bio-nlp/drugprocessor.py at master · librairy/bio-nlp · 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
import pysolr
import requests
import csv
from annoy import AnnoyIndex

class DrugProcessor:

    def __init__(self):
        # Setup a Solr instance. The timeout is optional.
        self.atc_solr = pysolr.Solr('http://librairy.linkeddata.es/data/atc', timeout=20)

        self.cord19_solr_url = 'http://librairy.linkeddata.es/data/covid-paragraphs'
        self.cord19_solr = pysolr.Solr(self.cord19_solr_url, timeout=20)

        self.index = AnnoyIndex(1046, 'angular') #186
        self.index.load('index.annoy')

        with open('index.dictionary', mode='r') as infile:
            reader = csv.reader(infile)
            self.index_dict = dict((rows[0],int(rows[1])) for rows in reader)

        self.index_inv_dict = {}
        for key in self.index_dict.keys():
            value = self.index_dict[key]
            self.index_inv_dict[value]=key


    def find_drugs(self, keyword):
        query = "label_t:\""+keyword+"\" or code_s:"+keyword + " or id:" + keyword
        results = self.atc_solr.search(query)
        drugs= []
        for result in results:
            drug = {}
            if ('label_t' in result):
                drug['name'] = result['label_t']
            if ('id' in result):
                drug['code'] = result['id']
            if ('level_i' in result):
                drug['level'] = result['level_i']
            drugs.append(drug)
        #print("found drugs",drugs)
        return drugs

    def get_drugs_as_terms(self, size,level):
        params = {}
        drug_level = 5
        if (int(level) > 0):
            drug_level=level
        field = 'bionlp_drugs_C'+str(drug_level)
        params['terms.fl']=field
        params['terms.sort']='count'
        params['terms.mincount']=1
        params['terms.limit']=size

        url = self.cord19_solr_url + "/terms"
        resp = requests.get(url=url, params=params)
        data = resp.json()
        #print(data)
        drugs = []
        results = data['terms'][field]
        i = 0
        while (i<len(results)):
            drug_code = results[i]
            #print("drug code", drug_code)
            i+=1
            frequency = results[i]
            i+=1
            drugs_candidates = self.find_drugs(drug_code.upper())
            if (len(drugs_candidates) >0 ):
                drug = drugs_candidates[0]
                drug['freq'] = frequency
                drugs.append(drug)
        return drugs


    def get_drugs(self, query, size, level):
        if (query == "*:*"):
            return self.get_drugs_as_terms(size,level)
        counter = 0
        completed = False
        window_size=50000
        cursor = "*"
        drugs = {}
        while (not completed):
            old_counter = counter
            try:
                fields = ["bionlp_drugs_N"+str(l)  for l in range(1,6)]
                if (int(level) >= 0):
                    fields = ["bionlp_drugs_N"+str(level)]
                #print("searching drugs in paragraphs by: ", query , " with fields: ", fields)
                paragraphs = self.cord19_solr.search(q=query,fl=",".join(fields),rows=window_size,cursorMark=cursor,sort="id asc")
                cursor = paragraphs.nextCursorMark
                counter += len(paragraphs)
                for paragraph in paragraphs:
                    for field in fields:
                        if (field in paragraph):
                            for drug in paragraph[field]:
                                if (not drug in drugs):
                                    drugs[drug] = 0
                                drugs[drug] = drugs[drug] +1
                if (old_counter == counter):
                    break
            except Exception as e:
                print(repr(e))
        result = []
        for w in sorted(drugs, key=drugs.get, reverse=True):
            found_drugs = self.find_drugs(w)
            if (len(found_drugs) > 0):
                drug = found_drugs[0]
                drug['freq'] = drugs[w]
                result.append(drug)
            if (len(result) >= size):
                break
        return result

    def get_related_drugs(self, code, size, level):
        search_word=code
        drugs = []
        if (not search_word in self.index_dict):
            return drugs
        ref_index = self.index_dict[search_word]
        for neighbour in self.index.get_nns_by_item(ref_index, int(size)*3):
            #print("neighbour",neighbour)
            if (neighbour in self.index_inv_dict):
                neighbour_code=self.index_inv_dict[neighbour]
                if (code != neighbour_code):
                    neighbours = self.find_drugs(neighbour_code)
                    if (len(neighbours) > 0):
                        drug = neighbours[0]
                        if (level == -1 ):
                            drugs.append(drug)
                        elif (drug['level'] == level) :
                            drugs.append(drug)

            if (len(drugs) == size):
                break
        return drugs