forked from Aveek-Saha/Movie-Script-Database
-
Notifications
You must be signed in to change notification settings - Fork 0
/
parse_files.py
450 lines (392 loc) · 18.1 KB
/
parse_files.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
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
import subprocess
import glob
import os
from os import listdir, makedirs
from os.path import isfile, join, sep, getsize, exists
import numpy as np
import argparse
import re
import time
from tqdm import tqdm
# PROCESS ARGUMENTS
def read_args():
parser = argparse.ArgumentParser(
description='Script that parses a movie script pdf/txt into its constituent classes')
parser.add_argument(
"-i", "--input", help="Path to script PDF/TXT to be parsed", required=True)
parser.add_argument(
"-o", "--output", help="Path to directory for saving output", required=True)
parser.add_argument("-a", "--abridged",
help="Print abridged version (on/off)", default='off')
parser.add_argument(
"-t", "--tags", help="Print class label tags (on/off)", default='off')
parser.add_argument(
"-c", "--char", help="Print char info file (on/off)", default='off')
args = parser.parse_args()
if args.abridged not in ['on', 'off']:
raise AssertionError("Invalid value. Choose either off or on")
if args.tags not in ['on', 'off']:
raise AssertionError("Invalid value. Choose either off or on")
if args.char not in ['on', 'off']:
raise AssertionError("Invalid value. Choose either off or on")
return os.path.abspath(args.input), os.path.abspath(args.output), args.abridged, args.tags, args.char
# READ FILE
def read_file(file_orig):
if file_orig.endswith(".pdf"):
file_name = file_orig.replace('.pdf', '.txt')
subprocess.call('pdftotext -layout ' + file_orig +
' ' + file_name, shell=True)
elif file_orig.endswith(".txt"):
file_name = file_orig
else:
raise AssertionError("File should be either pdf or txt")
fid = open(file_name, 'r')
script_orig = fid.read().splitlines()
fid.close()
if file_orig.endswith(".pdf"):
subprocess.call('rm ' + file_name, shell=True)
return script_orig
# DETECT SCENE BOUNDARIES:
# LOOK FOR ALL-CAPS LINES CONTAINING "INT." OR "EXT."
def get_scene_bound(script_noind, tag_vec, tag_set, bound_set):
bound_ind = [i for i, x in enumerate(script_noind) if tag_vec[i] not in tag_set and
x.isupper() and
any([y in x.lower() for y in bound_set])]
if len(bound_ind) > 0:
for x in bound_ind:
tag_vec[x] = 'S'
return tag_vec, bound_ind
# DETECT TRANSITIONS:
# LOOK FOR ALL-CAPS LINES PRECEDED BY NEWLINE, FOLLOWED BY NEWLINE AND CONTAINING "CUT " OR "FADE "
def get_trans(script_noind, tag_vec, tag_set, trans_thresh, trans_set):
re_func = re.compile('[^a-zA-Z ]')
trans_ind = [i for i, x in enumerate(script_noind) if tag_vec[i] not in tag_set and x.isupper()
and len(re_func.sub('', x).split()) < trans_thresh
and any([y in x for y in trans_set])]
if len(trans_ind) > 0:
for x in trans_ind:
tag_vec[x] = 'T'
return tag_vec, trans_ind
# DETECT METADATA:
# LOOK FOR CONTENT PRECEDING SPECIFIC PHRASES THAT INDICATE BEGINNING OF MOVIE
def get_meta(script_noind, tag_vec, tag_set, meta_thresh, meta_set, sent_thresh, bound_ind, trans_ind):
re_func = re.compile('[^a-zA-Z ]')
met_ind = [i for i, x in enumerate(script_noind) if tag_vec[i] not in tag_set
and i != 0 and i != (len(script_noind) - 1)
and len(x.split()) < meta_thresh
and len(re_func.sub('', script_noind[i - 1]).split()) == 0
and len(re_func.sub('', script_noind[i + 1]).split()) == 0
and any([y in x for y in meta_set])]
sent_ind = [i for i, x in enumerate(script_noind) if tag_vec[i] not in tag_set
and i != 0 and i != (len(script_noind) - 1)
and len(x.split()) > sent_thresh
and len(script_noind[i - 1].split()) == 0
and len(script_noind[i + 1].split()) > 0]
meta_ind = sorted(met_ind + bound_ind + trans_ind + sent_ind)
if len(meta_ind) > 0:
for i, x in enumerate(script_noind[: meta_ind[0]]):
if len(x.split()) > 0:
tag_vec[i] = 'M'
return tag_vec
# DECOMPOSE LINE WITH DIALOGUE AND DIALOGUE METADATA INTO INDIVIDUAL CLASSES
def separate_dial_meta(line_str):
if '(' in line_str and ')' in line_str:
bef_par_str = ' '.join(line_str.split('(')[0].split())
in_par_str = ' '.join(line_str.split('(')[1].split(')')[0].split())
rem_str = ')'.join(line_str.split(')')[1:])
else:
bef_par_str = line_str
in_par_str = ''
rem_str = ''
return bef_par_str, in_par_str, rem_str
# DETECT CHARACTER-DIALOGUE BLOCKS:
# CHARACTER IS ALL-CAPS LINE PRECEDED BY NEWLINE AND NOT FOLLOWED BY A NEWLINE
# DIALOGUE IS WHATEVER IMMEDIATELY FOLLOWS CHARACTER
# EITHER CHARACTER OR DIALOGUE MIGHT CONTAIN DILAOGUE METADATA; WILL BE DETECTED LATER
def get_char_dial(script_noind, tag_vec, tag_set, char_max_words):
char_ind = [i for i, x in enumerate(script_noind) if tag_vec[i] not in tag_set and x.isupper()
and i != 0 and i != (len(script_noind) - 1)\
# and len(script_noind[i - 1].split()) == 0\
and len(script_noind[i + 1].split()) > 0\
and len(x.split()) < char_max_words\
and any([separate_dial_meta(x)[y] for y in [0, 2]])]
for x in char_ind:
tag_vec[x] = 'C'
dial_flag = 1
while dial_flag > 0:
line_ind = x + dial_flag
if len(script_noind[line_ind].split()) > 0:
dial_str, dial_meta_str, rem_str = separate_dial_meta(
script_noind[line_ind])
if dial_str != '' or rem_str != '':
tag_vec[line_ind] = 'D'
else:
tag_vec[line_ind] = 'E'
dial_flag += 1
else:
dial_flag = 0
return tag_vec
# DETECT SCENE DESCRIPTION
# LOOK FOR REMAINING LINES THAT ARE NOT PAGE BREAKS
def get_scene_desc(script_noind, tag_vec, tag_set):
desc_ind = [i for i, x in enumerate(script_noind) if tag_vec[i] not in tag_set and
len(x.split()) > 0 and
not x.strip('.').isdigit()]
for x in desc_ind:
tag_vec[x] = 'N'
return tag_vec
# COMBINE MULTI-LINE CLASSES, SPLIT MULTI-CLASS LINES
def combine_tag_lines(tag_valid, script_valid):
tag_final = []
script_final = []
for i, x in enumerate(tag_valid):
if x in ['M', 'T', 'S']:
# APPEND METADATA, TRANSITION AND SCENE BOUNDARY LINES AS THEY ARE
tag_final.append(x)
script_final.append(script_valid[i])
elif x in ['C', 'D', 'N']:
# IF CHARACTER, DIALOGUE OR SCENE DESCRIPTION CONSIST OF MULTIPLE LINES, COMBINE THEM
if i == 0 or x != tag_valid[i - 1]:
# INITIALIZE IF FIRST OF MULTIPLE LINES
to_combine = []
to_combine += script_valid[i].split()
if i == (len(tag_valid) - 1) or x != tag_valid[i + 1]:
combined_str = ' '.join(to_combine)
if x == 'N':
# IF SCENE DESCRIPTION, WRITE AS IT IS
tag_final.append(x)
script_final.append(combined_str)
else:
_, in_par, _ = separate_dial_meta(combined_str)
if in_par != '':
# IF DIALOGUE METADATA PRESENT, EXTRACT IT
dial_meta_str = ''
char_dial_str = ''
while '(' in combined_str and ')' in combined_str:
before_par, in_par, combined_str = separate_dial_meta(
combined_str)
char_dial_str += ' ' + before_par
dial_meta_str += ' ' + in_par
char_dial_str += ' ' + combined_str
char_dial_str = ' '.join(char_dial_str.split())
dial_meta_str = ' '.join(dial_meta_str.split())
if x == 'C':
# IF CHARACTER, APPEND DIALOGUE METADATA
tag_final.append(x)
script_final.append(
' '.join(char_dial_str.split()))
tag_final.append('E')
script_final.append(dial_meta_str)
elif x == 'D':
# IF DIALOGUE, PREPEND DIALOGUE METADATA
tag_final.append('E')
script_final.append(dial_meta_str)
tag_final.append(x)
script_final.append(
' '.join(char_dial_str.split()))
else:
# IF NO DIALOGUE METADATA, WRITE AS IT IS
tag_final.append(x)
script_final.append(combined_str)
elif x == 'E':
# IF DIALOGUE METADATA LINE, WRITE WITHOUT PARENTHESIS
split_1 = script_valid[i].split('(')
split_2 = split_1[1].split(')')
dial_met = split_2[0]
tag_final.append('E')
script_final.append(dial_met)
return tag_final, script_final
# MERGE CONSECUTIVE IDENTICAL CLASSES
def merge_tag_lines(tag_final, script_final):
merge_ind = [i for i in range(len(tag_final) - 1) if tag_final[i + 1] != 'M' and
(tag_final[i] == tag_final[i + 1])]
merge_ind += [len(tag_final) - 1]
merge_unique = [merge_ind[i] for i in np.where(np.diff(merge_ind) != 1)[0]]
if len(merge_unique) > 0:
tag_merged = []
script_merged = []
last_ind = 0
for merge_ind in range(len(merge_unique)):
# PREPEND PREVIOUS
tag_ind = merge_unique[merge_ind]
tag_merged += tag_final[last_ind: tag_ind]
script_merged += script_final[last_ind: tag_ind]
# MERGE CURRENT
check_tag = tag_final[tag_ind]
tag_end = [i for i, x in enumerate(
tag_final[(tag_ind + 1):]) if x != check_tag][0]
tag_merged += [check_tag]
script_merged += [' '.join(script_final[tag_ind: (tag_ind + tag_end + 1)])]
last_ind = (tag_ind + tag_end + 1)
if tag_ind == merge_unique[-1]:
tag_merged += tag_final[last_ind:]
script_merged += script_final[last_ind:]
else:
tag_merged = tag_final
script_merged = script_final
return tag_merged, script_merged
# REARRANGE DIALOGUE METADATA TO ALWAYS APPEAR AFTER DIALOGUE
def rearrange_tag_lines(tag_merged, script_merged):
dial_met_ind = [i for i in range(1, (len(tag_merged) - 1)) if tag_merged[i] == 'E' and
tag_merged[i - 1] in ['C', 'D'] and
tag_merged[i + 1] == 'D']
if len(dial_met_ind) > 0:
for tag_ind in dial_met_ind:
temp_ind = min([i for i, x in enumerate(
tag_merged[(tag_ind + 1):]) if x != 'D'])
# REARRANGE TAGS
tag_merged[tag_ind: (
tag_ind + temp_ind)] = tag_merged[(tag_ind + 1): (tag_ind + temp_ind + 1)]
tag_merged[tag_ind + temp_ind] = 'E'
# REARRANGE LINES
temp_line = script_merged[tag_ind]
script_merged[tag_ind: (
tag_ind + temp_ind)] = script_merged[(tag_ind + 1): (tag_ind + temp_ind + 1)]
script_merged[tag_ind + temp_ind] = temp_line
return tag_merged, script_merged
# CHECK FOR UN-MERGED CLASSES
def find_same(tag_valid):
same_ind = [i for i in range(len(tag_valid) - 1) if (tag_valid[i] != 'M' and
(tag_valid[i] == tag_valid[i + 1]))]
return len(same_ind)
# CHECK FOR DIALOGUE METADATA PRECEDING DIALOGUE
def find_arrange(tag_valid):
arrange_ind = [i for i in range(len(tag_valid) - 1) if tag_valid[i] == 'E' and
tag_valid[i + 1] == 'D']
return len(arrange_ind)
# PARSER FUNCTION
def parse(file_orig, save_dir, abr_flag, tag_flag, char_flag, save_name=None, abridged_name=None, tag_name=None):
#------------------------------------------------------------------------------------
# DEFINE
tag_set = ['S', 'N', 'C', 'D', 'E', 'T', 'M']
meta_set = ['BLACK', 'darkness']
bound_set = ['int. ', 'ext. ', 'int ', 'ext ']
trans_set = ['CUT ', 'FADE ', 'cut ']
char_max_words = 7
meta_thresh = 2
sent_thresh = 5
trans_thresh = 5
# READ PDF/TEXT FILE
script_orig = read_file(file_orig)
# REMOVE INDENTS
alnum_filter = re.compile('[\W_]+', re.UNICODE)
script_noind = []
for script_line in script_orig:
if len(script_line.split()) > 0 and alnum_filter.sub('', script_line) != '':
script_noind.append(' '.join(script_line.split()))
else:
script_noind.append('')
num_lines = len(script_noind)
tag_vec = np.array(['0' for x in range(num_lines)])
#------------------------------------------------------------------------------------
# DETECT SCENE BOUNDARIES
tag_vec, bound_ind = get_scene_bound(
script_noind, tag_vec, tag_set, bound_set)
# DETECT TRANSITIONS
tag_vec, trans_ind = get_trans(
script_noind, tag_vec, tag_set, trans_thresh, trans_set)
# DETECT METADATA
tag_vec = get_meta(script_noind, tag_vec, tag_set, meta_thresh,
meta_set, sent_thresh, bound_ind, trans_ind)
# DETECT CHARACTER-DIALOGUE
tag_vec = get_char_dial(script_noind, tag_vec, tag_set, char_max_words)
# DETECT SCENE DESCRIPTION
tag_vec = get_scene_desc(script_noind, tag_vec, tag_set)
#------------------------------------------------------------------------------------
# SAVE TAGS TO FILE
if tag_flag == 'on':
if tag_name is None:
tag_name = os.path.join(save_dir, '.'.join(
file_orig.split('/')[-1].split('.')[: -1]) + '_tags.txt')
else:
tag_name = os.path.join(save_dir, tag_name)
np.savetxt(tag_name, tag_vec, fmt='%s', delimiter='\n')
# REMOVE UN-TAGGED LINES
tag_valid = []
script_valid = []
for i, x in enumerate(tag_vec):
if x != '0':
tag_valid.append(x)
script_valid.append(script_noind[i])
# FORMAT TAGS, LINES
tag_valid, script_valid = combine_tag_lines(tag_valid, script_valid)
max_rev = 0
while find_same(tag_valid) > 0 or find_arrange(tag_valid) > 0:
tag_valid, script_valid = merge_tag_lines(tag_valid, script_valid)
tag_valid, script_valid = rearrange_tag_lines(tag_valid, script_valid)
max_rev += 1
if max_rev == 1000:
raise AssertionError(
"Too many revisions. Something must be wrong.")
#------------------------------------------------------------------------------------
# WRITE PARSED SCRIPT TO FILE
if save_name is None:
save_name = os.path.join(save_dir, '.'.join(
file_orig.split('/')[-1].split('.')[: -1]) + '_parsed.txt')
else:
save_name = os.path.join(save_dir, "full", save_name)
fid = open(save_name, 'w')
for tag_ind in range(len(tag_valid)):
_ = fid.write(tag_valid[tag_ind] + ': ' + script_valid[tag_ind] + '\n')
fid.close()
#------------------------------------------------------------------------------------
# CREATE CHARACTER=>DIALOGUE ABRIDGED VERSION
if abr_flag == 'on':
fid = open(save_name, 'r')
parsed_script = fid.read().splitlines()
fid.close()
if abridged_name is None:
abridged_name = os.path.join(save_dir, '.'.join(
file_orig.split('/')[-1].split('.')[: -1]) + '_abridged.txt')
else:
abridged_name = os.path.join(save_dir, "abridged", abridged_name)
abridged_ind = [i for i, x in enumerate(parsed_script) if x.startswith('C:') and
parsed_script[i + 1].startswith('D:')]
fid = open(abridged_name, 'w')
for i in abridged_ind:
char_str = ' '.join(parsed_script[i].split('C:')[1].split())
dial_str = ' '.join(parsed_script[i + 1].split('D:')[1].split())
_ = fid.write(''.join([char_str, '=>', dial_str, '\n']))
fid.close()
#------------------------------------------------------------------------------------
# CREATE CHAR INFO FILE
if char_flag == 'on':
tag_str_vec = np.array(tag_valid)
script_vec = np.array(script_valid)
char_ind = np.where(tag_str_vec == 'C')[0]
char_set = sorted(set(script_vec[char_ind]))
charinfo_vec = []
for char_id in char_set:
spk_ind = list(
set(np.where(script_vec == char_id)[0]) & set(char_ind))
if len(spk_ind) > 0:
num_lines = len([i for i in spk_ind if i != (len(tag_str_vec) - 1) and
tag_str_vec[i + 1] == 'D'])
charinfo_str = char_id + ': ' + \
str(num_lines) + '|'.join([' ', ' ', ' ', ' '])
charinfo_vec.append(charinfo_str)
charinfo_name = os.path.join(save_dir, '.'.join(
file_orig.split('/')[-1].split('.')[: -1]) + '_charinfo.txt')
np.savetxt(charinfo_name, charinfo_vec, fmt='%s', delimiter='\n')
# MAIN FUNCTION
if __name__ == "__main__":
DIR_FINAL = join("scripts", "final")
DIR_OUT = join("scripts", "parsed")
DIR_OUT_FULL = join(DIR_OUT, "full")
DIR_OUT_ABRIDGED = join(DIR_OUT, "abridged")
if not os.path.exists(DIR_OUT):
os.makedirs(DIR_OUT)
if not os.path.exists(DIR_OUT_FULL):
os.makedirs(DIR_OUT_FULL)
if not os.path.exists(DIR_OUT_ABRIDGED):
os.makedirs(DIR_OUT_ABRIDGED)
files = [join(DIR_FINAL, f) for f in listdir(DIR_FINAL)
if isfile(join(DIR_FINAL, f)) and getsize(join(DIR_FINAL, f)) > 3000]
for f in tqdm(files):
file_orig, save_dir, abr_flag, tag_flag, char_flag, save_name, abridged_name = f, DIR_OUT, 'on', 'off', 'off', f.split(
sep)[-1], f.split(sep)[-1].split('.txt')[0] + '_abridged.txt'
try:
parse(file_orig, save_dir, abr_flag,
tag_flag, char_flag, save_name, abridged_name)
except:
pass