MulCo-Alpha/semantic_graph.py at main · CC-RMD-EpiBio/MulCo-Alpha · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
from graph_classes import *
import sys
import pickle
from datetime import datetime, timedelta
import spacy

class SemanticGraph:
    def __init__(self):
        self.spacy_model = spacy.load("en_core_web_sm")

    def forward(self,doc_sents,doc_events,doc_timexes,doc_timex_vals,doc_id):
        self.get_spacy(doc_sents)
        self.get_graph(doc_sents,doc_events,doc_timexes,doc_timex_vals,doc_id)
    def get_spacy(self,doc_sents):
        full_doc = []

        for sent_num,sent in list(sorted(doc_sents.items(), key = lambda x:x[0])):
            full_doc.extend(sent.split())

        tokenized_doc = spacy.tokens.doc.Doc(self.spacy_model.vocab,words=full_doc)

        for name, proc in self.spacy_model.pipeline:
            tokenized_doc = proc(tokenized_doc)
        self.spacy = tokenized_doc

    def get_graph(self,doc_sents,doc_events,doc_timexes,doc_timex_vals,doc_id):
        self.semantic_graph = Graph()

        print(doc_id)
        try:
            doc_date_start = datetime.strptime(doc_id[3:11],"%Y%m%d")
        except:
            try:
                doc_date_start = datetime.strptime(doc_id[2:8],"%y%m%d")
            except:
                print("BAD DATE")
        doc_date_end = doc_date_start + timedelta(days=1)
        doc_date = (doc_date_start,doc_date_end)

        self.get_spacy(doc_sents)

        for tid in doc_timex_vals:
            if tid in doc_timexes.keys():
                _,timex_val = doc_timex_vals[tid]

                cmpr = self.date_compare(doc_date,timex_val)
                self.semantic_graph.add_edge(True,"D",tid,"DCT-Timex",cmpr)

                for tid2 in doc_timex_vals:
                    if tid2 in doc_timexes.keys():
                        _,timex_val_2 = doc_timex_vals[tid2]
                        if tid != tid2:
                            cmpr = self.date_compare(timex_val,timex_val_2)
                            self.semantic_graph.add_edge(True,tid,tid2,"Timex-Timex",cmpr)

        event_preds = {}
        for eid in doc_events:
            _,e_word,_ = doc_events[eid]

            event_spacy = self.spacy[e_word]
            event_preds[e_word] = (eid,False)

        for tid in doc_timex_vals:
            if tid in doc_timexes.keys():
                _,t_word,_ = doc_timexes[tid]
                timex_spacy = self.spacy[t_word]
                pred = timex_spacy
                while pred.i not in event_preds.keys() and pred.dep_ != "ROOT":
                    pred = pred.head

                if pred.i in event_preds.keys():
                    eid,_ = event_preds[pred.i]
                    self.semantic_graph.add_edge(False,eid,tid,"Pred-Timex")
                    event_preds[pred.i] = (eid,True)

        for e_key in event_preds:
            eid,ebool = event_preds[e_key]
            if ebool == False:
                self.semantic_graph.add_edge(False,eid,eid,"Self-Loop")

    def date_compare(self,t1,t2):
        t1_start,t1_end = t1
        t2_start,t2_end = t2

        if t1_start is None or t2_start is None:
            return "None"

        t1_start = t1_start.date()
        t1_end = t1_end.date()
        t2_start = t2_start.date()
        t2_end = t2_end.date()

        if t1_start > t2_end:
            return "After"
        if t1_end < t2_start:
            return "Before"
        if t1_start == t2_start and t1_end == t2_end:
            return "Simultaneous"

        return "None"

if __name__ == "__main__":
    doc_file = sys.argv[1]
    full_dict = pickle.load(open(doc_file,"rb"))

    timex_file = sys.argv[3]
    all_timex_vals = pickle.load(open(timex_file,"rb"))

    graphs_dict = {}
    for document in full_dict:
        doc_id = document.split('.tml')[0]
        doc_dict = full_dict[document]
        doc_sents = doc_dict["sents"]
        doc_events = doc_dict["events"]
        doc_timexes = doc_dict["timexes"]
        doc_timex_vals = all_timex_vals[doc_id]

        doc_sents = {int(x):y for x,y in doc_sents.items()}
        doc_events = {x:[int(y[0]), int(y[1])-1, y[2]] for x,y in doc_events.items()}
        doc_timexes = {x:[int(y[0]), int(y[1])-1, y[2]] for x,y in doc_timexes.items()}

        doc_len = 0
        for sent_num, sent in list(sorted(doc_sents.items(), key = lambda x:x[0])):
            if sent_num == 0:
                doc_len += len(sent.split())
                continue
            for event in doc_events:
                if doc_events[event][0] == sent_num:
                    doc_events[event][1] = doc_events[event][1] + doc_len
            for timex in doc_timexes:
                if doc_timexes[timex][0] == sent_num:
                    doc_timexes[timex][1] = doc_timexes[timex][1] + doc_len
            doc_len += len(sent.split())

        semantic_model = SemanticGraph()
        semantic_model.forward(doc_sents,doc_events,doc_timexes,doc_timex_vals,doc_id)
        semantic_graph = semantic_model.semantic_graph
        graphs_dict[doc_id] = semantic_graph

    out_file = sys.argv[2]
    pickle.dump(graphs_dict,open(out_file,"wb+"))