obj_cmp"] and word[1] == "VBN":
sentence["obj_cmp"] = False
sentence["pss_vce"] = True
sentence["pdt"] = word
if sentence["sbj_toi"] and not sentence["sbj"]:
sentence["sbj"] = word
elif not sentence["pdt"]:
sentence["pdt"] = word
if not sentence["sbj"]:
return {"sbj": None, "pdt": None}
if sentence["sbj"][1] in ["NNS", "NNPS"] and word[1] == "VBZ":
sentence["svb_dis"] = True
sentence["err"] = True
if word[0].lower() in [
"am", "is", "are", "was", "were", "be", "been"
]:
sentence["obj_cmp"] = True
elif sentence["obj_toi"] and not sentence["obj"]:
sentence["obj"] = word
return sentence
def extract_1(tagged_corpus):
global TIER
TIER = 2.1
# from now on, these will be using a more concise method
sentence = {
"sbj_toi": False,
"sbj_dt": None,
"sbj_att": [],
"sbj_adt": [],
"sbj": None,
"sbj_tpa": [],
"pdt": None,
"obj_toi": False,
"obj_cmp": False, # is complement?
"obj_dt": None,
"obj_att": [],
"obj_adt": [],
"obj": None,
"obj_tpa": []
}
for word in tagged_corpus:
if word[0].lower() == "to":
if sentence["sbj"] and sentence["pdt"] and not sentence["obj"]:
sentence["obj_toi"] = True
elif not sentence["sbj"]:
sentence["sbj_toi"] = True
if word[1] == "DT":
if sentence["pdt"] and not sentence["obj_toi"]:
sentence["obj_dt"] = word
elif not sentence["sbj_toi"]:
sentence["sbj_dt"] = word
if NLP.Converter.penn_to_wn(word[1]) == "r":
if sentence["pdt"]:
if not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(obj_att) - 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["obj_att"].append([[word], None])
else:
sentence["obj_att"].append([[word], None])
else:
if sentence["obj"]:
sentence["obj_tpa"].append(word)
else:
if not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sbj_att) - 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["sbj_att"].append([[word], None])
else:
sentence["sbj_att"].append([[word], None])
else:
if sentence["sbj"]:
sentence["sbj_tpa"].append(word)
if NLP.Converter.penn_to_wn(word[1]) == "a" or word[1] == "PRP$":
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["obj_att"].append([[], word])
else:
sentence["obj_att"].append([[], word])
elif not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["sbj_att"].append([[], word])
else:
sentence["sbj_att"].append([[], word])
if sentence["obj_cmp"] and NLP.Converter.penn_to_wn(word[1]) == "a":
# if sentence["obj_cmp"] is not False, which is its default value
# then the predicate has to have been found already
# making checking for sentence["pdt"] unnecessary.
sentence["obj"] = word # this could work ...
res = utils.deep_in(sentence["obj_att"], word)
if res:
sentence["obj_att"].remove(res)
if NLP.Converter.penn_to_wn(word[1]) == "n" or \
word[1] == "PRP" or word[0].lower() == "it" or \
word[1] == "VBG" or \
NLP.Determiner.is_gerund(word[0]):
# subjects have to appear before objects
if NLP.Converter.penn_to_wn(word[1]) == "n":
if not sentence["pdt"] and not sentence["sbj_toi"]:
if sentence["sbj"]:
sentence["sbj_adt"].append(sentence["sbj"])
sentence["sbj"] = word
else:
sentence["sbj"] = word
elif not sentence["obj_toi"]:
if sentence["obj"]:
if not sentence["obj_cmp"]:
sentence["obj_adt"].append(sentence["obj"])
sentence["obj"] = word
else:
sentence["obj"] = word
else:
if sentence["pdt"] and not sentence["obj_toi"]:
sentence["obj"] = word
elif not sentence["sbj_toi"]:
sentence["sbj"] = word
if NLP.Converter.penn_to_wn(word[1]) == "v" and \
word[1] != "VBG":
if sentence["sbj_toi"] and not sentence["sbj"]:
sentence["sbj"] = word
elif not sentence["pdt"]:
sentence["pdt"] = word # the all important line
# this is what tier 2.1 is all about
if word[0].lower() in [
"am", "is", "are", "was", "were", "be", "been"
]:
sentence["obj_cmp"] = True
elif sentence["obj_toi"] and not sentence["obj"]:
sentence["obj"] = word
return sentence # much more concise!
def extract_2(tagged_corpus):
global TIER
TIER = 2.2
sentence = {
"sbj_toi": False,
"sbj_dt": None,
"sbj_att": [],
"sbj_adt": [],
"sbj": None,
"sbj_tpa": [],
"pdt": None,
"obj_toi": False,
"obj_cmp": False,
"obj_dt": None,
"obj_att": [],
"obj_adt": [],
"obj": None,
"obj_tpa": [],
"disagree": False
}
for word in tagged_corpus:
if word[0].lower() == "to":
if sentence["sbj"] and sentence["pdt"] and not sentence["obj"]:
sentence["obj_toi"] = True
elif not sentence["sbj"]:
sentence["sbj_toi"] = True
if word[0].lower() in ["a", "the"]:
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_adt"] or sentence["obj_att"]:
return {"sbj": None, "pdt": None}
sentence["obj_dt"] = word
elif not sentence["sbj_toi"]:
if sentence["sbj_adt"] or sentence["sbj_att"]:
return {
"sbj": None,
"pdt": None
} # so that the error will be noted and passed over
sentence["sbj_dt"] = word
if NLP.Converter.penn_to_wn(word[1]) == "r":
if sentence["pdt"]:
if not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"])
- 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["obj_att"].append([[word], None])
else:
sentence["obj_att"].append([[word], None])
else:
if sentence["obj"]:
sentence["obj_tpa"].append(word)
else:
if not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"])
- 1]
if closest[1] is None:
closest[0].append(word)
else:
sentence["sbj_att"].append([[word], None])
else:
sentence["sbj_att"].append([[word], None])
else:
if sentence["sbj"]:
sentence["sbj_tpa"].append(word)
if NLP.Converter.penn_to_wn(word[1]) == "a" or word[1] == "PRP$":
if sentence["pdt"] and not sentence["obj_toi"]:
if sentence["obj_att"]:
closest = sentence["obj_att"][len(sentence["obj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["obj_att"].append([[], word])
else:
sentence["obj_att"].append([[], word])
elif not sentence["sbj_toi"]:
if sentence["sbj_att"]:
closest = sentence["sbj_att"][len(sentence["sbj_att"]) - 1]
if closest[1] is None:
closest[1] = word
else:
sentence["sbj_att"].append([[], word])
else:
sentence["sbj_att"].append([[], word])
if sentence["obj_cmp"] and NLP.Converter.penn_to_wn(word[1]) == "a":
sentence["obj"] = word
res = utils.deep_in(sentence["obj_att"], word)
if res:
sentence["obj_att"].remove(res)
if NLP.Converter.penn_to_wn(word[1]) == "n" or \
word[1] == "PRP" or word[0].lower() == "it" or \
word[1] == "VBG" or \
NLP.Determiner.is_gerund(word[0]):
if NLP.Converter.penn_to_wn(word[1]) == "n":
if not sentence["pdt"] and not sentence["sbj_toi"]:
if sentence["sbj"]:
sentence["sbj_adt"].append(sentence["sbj"])
sentence["sbj"] = word
else:
sentence["sbj"] = word
elif not sentence["obj_toi"]:
if sentence["obj"]:
if not sentence["obj_cmp"]:
sentence["obj_adt"].append(sentence["obj"])
sentence["obj"] = word
else:
sentence["obj"] = word
else:
if sentence["pdt"] and not sentence["obj_toi"]:
sentence["obj"] = word
elif not sentence["sbj_toi"]:
sentence["sbj"] = word
if NLP.Converter.penn_to_wn(word[1]) == "v" and \
word[1] != "VBG":
if sentence["sbj_toi"] and not sentence["sbj"]:
sentence["sbj"] = word
elif not sentence["pdt"]:
sentence["pdt"] = word
if not sentence["sbj"]:
return {
"sbj": None,
"pdt": None
} # so that the error will be noted and passed over
# will first do some basic determination based on plurality tags
if sentence["sbj"][1] in ["NNS", "NNPS"] and word[1] == "VBZ":
sentence["disagree"] = True
if word[0].lower() in [
"am", "is", "are", "was", "were", "be", "been"
]:
sentence["obj_cmp"] = True
elif sentence["obj_toi"] and not sentence["obj"]:
sentence["obj"] = word
return sentence