forked from worldveil/dejavu
-
Notifications
You must be signed in to change notification settings - Fork 7
/
fingerprintWorker.py
392 lines (302 loc) · 12.5 KB
/
fingerprintWorker.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
import audioHelper as hlp
import os
import numpy as np
import time
import math
from fingerprint import Fingerprint
from audioHelper import AudioHelper
from datasketch import MinHash
import pickle
import os
##### SUPPORTED AUDIO FORMATS #####
VALID_EXT = ['.wav', '.ogg', '.mp3', '.flac', '.grid', '.mpeg']
CUSTOM_EXT = '.grid'
class Worker(object):
def __init__(self, db):
self.fgp_db = db
self.fgp_api = Fingerprint()
def mic_recognize(self, limit=None):
if limit is None:
limit = 10
print('Microphone listening for: {} seconds'.format(limit))
self.mic = AudioHelper()
result = set()
mic_data = self.mic.recognize(limit=limit)
for num_channels, channel in enumerate(mic_data):
hashes = self.fgp_api.fingerprint(channel, frame_rate=self.mic.samplerate, verbose=True, plot=True)
result |= set(hashes)
return result
def fingerprint_worker(self, file_path, limit=None, grid_only=False, verbose=False, plot=False):
#st = time.time()
song_name, extension = os.path.splitext(file_path)
# print('Fingerprinting: ', song_name, '\nFile extension: ', extension)
# using different extraction method for mp3
if extension is '.mp3' or '.mpeg':
# print(file_path)
num_channels, frame_rate, audio_data = hlp.retrieve_audio_mpeg(file_path, limit)
else:
num_channels, frame_rate, audio_data = hlp.retrieve_audio(file_path, limit)
#print('from fingerprint worker\n frame rate {}, data {}'.format(frame_rate, channels))
result = set()
for num_channels, channel in enumerate(audio_data):
# print('Channel number:', num_channels+1)
hashes = self.fgp_api.fingerprint(channel, frame_rate=frame_rate, verbose=verbose, plot=plot)
if grid_only:
return self.fgp_api.fingerprint(channel, frame_rate=frame_rate, grid_only=grid_only, plot=plot)
result |= set(hashes)
#ft = time.time() - st
#print('Elapsed fingerprinting time: ', ft)
#print('Generated {} hashes'.format(len(result)))
return song_name, result
def insert_wav_to_db(self, song_n):
#db.connect()
song_name, list_hash = self.fingerprint_worker(song_n, limit=None)
print('Song name: ', song_name)
print('Number of generated hashes: ', len(list_hash))
self.fgp_db.insert_song(song_name, 1)
for h in list_hash:
self.fgp_db.insert_fingerprint(h[0], song_name, h[1])
def get_max_track_frequency(self, list_tracks):
"""Interates through a list of tuples (track, frequency of track) and returns the maximum value"""
max_t_frequ = 0
for t in list_tracks.keys():
if list_tracks[t] > max_t_frequ:
max_t_frequ = list_tracks[t]
return max_t_frequ
def align_matches_weighted(self, list_matches):
candidates = dict()
for tup in list_matches:
track_name, time_delta = tup
if time_delta not in candidates:
candidates[time_delta] = dict()
if track_name not in candidates[time_delta]:
candidates[time_delta][track_name] = 1
else:
candidates[time_delta][track_name] += 1
weighted_candidates = []
# each candidate is a tuple of (weight, (k,v))
# default weight = 1
# formula = (e ^ -(|time_delta|)) + max time delta value over a candidate list
for k, v in candidates.items():
cand_weight = float(math.e ** (-abs(k))) * 1000
max_t_freq = self.get_max_track_frequency(v)
cand_tup = (cand_weight + max_t_freq, k, v)
weighted_candidates.append(cand_tup)
weighted_candidates = sorted(weighted_candidates, key=lambda weight: weight[0])
res = [elem for elem in weighted_candidates if elem[0] > 100.0]
# escape case where list of candidates is empty
if len(res) == 0:
return {'song id': 0,
'song name': 'No results found',
'is fingerprinted': 0}, candidates, res
prime_candidate = res[-1]
prime_weight = prime_candidate[0]
max_count = 0
query_track = ''
# query the track with most hits
for k, v in prime_candidate[2].items():
if v > max_count:
max_count = v
query_track = k
query_hit, id, name, is_fng = self.fgp_db.get_song_by_name(query_track)
# cut-off weight for candidates
CUT_OFF_WEIGHT_1 = 368.87944117144235
CUT_OFF_WEIGHT_2 = 1010
if prime_weight <= CUT_OFF_WEIGHT_2 and max_count <= 10:
track = {
'song id': 0,
'song name': 'No results found',
'is fingerprinted': 0,
}
return track, candidates, res
track = {
'song id': id,
'song name': name,
'is fingerprinted': int(is_fng),
}
return track, candidates, res
def fingerprint_songs(self, user_path='', num_tracks=None):
dir_structure = self.build_dir_map(user_path)
# get fingerprinted files
number_fgp, already_fingerprinted = self.get_wavs_by_fgp(1)
#print(already_fingerprinted)
#print('Number of fingerprints=', number_fgp)
song_counter = 0
# go through each file in the directory
for file in dir_structure.keys():
# don't re-fingerprint files
if file in already_fingerprinted:
print('Skipping: {}'.format(file))
continue
if song_counter == num_tracks:
print('Added {} tracks to database.'.format(song_counter))
self.fgp_db.connection.close()
return
# path of dir + actual file
path = dir_structure[file] + '\\' + file
# avoid invalid extensions
_pth, ext = os.path.splitext(path)
if ext not in VALID_EXT:
continue
# insert song returns true if it managed, false otherwise
res = self.fgp_db.insert_song(file, 1)
if res:
song_counter += 1
# generate and insert hashes
_, list_hashes = self.fingerprint_worker(path)
formatted_list = []
for h in list_hashes:
# db.insert_fingerprint(h[0], file, h[1])
formatted_list.append((h[0], file, h[1]))
res = self.fgp_db.dump_fingerprints(formatted_list)
# stop everything in case of failure
if not res:
self.fgp_db.delete_songs([file])
print('Fingerprinting failed for: {}'.format([file]))
return
else:
print('Fingerprinting skipped')
continue
print('Number of wavs: ', song_counter)
def get_wavs_by_fgp(self, is_fgp=0):
res = list(self.fgp_db.get_songs_by_fgp_status(is_fgp))
clean_list = []
for elem in res:
temp = str(elem)[2:-3]
clean_list.append(temp)
# print(clean_list)
number_of_tracks = len(clean_list)
return number_of_tracks, clean_list
######################################################################
#
# GRIDHASH ALGORITHM
#
######################################################################
##### DIRECTORY STRUCTURE METHODS #####
def _get_dir_structure(self, dir_path):
"""Returns all files from a specified directory"""
files = []
for (dirpath, dirname, filenames) in os.walk(dir_path):
files.append([dirpath, filenames])
return files
def has_valid_extension(self, path_to_file):
"""Checks if file extension is valid
Valid extensions: '.wav', '.ogg', '.mp3', '.flac', '.grid', '.mpeg'
"""
path, ext = os.path.splitext(path_to_file)
if ext in VALID_EXT:
return True
return False
def build_dir_map(self, root):
"""creates a dictionary directory structure.
It maps files to their relative path.
file.wav -> c//dir/dir2/dir_with_wavs
Attributes:
root - where to start looking
Return:
map - dictionary structure
"""
dir_struct = self._get_dir_structure(root)
map = dict()
for tup in dir_struct:
current_directory = tup[0]
files_in_dir = tup[1]
for f in files_in_dir:
path = os.path.join(current_directory, f)
# add key if not already in dict and if file has a valid extension
if f not in map and self.has_valid_extension(path):
map[f] = current_directory
return map
##### IO METHODS #####
def export_file(self, file_name, data, dest_dir=''):
"""Stores gridHash file to specified location
Attributes:
file_name - name of file
data - information to package to the file
dest_dir - file path
"""
name = file_name[:-4] + CUSTOM_EXT
path = os.path.join(dest_dir, name)
with open(path, mode='wb') as f:
try:
min_data = self.get_minHash(data)
pickle.dump(min_data, f)
f.close()
print('Exported: {}'.format(name))
return True
except:
print('Export failed: {}'.format(name))
return False
def load_grid(self, file_name, local_dir=''):
"""Loads gridHash file from specified location.
Attributes:
file_name - name of file to load
local_dir - load path
Return:
data - retrieved information
"""
path = os.path.join(local_dir, file_name)
filename, ext = os.path.splitext(path)
if ext != CUSTOM_EXT:
path = path[:-len(ext)] + CUSTOM_EXT
with open(path, 'rb') as f:
data = pickle.load(f)
return data
##### minHash generators ######
def get_minHash(self, input_set):
"""Generates minHash object from input set
Attributes:
input_set - list of strings to minHash
Returns:
minHash object
"""
min_h = MinHash()
for itm in input_set:
min_h.update(itm.encode('utf8'))
return min_h
def export_many(self, files_in, files_out, limit=0):
"""Exports multiple gridHash objects"""
# initialize counter for files to be indexed
counter = 0
# build directory maps
dir_map = self.build_dir_map(files_in)
indexed = self.build_dir_map(files_out)
# if no number of files is specified, process all files
if limit == 0:
limit = len(dir_map.keys())
print('Info:\n',
'There are {} available audio files.\n'.format(len(dir_map.keys())),
'There are {} available gridHash files.\n'.format(len(indexed.keys()))
)
# go file by file
for tr in dir_map.keys():
if counter < limit:
# check if the file has not already been exported
pre = tr[:-4] + CUSTOM_EXT
if pre not in indexed.keys():
_path = os.path.join(dir_map[tr], tr)
# ensure a valid extension
if self.has_valid_extension(_path):
set_data = self.fingerprint_worker(_path, grid_only=True, plot=False)
#print(tr, set_data)
# generate gridhash
res = self.export_file(tr, set_data, dest_dir=files_out)
if res:
counter += 1
else:
return
else:
print('Skipping: {} file already exists'.format(tr))
print('Exported {} grids'.format(counter))
def compute_jaccard(self, s1, s2, grid_folder):
"""Computes jaccard distance between two gridHash files"""
dir_map = self.build_dir_map(grid_folder)
c1 = None
c2 = None
for itm in dir_map.keys():
if itm == s1:
c1 = self.load_grid(itm, local_dir=grid_folder)
if itm == s2:
c2 = self.load_grid(itm, local_dir=grid_folder)
sim = c1.jaccard(c2)
return sim