-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpdf2chem.py
331 lines (285 loc) · 14.2 KB
/
pdf2chem.py
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
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
# Imports
import chemdataextractor as cde
import cirpy
import time
import sys
import re
import pandas as pd
import os
import textract
from datetime import datetime
# future work - a choice for whether to query 3-letter words
# if curation_type == "Regular (ignores 3-letter words)":
# regex_number = 4
# else:
# regex_number = 3
# curation_type = "Exhaustive (includes 3-letter words)"
pdf_path = os.getcwd()
regex_number = 3
IN_COLAB = "google.colab" in sys.modules
if IN_COLAB:
pdf_method = "pdfminer"
else:
pdf_method = "pdftotext"
print("We\'ll use {} as the pdf extraction method.".format(pdf_method))
# element symbols and false positives hard-coded until stable hosting is found
element_symbols = ['h', 'he', 'li', 'be', 'b', 'c', 'n', 'o', 'f', 'ne', 'na',
'mg', 'al', 'si', 'p', 's', 'cl', 'ar', 'k', 'ca',
'sc', 'ti', 'v', 'cr', 'mn', 'fe', 'co', 'ni', 'cu',
'zn', 'ga', 'ge', 'as', 'se', 'br', 'kr', 'rb', 'sr',
'y', 'zr', 'nb', 'mo', 'tc', 'ru', 'rh', 'pd', 'ag',
'cd', 'in', 'sn', 'sb', 'te', 'i', 'xe', 'cs', 'ba',
'la', 'ce', 'pr', 'nd', 'pm', 'sm', 'eu', 'gd', 'tb',
'dy', 'ho', 'er', 'tm', 'yb', 'lu', 'hf', 'ta', 'w',
're', 'os', 'ir', 'pt', 'au', 'hg', 'tl', 'pb', 'bi',
'po', 'at', 'rn', 'fr', 'ra', 'ac', 'th', 'pa', 'u',
'np', 'pu', 'am', 'cm', 'bk', 'cf', 'es', 'fm', 'md',
'no', 'lr', 'rf', 'db', 'sg', 'bh', 'hs', 'mt', 'ds ',
'rg ', 'cn ', 'nh', 'fl', 'mc', 'lv', 'ts', 'og']
false_positives = ['reno', 'lower', 'format', 'lead', 'nci', 'cc', 'isi',
'doi', "\\'b", 'is', 'ph', 'mv', 'zone', 'based', 'on',
'final', 'kato', 'cm', 'life', 'versus', 'www', 'can',
'ate', 'mm', 'crystal', 'sem', 'an', 's1', 'force', 'may',
'any', 'lau', 'voltage', 'kc', 'mino', 'm. h.', 'set',
'selective', 'c.p.k.', 'same', 'page 10', 'm-1', 'ai',
'c1', 'm2', 'et', 'fulfill', 'dry', 'via', 'may', 'pka',
'any', 'edge', 'b.v.', 'final', 'rt', '2b', 'h.y.', 'y.k.',
'v.v.', 'w.y.', 'good', 'region', 'cycle', 'des', 'force',
'may', 'dsc', 'chcl', 'counter', 'van', 'see', 'best',
'green', 'equal', 'result', 'challenge', 'substance',
'spectrum', 'der', 'its', 'glass', 'all', 'new', 'mix',
'so', 'soc.', 'arm', 'nm', 'ran', 'enable', 'sd', 'saa',
'map', 'ac1', 'fab', 'act', 'b7', 'liu', 'check', 'dual',
'via', 'den', 'fc', 'if', 'rapid', 'san', 'van', 'control',
'see', 'harry', 'adam', 'line', 'ac-1', 'sig', 'recruit',
'bli', 'test', 'tau', 'acs', 'iap', 'box', 'campaign',
'target', 'gfp', 'new', 'cv', 'rt', 'lid', 'compound',
'selective', 'rfb', 'ment', 'est', 'mm', 'con', 'con-',
's4', 'harry', 'ip', 'lp', 'ple', 'ml', 'prone', 'pka',
'sum', 'derivative', 'ten', 'min', 'vortex', 'gradual',
'tot', 'ber', 'red', 'ing', 'para', 'phs', 'gen', 'dft',
'nals', 'enable', 'set', 'versus', 'ma', 'the', 'and',
'eo', 'cps', 'ep', 'are', 'same', 'cos', 'age', 'sem',
's4', 'cycle', 'far', 'cal', 'overall', 'net', 'et', 'ml',
's1', 'prone', 'capture', 'or', 'rise', 'but', 'diurnal',
'dry', 'may', 'of', 'off', 'dp', 'if', 'dants', 'van',
'eden', 'line', 'tx', 'top', 'va', 'per', 'ny', 'on',
'ing', 'cp', 'for', 'dc', 'air', 'nhe', 'gas', 'zonal',
'all', 'new', 'based', 'had', 'ph', 'cm3', 'pyrite', 'soc',
'ser', 'acc', 'res', 'eds', 'mp', 'pro', 'inc', 'im', 'bv',
'disodium', 'ab', 'ed', 'carboxylate', '1mm', 'nat', 'eq',
'acc', 'sci', 'mol', 'int', 'sc-s', 'scs', 'gu', 'atm',
'shi', '2az', 'abbott', 'ms', 'wang', 'pdc', 'franklin',
'bay', 'dess', 'hbd', 'retard', 'intercept', 'iii',
'acid', 'fraction', 'aldrich', 'triton', 'cda', 'cyano',
'vinyl', 'flux', 'ethyl', 'methyl', 'mit', 'trigger',
'accelerate', 'ants', 'pentyl', 'laser', 'india', 'dos',
'los', 'acetyl', 'dec', 'sheets', 'tem', 'dimethyl',
'serial', 'tag', 'tandem', 'trap', 'mic', 'exciton',
'aldehyde', 'combat', 'roi', 'probiotic', 'antiviral',
'cada', 'beam', 'austin', 'lactone', 'lumen', 'diethyl',
'optimal', 'sulfoxide', 'gm3', 'gel', 'blockade', 'omega',
'cubes', 'bin', 'alcohols', 'alcohol', 'benchmark',
'portal', 'matrix', 'apex', 'bacterial', 'cube', 'linker',
'cascade', 'optimum', 'carbonyl oxygen', 'facet', 'shield']
if regex_number == 3:
false_positives = [word for word in false_positives if not
re.search("[a-zA-Z0-9+-]{3}", word) or
re.search("[a-zA-Z0-9+-]{4}", word)]
"""# Define functions"""
def quick_curate(pdf_path, pdf_method, false_positives, regex_number):
# extract the text from the pdf
# the pdf_method should adapt to both local and hosted
# runtime compatibility
text = textract.process(pdf_path, method=pdf_method)
# queue up and reset list used to process the paper
temp_word_list = []
# strip new line and other markup from pdf mining
text = str(text).replace("\\n-", '').replace('\-\n', '')
text = str(text).replace('\-\n-', '').replace('\\n', ' ')
text = str(text).replace('\n', ' ').replace('.', '')
text = str(text).replace('*', "").replace('ISSN', '').replace('NSF', '')
text = str(text).replace('NIH', '').replace("b'", '').replace(r"\r", '')
# split by white spaces
temp_word_list = re.split("\s+", str(text))
# try to remove reference section by cutting off everything after the last
# mention of reference
ref = [i for i, w in enumerate(temp_word_list) if
w.lower().startswith('reference')]
#print(ref)
try:
temp_word_list = temp_word_list[:(ref[-1])]
except Exception as e:
pass
# reconnect any words that got hyphenated and cut off at the end
# of a column
for i, word in enumerate(temp_word_list):
if re.search('[-]+$', word):
temp_word_list[i] = word.replace('-', '') + temp_word_list[i+1]
del(temp_word_list[i+1])
print('The initial list for {} has {} words.'.format(pdf_path,
len(temp_word_list)))
# reconstruct a text string from the cleaned list,
# as cde's NLP works on strings
cleaned_text = ''
for word in temp_word_list:
cleaned_text += word
cleaned_text += ' '
# have cde do NLP on the string and convert the results
# into a list of strings
doc = cde.Document(cleaned_text)
chemicals_all = [span for span in doc.cems]
chem_strings = [str(word).lower().replace('\n', ' ')
for word in chemicals_all]
# remove any blanks or null values
chem_strings = [word for word in chem_strings if word]
# remove anything left with a backslash in it
chem_strings = [word for word in chem_strings if not
re.search('[\\\+]', word)]
print('We\'ll attempt to resolve {} \
potential chemicals.'.format(len(chem_strings)))
# reset lists used for processing query hits and misses
smiles_list = []
already_queried = []
missed_items = []
for item in chem_strings:
# if Sn is found, it's probably tin, not S=C
if item.lower() == "sn":
smiles_list.append('SnH4')
print(item, smiles_list[-1])
continue
# keeping element symbols, such as H, C, or Na
# this may turn into an option
if item in element_symbols:
smiles_list.append(cirpy.resolve(item, 'smiles'))
print(item, smiles_list[-1])
continue
# Future work - include options for user to specify exclusion of
# 3-letter words
# adapt the regex code that leaves out short words/abbreviations
# to the user input above
# if regex_number == 4:
#
# if not re.search("[a-zA-Z0-9+-]{4}", item):
# smiles_list.append(None)
# print('Found a word that\'s a\
# likely false positive: {}'.format(item))
# missed_items.append(item)
# continue
if regex_number == 3:
if not re.search("[a-zA-Z0-9+-]{3}", item):
smiles_list.append(None)
print('Found a word that\'s a likely false \
positive: {}'.format(item))
missed_items.append(item)
continue
# save time by not querying chemicals that are in the text many times
if item in already_queried:
smiles_list.append(None)
print('We\'ve already queried this one: {}'.format(item))
# don't query the chemical if it's a known false positive
# these include author names and a few other odds and ends
elif item.strip('.').strip(',').lower() in false_positives:
smiles_list.append(None)
print('Found one known to be a false positive: {}'.format(item))
# if the item passes all the tests,
# attempt to resolve it via NIH's CIR
else:
try:
smiles_list.append(cirpy.resolve(item, 'smiles'))
print(item, smiles_list[-1])
time.sleep(0.21)
# except loop in here to account for
# internet stability issues and the like
except Exception as e:
try:
print(e)
print('Exception raised. Pausing for \
2 seconds and trying again')
time.sleep(2)
smiles_list.append(cirpy.resolve(item, 'smiles'))
print(smiles_list[-1])
except Exception as e:
try:
print(e)
print('Exception raised. Pausing for another \
2 seconds and trying again')
time.sleep(2)
smiles_list.append(cirpy.resolve(item, 'smiles'))
print(smiles_list[-1])
except Exception as e:
try:
print(e)
print('Exception raised. Pausing for one more \
stretch and trying again')
time.sleep(2)
smiles_list.append(cirpy.resolve(item, 'smiles'))
print(smiles_list[-1])
except Exception as e:
print(e)
print('It still raised an exception. Here\'s \
how far it got:')
print(smiles_list)
print(len(smiles_list))
print('This item will be added to a list \
called missed items.')
print(item)
smiles_list.append('Check')
missed_items.append(item)
already_queried.append(item)
# tidy these up into pandas dataframes and export them as csv files
chem_df = pd.DataFrame(zip(chem_strings, smiles_list),
columns=('Name', 'SMILES'))
chem_df = chem_df.dropna()
chem_df.to_csv(os.path.splitext(pdf_path)[0]+'_'+datetime.today().
strftime('%Y%m%d')+'_names_and_SMILES.csv')
if missed_items:
missed_df = pd.DataFrame(missed_items, columns=['Missed'])
missed_df = missed_df.drop_duplicates()
missed_df.to_csv(os.path.splitext(pdf_path)[0]+'_'+datetime.
today().strftime('%Y%m%d')+'_zzz_missed_items.csv')
def aggregate_csv_files():
# combines all results files into a single csv file
all_chemicals = pd.concat([pd.read_csv(filename) for filename in
os.listdir(pdf_dir) if
re.search('csv$', filename)])
all_chemicals.to_csv(datetime.today().
strftime('%Y%m%d')+"combined_csv.csv",
index=False, encoding='utf-8-sig')
"""# Curate pdfs"""
#@title ## Curator output will appear below
def curate_folder(pdf_dir = os.getcwd()):
"""
Extract known chemicals from a folder of pdf files, and export a
.csv file of SMILESstrings, a machine-readable chemical format
for each file and a combined .csv for all the pdf files.
Extended Summary
----------------
Extract text from a pdf file. Use chemdataextractor's NLP to identify
chemical entities. Attempt to resolve each entity at NIH's CACTVS service.
Organize chemicals recognized by PubChem into a dataframe.
Export the chemical names and SMILES strings as a .csv files
Repeat for each pdf file in the folder
Parameters
----------
pdf_dir : string, optional
path to a folder of pdf files (the default is the current working
directory)
"""
pd.DataFrame(data=None, columns=('Name', 'SMILES'))
assert os.path.exists(pdf_dir), "I did not find the \
directory at, "+str(pdf_dir)
os.chdir(pdf_dir)
for filename in os.listdir(pdf_dir):
if re.search('pdf$', filename):
try:
chemicals = quick_curate(filename, pdf_method,
false_positives, regex_number)
except Exception as e:
print('An exception was raised for ' + filename)
print(e)
try:
aggregate_csv_files()
except Exception as e:
print(e)
"An error occurred while trying to combine the output csv files."