File size: 31,404 Bytes
c7e47b2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""Class definitions for MedleyDB multitracks."""

from __future__ import print_function

import csv
import os
import sox
import yaml

from . import INST_TAXONOMY
from . import INST_F0_TYPE
from . import MIXING_COEFFICIENTS
from . import ANNOT_PATH
from . import METADATA_PATH
from . import AUDIO_PATH
from . import TRACK_LIST_V1
from . import TRACK_LIST_V2
from . import TRACK_LIST_EXTRA
from . import TRACK_LIST_BACH10

_YESNO = dict(yes=True, no=False)
_TRACKID_FMT = "%s_%s"
_METADATA_FMT = "%s_METADATA.yaml"
_STEMDIR_FMT = "%s_STEMS"
_RAWDIR_FMT = "%s_RAW"
_MIX_FMT = "%s_MIX.wav"
_STEM_FMT = "%s_STEM_%%s.wav"
_RAW_FMT = "%s_RAW_%%s_%%s.wav"

_AUDIODIR_FMT = "%s"

_ACTIVATION_CONF_PATH = os.path.join(ANNOT_PATH, 'Activation_Confidence')
_ACTIVATION_CONF_ORIG_PATH = os.path.join(_ACTIVATION_CONF_PATH, 'original')
_ACTIVATION_CONF_V2_PATH = os.path.join(_ACTIVATION_CONF_PATH, 'v2')
_MELODY_PATH = os.path.join(ANNOT_PATH, 'Melody')
_INTERVALS_PATH = os.path.join(_MELODY_PATH, 'Intervals')
_MELODY1_PATH = os.path.join(_MELODY_PATH, 'Melody1')
_MELODY2_PATH = os.path.join(_MELODY_PATH, 'Melody2')
_MELODY3_PATH = os.path.join(_MELODY_PATH, 'Melody3')
_RANKING_PATH = os.path.join(_MELODY_PATH, 'Ranking')
_PITCH_PATH = os.path.join(ANNOT_PATH, 'Pitch')
_PITCH_PYIN_PATH = os.path.join(ANNOT_PATH, 'Pitch_Pyin')
_SOURCEID_PATH = os.path.join(ANNOT_PATH, 'Source_ID')

_ANNOTDIR_FMT = "%s_ANNOTATIONS"
_ACTIVCONF_FMT = "%s_ACTIVATION_CONF.lab"
_INTERVAL_FMT = "%s_INTERVALS.txt"
_MELODY1_FMT = "%s_MELODY1.csv"
_MELODY2_FMT = "%s_MELODY2.csv"
_MELODY3_FMT = "%s_MELODY3.csv"
_RANKING_FMT = "%s_RANKING.txt"
_SOURCEID_FMT = "%s_SOURCEID.lab"
_PITCHDIR_FMT = "%s_PITCH"
_PITCH_FMT = "%s.csv"
_PITCH_PYIN_FMT = "%s_vamp_pyin_pyin_smoothedpitchtrack.csv"


class MultiTrack(object):
    """MultiTrack Class definition.

    This class loads all available metadata, annotations, and filepaths for a
    given multitrack directory.

    Parameters
    ----------
    track_id : str
        Track id in format 'Artist_Title'.

    Attributes
    ----------
    artist : str
        The artist of the multitrack
    title : str
        The title of the multitrack
    track_id : str
        The unique identifier of the multitrack. In the form 'Artist_Title'
    annotation_dir : str
        Path to multitrack's annotation directory
    audio_path : str
        Path to multitrack's top level audio directory
    mix_path : str
        Path to multitrack's mix file.
    melody_rankings : dictionary
        Dictionary of melody rankings keyed by stem id
    melody1_fpath : str
        Path to melody 1 annotation file
    melody2_fpath : str
        Path to melody 2 annotation file
    melody3_fpath : str
        Path to melody 3 annotation file
    melody_intervals_fpath : str
        Path to melody intervals file
    melody_rankings_fpath : str
        Path to melody rankings file
    activation_conf_fpath : str
        Path to original activation confidence file
    activation_conf_v2_fpath : str
        Path to version 2 activation confidence file
    source_id_fpath : str
        Path to source id file
    mixing_coefficients : dictionary
        Dictionary of mixing weights keyed by stem id
    stems : dictionary
        Dictionary of stem Track objects keyed by stem id
    raw_audio : dictionary
        Dictionary of dictionaries keyed by stem id
    stem_instruments : list
        List of stem instrument labels
    raw_instruments : list
        List of raw audio instrument labels
    duration : float or None
        Duration of mix, or None if audio cannot be found
    is_excerpt : bool
        True if multitrack is an excerpt
    has_bleed : bool
        True if multitrack has bleed
    is_instrumental : bool
        True if multitrack is instrumental
    origin : str
        Origin of multitrack
    genre : str
        Genre of multitrack
    metadata_version : str
        Metadata version
    has_melody : bool
        True if multitrack has at least one melody stem
    predominant_stem : Track or None
        Track object for the predominant stem if availalbe, otherwise None
    dataset_version : string
        Iteration a multitrack came from.
        (E.g. "V1" for MedleyDB dataset_version 1,
        "V2" for MedleyDB dataset_version 2)
    _stem_activations : np.array
        Matrix of stem activations
    _stem_activations_idx : dictionary
        Dictionary mapping stem index to column of the stem_activations matrix
    _meta_path : str
        Path to metadata file.
    _stem_dir_path : str
        Path to multitrack's stem file directory
    _raw_dir_path : str
        Path to multitrack's raw file directory
    _stem_fmt : str
        Format of stem file basenames
    _raw_fmt : str
        format of raw file basenames
    _metadata : dict
        dictionary of data loaded from metadata file
    _melody1_annotation : np.array or None
        Melody 1 annotation if exists, otherwise None
    _melody2_annotation : np.array or None
        Melody 2 annotation if exists, otherwise None
    _melody3_annotation : np.array or None
        Melody 3 annotation if exists, otherwise None

    Examples
    --------
        >>> mtrack = Multitrack('LizNelson_Rainfall')
        >>> another_mtrack = Multitrack('ArtistName_TrackTitle')

    """

    def __init__(self, track_id):
        """MultiTrack object __init__ method.
        """

        # Artist, Title & Track Directory #
        self.artist = track_id.split('_')[0]
        self.title = track_id.split('_')[1]
        self.track_id = track_id
        self.dataset_version = get_dataset_version(track_id)

        # Filenames and Filepaths #
        self._meta_path = os.path.join(
            METADATA_PATH, _METADATA_FMT % self.track_id
        )

        # break if metadata file cannot be found
        if not os.path.exists(self._meta_path):
            raise IOError("Cannot find metadata for %s" % self.track_id)

        if AUDIO_PATH:
            self.audio_path = os.path.join(
                AUDIO_PATH, _AUDIODIR_FMT % track_id
            )
            self._stem_dir_path = os.path.join(
                self.audio_path, _STEMDIR_FMT % self.track_id
            )
            self._raw_dir_path = os.path.join(
                self.audio_path, _RAWDIR_FMT % self.track_id
            )
            self.mix_path = os.path.join(
                self.audio_path, _MIX_FMT % self.track_id
            )
        else:
            self.audio_path = None
            self._stem_dir_path = None
            self._raw_dir_path = None
            self.mix_path = None

        # Stem & Raw File Formats #
        self._stem_fmt = _STEM_FMT % self.track_id
        self._raw_fmt = _RAW_FMT % self.track_id

        # Yaml Dictionary of Metadata #
        self._metadata = self._load_metadata()

        self.activation_conf_fpath = os.path.join(
            _ACTIVATION_CONF_ORIG_PATH, _ACTIVCONF_FMT % self.track_id
        )
        self.activation_conf_v2_fpath = os.path.join(
            _ACTIVATION_CONF_V2_PATH, _ACTIVCONF_FMT % self.track_id
        )
        self.source_id_fpath = os.path.join(
            _SOURCEID_PATH, _SOURCEID_FMT % self.track_id
        )
        self.melody_intervals_fpath = os.path.join(
            _INTERVALS_PATH, _INTERVAL_FMT % self.track_id
        )
        self.melody_rankings_fpath = os.path.join(
            _RANKING_PATH, _RANKING_FMT % self.track_id
        )
        self.melody_rankings = self._get_melody_rankings()

        if self.track_id in MIXING_COEFFICIENTS:
            self.mixing_coefficients = MIXING_COEFFICIENTS[self.track_id]
        else:
            self.mixing_coefficients = None

        # Stem & Raw Dictionaries. Lists of filepaths. #
        self.stems, self.raw_audio = self._parse_metadata()

        # Lists of Instrument Labels #
        self.stem_instruments = sorted(
            [inst for sublist in self.stems.values()
             for inst in sublist.instrument]
        )
        self.raw_instruments = sorted(
            [inst for sublist in get_dict_leaves(self.raw_audio)
             for inst in sublist.instrument]
        )

        # Basic Track Information #
        self._duration = None
        self.is_excerpt = _YESNO[self._metadata['excerpt']]
        self.has_bleed = _YESNO[self._metadata['has_bleed']]
        self.is_instrumental = _YESNO[self._metadata['instrumental']]
        self.origin = self._metadata['origin']
        self.genre = self._metadata['genre']
        self.metadata_version = self._metadata['version']

        # Melody annotation paths
        self.melody1_fpath = os.path.join(
            _MELODY1_PATH, _MELODY1_FMT % self.track_id
        )
        self.melody2_fpath = os.path.join(
            _MELODY2_PATH, _MELODY2_FMT % self.track_id
        )
        self.melody3_fpath = os.path.join(
            _MELODY3_PATH, _MELODY3_FMT % self.track_id
        )

        self.has_melody = os.path.exists(self.melody1_fpath)

        self._melody1_annotation = None
        self._melody2_annotation = None
        self._melody3_annotation = None

        self.predominant_stem = self._get_predominant_stem()
        self._stem_activations = None
        self._stem_activations_idx = None
        self._stem_activations_v2 = None
        self._stem_activations_idx_v2 = None

    @property
    def duration(self):
        """float: Duration of the mix
        """
        if self._duration is None:
            if self.mix_path is not None and os.path.exists(self.mix_path):
                self._duration = get_duration(self.mix_path)

        return self._duration

    @property
    def melody1_annotation(self):
        """np.array: Melody 1 annotation.
        """
        if self._melody1_annotation is None:
            self._melody1_annotation, _ = read_annotation_file(
                self.melody1_fpath, header=False
            )
        return self._melody1_annotation

    @property
    def melody2_annotation(self):
        """np.array: Melody 2 annotation.
        """
        if self._melody2_annotation is None:
            self._melody2_annotation, _ = read_annotation_file(
                self.melody2_fpath, header=False
            )
        return self._melody2_annotation

    @property
    def melody3_annotation(self):
        """np.array: Melody 3 annotation.
        """
        if self._melody3_annotation is None:
            self._melody3_annotation, _ = read_annotation_file(
                self.melody3_fpath, header=False
            )
        return self._melody3_annotation

    @property
    def stem_activations(self):
        """np.array: Matrix of stem activations
        """
        if self._stem_activations is None:
            (self._stem_activations,
             self._stem_activations_idx) = self._get_activation_annotations()
        return self._stem_activations

    @property
    def stem_activations_idx(self):
        """dictionary : Dictionary mapping stem index to column of the
        stem_activations matrix.
        """
        if self._stem_activations_idx is None:
            (self._stem_activations,
             self._stem_activations_idx) = self._get_activation_annotations()
        return self._stem_activations_idx

    @property
    def stem_activations_v2(self):
        """np.array: Matrix of stem activations (annotations version 2)
        """
        if self._stem_activations_v2 is None:
            (self._stem_activations_v2,
             self._stem_activations_idx_v2
            ) = self._get_activation_annotations(version='v2')
        return self._stem_activations_v2

    @property
    def stem_activations_idx_v2(self):
        """dictionary : Dictionary mapping stem index to column of the
        stem_activations matrix. (annotations version 2)
        """
        if self._stem_activations_idx_v2 is None:
            (self._stem_activations_v2,
             self._stem_activations_idx_v2
            ) = self._get_activation_annotations(version='v2')
        return self._stem_activations_idx_v2

    def _load_metadata(self):
        """Load the metadata file.

        Returns
        -------
        metadata : dict
            Dictionary of data read directly from the YAML metadata file.

        """
        with open(self._meta_path, 'r') as f_in:
            metadata = yaml.load(f_in)
        return metadata

    def _parse_metadata(self):
        """Parse metadata dictionary.

        Returns
        -------
        stems : dict
            Dictionary of Track objects keyed by stem_id
        raw_audio : dict
            Dictionary of dictionaries of Track objects keyed by stem_id
            and raw_id

        """
        stems = dict()
        raw_audio = dict()
        stem_dict = self._metadata['stems']

        for k in stem_dict:
            stem_idx = int(k[1:])

            instrument = stem_dict[k]['instrument']
            component = stem_dict[k]['component']

            if stem_idx in self.melody_rankings:
                ranking = self.melody_rankings[stem_idx]
            else:
                ranking = None

            if AUDIO_PATH:
                file_name = stem_dict[k]['filename']
                audio_path = os.path.join(self._stem_dir_path, file_name)
            else:
                audio_path = None

            file_id = "%s_STEM_%s" % (self.track_id, k[1:])

            mixing_coefficient = self._get_mixing_coefficient(stem_idx)

            track = Track(instrument=instrument, audio_path=audio_path,
                          component=component, stem_idx=stem_idx,
                          ranking=ranking, mix_path=self.mix_path,
                          file_id=file_id,
                          mixing_coefficient=mixing_coefficient)

            stems[stem_idx] = track
            raw_dict = stem_dict[k]['raw']

            for j in raw_dict:
                raw_idx = int(j[1:])
                instrument = raw_dict[j]['instrument']

                if AUDIO_PATH:
                    file_name = raw_dict[j]['filename']
                    audio_path = os.path.join(self._raw_dir_path, file_name)
                else:
                    audio_path = None

                track = Track(instrument=instrument, audio_path=audio_path,
                              stem_idx=stem_idx, raw_idx=raw_idx,
                              mix_path=self.mix_path, ranking=ranking)
                if stem_idx not in raw_audio:
                    raw_audio[stem_idx] = {}

                raw_audio[stem_idx][raw_idx] = track

        return stems, raw_audio

    def _get_mixing_coefficient(self, stem_idx):
        """Get best availalbe mixing coefficient for a stem.

        Parameters
        ----------
        stem_idx : int
            Stem index

        Returns
        -------
        mixing_coefficient : float
            Stem's mixing coefficient

        """
        if self.mixing_coefficients is not None:
            use_manual = (
                'manual' in self.mixing_coefficients.keys() and
                self.mixing_coefficients['manual'] != {}
            )

            if use_manual:
                mixing_coefficient = (
                    self.mixing_coefficients['manual'][stem_idx]
                )
            else:
                mixing_coefficient = (
                    self.mixing_coefficients['audio'][stem_idx] +
                    self.mixing_coefficients['stft'][stem_idx]
                ) * 0.5

        else:
            mixing_coefficient = None

        return mixing_coefficient

    def _get_melody_rankings(self):
        """Get rankings from the melody rankings annotation file.

        Returns
        -------
        melody_rankings : dict
            Dictonary of melody rankings keyed by stem_id

        """
        melody_rankings = {}
        if os.path.exists(self.melody_rankings_fpath):
            with open(self.melody_rankings_fpath) as f_handle:
                linereader = csv.reader(f_handle)
                for line in linereader:
                    stem_idx = int(line[0].split('_')[-1].split('.')[0])
                    ranking = int(line[1])
                    melody_rankings[stem_idx] = ranking
        return melody_rankings

    def _get_predominant_stem(self):
        """Get predominant stem if files exists.

        Returns
        -------
        predominant_stem : Track or None
            If a predominant stem is labeled, returns the Track object
            corresponding to the predominant stem. Otherwise None.

        """

        if len(self.melody_rankings) > 0:
            predominant_idx = [
                k for k, v in self.melody_rankings.items() if v == 1
            ]
            if len(predominant_idx) > 0:
                predominant_idx = predominant_idx[0]
                return self.stems[predominant_idx]
            else:
                return None
        else:
            return None

    def _get_activation_annotations(self, version=None):
        """Get activation confidence annotation if file exists.

        Parameters
        ----------
        version : str
            If 'v2', uses the version 2 annotations. Otherwise uses acitvation
            annotations from the original release.

        Returns
        -------
        activations : list
            List of lists of activation confidences
        idx_dict : dict
            Dictionary of column ids in the activations table keyed by stem_id

        """
        if version == 'v2':
            fpath = self.activation_conf_v2_fpath
        else:
            fpath = self.activation_conf_fpath

        if os.path.exists(fpath):
            activations, header = read_annotation_file(
                fpath, header=True
            )
            idx_dict = {}
            for i, stem_str in enumerate(header):
                if stem_str == 'time':
                    continue
                else:
                    stem_idx = format_index(stem_str)
                    idx_dict[stem_idx] = i
            return activations, idx_dict
        else:
            return None, None

    def melody_stems(self):
        """Get list of stems that contain melody.

        Returns
        -------
        melody_stems : list
            List of Track objects where component='melody'.

        """
        stem_objects = self.stems.values()
        return [track for track in stem_objects if track.component == 'melody']

    def bass_stems(self):
        """Get list of stems that contain bass.

        Returns
        -------
        bass_stems: list
            List of Track objects where component='bass'.

        """
        stem_objects = self.stems.values()
        return [track for track in stem_objects if track.component == 'bass']

    def num_stems(self):
        """Number of stems.

        Returns
        -------
        n_stems : int
            Number of stems.

        """
        return len(self.stems)

    def num_raw(self):
        """Number of raw audio files.

        Returns
        -------
        n_raw : int
            Number of raw audio files.

        """
        return len(get_dict_leaves(self.raw_audio))

    def stem_filepaths(self):
        """Get list of filepaths to stem files.

        Returns
        -------
        stem_fpaths : list
            List of filepaths to stems.

        """
        return [track.audio_path for track in self.stems.values()]

    def raw_filepaths(self):
        """Get list of filepaths to raw audio files.

        Returns
        -------
        raw_fpaths : list
            List of filepaths to raw audio files.

        """
        return [track.audio_path for track in get_dict_leaves(self.raw_audio)]

    def activation_conf_from_stem(self, stem_idx, version=None):
        """Get activation confidence from given stem.

        Parameters
        ----------
        stem_idx : int
            stem index (eg. 2 for stem S02)
        version : str
            If 'v2', uses the version 2 annotations. Otherwise uses acitvation
            annotations from the original release.

        Returns
        -------
        activation_confidence : list
            List of time, activation confidence pairs

        """
        if version == 'v2':
            activation_idx = self.stem_activations_idx_v2
            stem_activations = self.stem_activations_v2
        else:
            activation_idx = self.stem_activations_idx
            stem_activations = self.stem_activations

        activations = []
        if stem_idx in activation_idx:
            activ_conf_idx = activation_idx[stem_idx]
            for step in stem_activations:
                activations.append([step[0], step[activ_conf_idx]])
        else:
            activations = None

        return activations


class Track(object):
    """Track class definition.
    Used for stems and for raw audio tracks.

    Parameters
    ----------
    instrument : list of str
        The track's instrument label(s).
    audio_path : str
        Path to corresponding audio file.
    stem_idx : int or str
        stem index, either as int or str
        For ArtistName_TrackTitle_STEM_05.wav, either 5 or 'S05'
    mix_path : str
        Path to corresponding mix audio file.
    file_id : str or None, default=None
        The file's basename - e.g. Artist_Track_STEM_01
    raw_idx : int str or None, default=None
        Raw index, either as int or str
        For ArtistName_TrackTitle_RAW_05_02.wav, either 2 or 'R02'
    component : str, default=''
        stem's component label, if exists.
    ranking : int or None, default=None
        The Track's melodic ranking
    mixing_coefficient : float or None, default=None
        The Tracks's mixing coefficient

    Attributes
    ----------
    instrument : str
        The track's instrument label
    f0_type : str
        The track's f0 type. One of
            - 'm' for monophonic sources
            - 'p' for polyphonic sources
            - 'u' for unpitched sources
    audio_path : str
        Path to corresponding audio file
    component : str or None
        The Track's component label, if exists
        E.g. 'melody', 'bass'
    ranking : int or None
        The Track's melodic ranking, if exists
    stem_idx : int
        The Track's stem index
    raw_idx : int or None
        The Track's raw index, if exists
    mixing_coefficient : float or None
        The Tracks's mixing coefficient, if exists
    duration : float or None
        The track's duration in seconds, if the audio is availalbe
    mix_path : str
        The path to the track's corresponding mix
    pitch_path : str
        The path to the track's pitch annotation
    pitch_annotation : list or None
        List of pairs of time (seconds), frequency (Hz)
    _pitch_annotation : list or None
        List of pairs of time (seconds), frequency (Hz)

    """
    def __init__(self, instrument, audio_path, stem_idx, mix_path,
                 file_id=None, raw_idx=None, component='', ranking=None,
                 mixing_coefficient=None):
        """Track object __init__ method.
        """
        if isinstance(instrument, list):
            self.instrument = instrument
        else:
            self.instrument = [instrument]

        self.f0_type = [get_f0_type(inst) for inst in self.instrument]

        self.audio_path = audio_path
        self.component = component
        self.ranking = ranking
        self.stem_idx = format_index(stem_idx)
        self.raw_idx = format_index(raw_idx)

        if file_id is not None:
            self.pitch_path = os.path.join(_PITCH_PATH, _PITCH_FMT % file_id)
            self.pitch_pyin_path = os.path.join(
                _PITCH_PYIN_PATH, _PITCH_PYIN_FMT % file_id
            )
        else:
            self.pitch_path = None
            self.pitch_pyin_path = None

        self._duration = None
        self.mixing_coefficient = mixing_coefficient

        self.mix_path = mix_path
        self._pitch_annotation = None
        self._pitch_estimate_pyin = None

    @property
    def duration(self):
        """float: duration of audio file
        """
        if self._duration is None:
            if self.audio_path is not None and os.path.exists(self.audio_path):
                self._duration = get_duration(self.audio_path)
        return self._duration

    @property
    def pitch_annotation(self):
        """list: List of pairs of time (seconds), frequency (Hz)
        """
        if self._pitch_annotation is None:
            self._pitch_annotation, _ = read_annotation_file(
                self.pitch_path, num_cols=2, header=False
            )
        return self._pitch_annotation

    @property
    def pitch_estimate_pyin(self):
        """list: List of pairs of time (seconds), frequency (Hz)
        """
        if self._pitch_estimate_pyin is None:
            self._pitch_estimate_pyin, _ = read_annotation_file(
                self.pitch_pyin_path, num_cols=2, header=False
            )
        return self._pitch_estimate_pyin

    def __eq__(self, other):
        return self.__dict__ == other.__dict__

    def __ne__(self, other):
        return self.__dict__ != other.__dict__

    def __hash__(self):
        return hash((tuple(self.instrument),
                     self.audio_path,
                     self.component,
                     self.stem_idx,
                     self.raw_idx,
                     self.mix_path,
                     self.pitch_path))


def get_f0_type(instrument):
    """Get the f0 type of an instrument.

    Parameters
    ----------
    instrument : str
        Instrument label, e.g. 'flute'

    Returns
    -------
    f0_type : str
        The instrument's f0 type. One of
            - 'm' for monophonic sources
            - 'p' for polyphonic sources
            - 'u' for unpitched sources

    """
    if instrument in set(INST_F0_TYPE.keys()):
        return INST_F0_TYPE[instrument]
    else:
        return "?"


def _path_basedir(path):
    """Get the name of the lowest directory of a path.

    Parameters
    ----------
    path : str
        A file path.

    Returns
    -------
    basedir : str
        The base directory of a path

    """
    return os.path.basename(os.path.normpath(path))


def format_index(index):
    """Load stem or raw index. Reformat if in string form.

    Parameters
    ----------
    index : int or str
        Index in string or integer form.
        E.g. any of 1 or 'S01' or 'R01'

    Returns
    -------
    formatted_index : int
        Index in integer form

    """
    if isinstance(index, str):
        return int(index.strip('S').strip('R'))
    elif index is None:
        return None
    else:
        return int(index)


def get_dict_leaves(dictionary):
    """Get the set of all leaves of a dictionary.

    Parameters
    ----------
    dictionary : dict
        A dictionary or nested dictionary.

    Returns
    -------
    vals : set
        Set of leaf values.

    """
    vals = []
    for k in dictionary:
        if isinstance(dictionary[k], dict):
            for val in get_dict_leaves(dictionary[k]):
                vals.append(val)
        else:
            if hasattr(dictionary[k], '__iter__'):
                for val in dictionary[k]:
                    vals.append(val)
            else:
                vals.append(dictionary[k])

    return set(vals)


def get_duration(wave_fpath):
    """Get the duration of a wave file, in seconds.

    Parameters
    ----------
    wave_fpath : str
        Wave file.

    Returns
    -------
    duration : float
        Duration of wave file in seconds.

    """
    n_samples = float(sox.file_info.num_samples(wave_fpath))
    sample_rate = float(sox.file_info.sample_rate(wave_fpath))
    return n_samples / sample_rate


def read_annotation_file(fpath, num_cols=None, header=False):
    """Read an annotation file.
    The returned annotations can be directly converted to a numpy array,
    if desired.

    When reading files generated by Tony, set num_cols=2.
    Annotation files created by Tony can contain a third column that
    sometimes has a value (e.g [2]) and sometimes does not. It isn't
    important for annotation and can be ignored.

    Parameters
    ----------
    fpath : str
        Path to annotation file.
    num_cols : int or None, default=None
        Number of columns to read. If specified,
        will only read the return num_cols columns of the annotation file.

    Returns
    -------
    annotation : list
        List of rows of the annotation file.
    header : list
        Header row. Empty list if header=False.

    Examples
    --------
    >>> melody_fpath = 'ArtistName_TrackTitle_MELODY1.txt'
    >>> pitch_fpath = 'my_tony_pitch_annotation.csv'
    >>> melody_annotation, _ = read_annotation_file(melody_fpath)
    >>> activation_annotation, header = read_annotation_file(
            actvation_fpath, header=True
        )
    >>> pitch_annotation, _ = read_annotation_file(pitch_fpath, num_cols=2)

    """
    if fpath is not None and os.path.exists(fpath):
        with open(fpath) as f_handle:
            annotation = []
            linereader = csv.reader(f_handle)

            # skip the headers for non csv files
            if header:
                header = next(linereader)
            else:
                header = []

            for line in linereader:
                if num_cols:
                    line = line[:num_cols]
                annotation.append([float(val) for val in line])
        return annotation, header
    else:
        return None, None


def get_valid_instrument_labels(taxonomy=INST_TAXONOMY):
    """Get set of valid instrument labels based on a taxonomy.

    Parameters
    ----------
    taxonomy_file : str, default=INST_TAXONOMY
        Path to instrument taxonomy file.

    Returns
    -------
    valid_instrument_labels : set
        Set of valid instrument labels.

    Examples
    --------
    >>> valid_labels = get_valid_instrument_labels()
    >>> my_valid_labels = get_valid_instrument_labels('my_taxonomy.yaml')

    """
    valid_instrument_labels = get_dict_leaves(taxonomy)
    return valid_instrument_labels


def is_valid_instrument(instrument):
    """Test if an instrument is valid based on a taxonomy.
    This is case sensitive! Taxonomy instrument labels are all lowercase.

    Parameters
    ----------
    instrument : str
        Input instrument.

    Returns
    -------
    value : bool
        True if instrument is valid.

    Examples
    --------
    >>> is_valid_instrument('clarinet')
    True
    >>> is_valid_instrument('Clarinet')
    False
    >>> is_valid_instrument('mayonnaise')
    False

    """
    return instrument in get_valid_instrument_labels()


def get_dataset_version(track_id):
    """Get the dataset version a track id appears in.

    Parameters
    ----------
    track_id : str
        Track id

    Returns
    -------
    dataset_version : str
        The dataset version string

    """
    if track_id in TRACK_LIST_V1:
        dataset_version = 'V1'
    elif track_id in TRACK_LIST_V2:
        dataset_version = 'V2'
    elif track_id in TRACK_LIST_EXTRA:
        dataset_version = 'EXTRA'
    elif track_id in TRACK_LIST_BACH10:
        dataset_version = 'BACH10'
    else:
        dataset_version = ''

    return dataset_version