los_angeles_midi_dataset_metadata_maker.py

# -*- coding: utf-8 -*-
"""Los_Angeles_MIDI_Dataset_Metadata_Maker.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/github/asigalov61/Los-Angeles-MIDI-Dataset/blob/main/META-DATA/Los_Angeles_MIDI_Dataset_Metadata_Maker.ipynb

# Los Angeles MIDI Dataset Metadata Maker (ver. 3.1)

***

Powered by tegridy-tools: https://github.com/asigalov61/tegridy-tools

***

#### Project Los Angeles

#### Tegridy Code 2023

***

# (SETUP ENVIRONMENT)
"""

#@title Install all dependencies (run only once per session)

!git clone https://github.com/asigalov61/tegridy-tools
!pip install tqdm

#@title Import all needed modules

print('Loading needed modules. Please wait...')
import os

import math
import statistics
import random
from collections import Counter
import pickle

from tqdm import tqdm

if not os.path.exists('/content/Dataset'):
    os.makedirs('/content/Dataset')

print('Loading TMIDIX module...')
os.chdir('/content/tegridy-tools/tegridy-tools')

import TMIDIX

print('Done!')

os.chdir('/content/')
print('Enjoy! :)')

"""# (DOWNLOAD SOURCE MIDI DATASET)"""

# Commented out IPython magic to ensure Python compatibility.
#@title Download original LAKH MIDI Dataset

# %cd /content/Dataset/

!wget 'http://hog.ee.columbia.edu/craffel/lmd/lmd_full.tar.gz'
!tar -xvf 'lmd_full.tar.gz'
!rm 'lmd_full.tar.gz'

# %cd /content/

#@title Mount Google Drive
from google.colab import drive
drive.mount('/content/drive')

"""# (FILE LIST)"""

#@title Save file list
###########

print('Loading MIDI files...')
print('This may take a while on a large dataset in particular.')

dataset_addr = "/content/Dataset"
# os.chdir(dataset_addr)
filez = list()
for (dirpath, dirnames, filenames) in os.walk(dataset_addr):
    filez += [os.path.join(dirpath, file) for file in filenames]
print('=' * 70)

if filez == []:
    print('Could not find any MIDI files. Please check Dataset dir...')
    print('=' * 70)

print('Randomizing file list...')
random.shuffle(filez)

TMIDIX.Tegridy_Any_Pickle_File_Writer(filez, '/content/filez')

#@title Load file list
filez = TMIDIX.Tegridy_Any_Pickle_File_Reader('/content/filez')
print('Done!')

"""# (PROCESS)"""

#@title Process MIDIs with TMIDIX MIDI processor

print('=' * 70)
print('TMIDIX MIDI Processor')
print('=' * 70)
print('Starting up...')
print('=' * 70)

###########

START_FILE_NUMBER = 0
LAST_SAVED_BATCH_COUNT = 0

input_files_count = START_FILE_NUMBER
files_count = LAST_SAVED_BATCH_COUNT

melody_chords_f = []

stats = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]

print('Processing MIDI files. Please wait...')
print('=' * 70)

for f in tqdm(filez[START_FILE_NUMBER:]):
    try:
        input_files_count += 1

        fn = os.path.basename(f)
        fn1 = fn.split('.mid')[0]

        #=======================================================
        # START PROCESSING

        opus = TMIDIX.midi2opus(open(f, 'rb').read())

        opus_events_matrix = []

        itrack0 = 1

        while itrack0 < len(opus):
            for event in opus[itrack0]:
                    opus_events_matrix.append(event)
            itrack0 += 1

        #=======================================================

        ms_score = TMIDIX.opus2score(TMIDIX.to_millisecs(opus))

        ms_events_matrix = []

        itrack1 = 1

        while itrack1 < len(ms_score):
            for event in ms_score[itrack1]:
                if event[0] == 'note':
                    ms_events_matrix.append(event)
            itrack1 += 1

        ms_events_matrix.sort(key=lambda x: x[1])

        #=======================================================

        # Convering MIDI to score with MIDI.py module
        score = TMIDIX.opus2score(opus)

        # INSTRUMENTS CONVERSION CYCLE

        events_matrix = []
        full_events_matrix = []

        itrack = 1
        patches = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]

        while itrack < len(score):
            for event in score[itrack]:
                if event[0] == 'note' or event[0] == 'patch_change':
                    events_matrix.append(event)
                full_events_matrix.append(event)
            itrack += 1

        full_events_matrix.sort(key=lambda x: x[1])
        events_matrix.sort(key=lambda x: x[1])

        events_matrix1 = []

        for event in events_matrix:
            if event[0] == 'patch_change':
                patches[event[2]] = event[3]

            if event[0] == 'note':
                event.extend([patches[event[3]]])
                events_matrix1.append(event)

        if len(events_matrix1) > 32:

            events_matrix1.sort(key=lambda x: x[1])

            for e in events_matrix1:
                if e[0] == 'note':
                    if e[3] == 9:
                        e[4] = ((abs(e[4]) % 128) + 128)
                    else:
                        e[4] = (abs(e[4]) % 128)

            pitches_counts = [[y[0],y[1]] for y in Counter([y[4] for y in events_matrix1]).most_common()]
            pitches_counts.sort(key=lambda x: x[0], reverse=True)

            patches = sorted([y[6] for y in events_matrix1])
            patches_counts = [[y[0], y[1]] for y in Counter(patches).most_common()]
            patches_counts.sort(key = lambda x: x[0])

            midi_patches = sorted(list(set([y[3] for y in events_matrix if y[0] == 'patch_change'])))
            if len(midi_patches) == 0:
                midi_patches = [0]

            times = []
            pt = ms_events_matrix[0][1]
            start = True
            for e in ms_events_matrix:
                if (e[1]-pt) != 0 or start == True:
                    times.append((e[1]-pt))
                    start = False
                pt = e[1]

            times_sum = min(10000000, sum(times))

            durs = [e[2] for e in ms_events_matrix]
            vels = [e[5] for e in ms_events_matrix]

            avg_time = int(sum(times) / len(times))
            avg_dur = int(sum(durs) / len(durs))
            avg_vel = int(sum(vels) / len(vels))

            mode_time = statistics.mode(times)
            mode_dur = statistics.mode(durs)
            mode_vel = statistics.mode(vels)

            median_time = int(statistics.median(times))
            median_dur = int(statistics.median(durs))
            median_vel = int(statistics.median(vels))

            text_events_list = ['text_event',
                          'text_event_08',
                          'text_event_09',
                          'text_event_0a',
                          'text_event_0b',
                          'text_event_0c',
                          'text_event_0d',
                          'text_event_0e',
                          'text_event_0f']

            text_events_count = len([e for e in full_events_matrix if e[0] in text_events_list])
            lyric_events_count = len([e for e in full_events_matrix if e[0] == 'lyric'])

            chords = []
            pe = ms_events_matrix[0]
            cho = []
            for e in ms_events_matrix:
                if (e[1] - pe[1]) == 0:
                  if e[3] != 9:
                    if (e[4] % 12) not in cho:
                      cho.append(e[4] % 12)
                else:
                  if len(cho) > 0:
                    chords.append(sorted(cho))
                  cho = []
                  if e[3] != 9:
                    if (e[4] % 12) not in cho:
                      cho.append(e[4] % 12)

                pe = e

            if len(cho) > 0:
                chords.append(sorted(cho))

            ms_chords_counts = sorted([[list(key), val] for key,val in Counter([tuple(c) for c in chords if len(c) > 1]).most_common()], reverse=True, key = lambda x: x[1])
            if len(ms_chords_counts) == 0:
                ms_chords_counts = [[[0, 0], 0]]

            total_number_of_chords = len(set([y[1] for y in events_matrix1]))

            tempo_change_count = len([f for f in full_events_matrix if f[0] == 'set_tempo'])

            thirty_second_note = [e for e in events_matrix1][32]
            thirty_second_note_idx = full_events_matrix.index(thirty_second_note)

            data = []
            data.append(['total_number_of_tracks', itrack])
            data.append(['total_number_of_opus_midi_events', len(opus_events_matrix)])
            data.append(['total_number_of_score_midi_events', len(full_events_matrix)])
            data.append(['average_median_mode_time_ms', [avg_time, median_time, mode_time]])
            data.append(['average_median_mode_dur_ms', [avg_dur, median_dur, mode_dur]])
            data.append(['average_median_mode_vel', [avg_vel, median_vel, mode_vel]])
            data.append(['total_number_of_chords', total_number_of_chords])
            data.append(['total_number_of_chords_ms', len(times)])
            data.append(['ms_chords_counts', ms_chords_counts])
            data.append(['pitches_times_sum_ms', times_sum])
            data.append(['total_pitches_counts', pitches_counts])
            data.append(['midi_patches', midi_patches])
            data.append(['total_patches_counts', patches_counts])
            data.append(['tempo_change_count', tempo_change_count])
            data.append(['text_events_count', text_events_count])
            data.append(['lyric_events_count', lyric_events_count])
            data.append(['midi_ticks', score[0]])
            data.extend(full_events_matrix[:thirty_second_note_idx])
            data.append(full_events_matrix[-1])

            melody_chords_f.append([fn1, data])

            #=======================================================

            # Processed files counter
            files_count += 1

            # Saving every 5000 processed files
            if files_count % 10000 == 0:
              print('SAVING !!!')
              print('=' * 70)
              print('Saving processed files...')
              print('=' * 70)
              print('Processed so far:', files_count, 'out of', input_files_count, '===', files_count / input_files_count, 'good files ratio')
              print('=' * 70)
              count = str(files_count)
              TMIDIX.Tegridy_Any_Pickle_File_Writer(melody_chords_f, '/content/drive/MyDrive/LAMD_META_DATA_'+count)
              melody_chords_f = []
              print('=' * 70)

    except KeyboardInterrupt:
        print('Saving current progress and quitting...')
        break

    except Exception as ex:
        print('WARNING !!!')
        print('=' * 70)
        print('Bad MIDI:', f)
        print('Error detected:', ex)
        print('=' * 70)
        continue

# Saving last processed files...
print('=' * 70)
print('Saving processed files...')
print('=' * 70)
print('Processed so far:', files_count, 'out of', input_files_count, '===', files_count / input_files_count, 'good files ratio')
print('=' * 70)
count = str(files_count)
TMIDIX.Tegridy_Any_Pickle_File_Writer(melody_chords_f, '/content/drive/MyDrive/LAMD_META_DATA_'+count)

# Displaying resulting processing stats...
print('=' * 70)
print('Done!')
print('=' * 70)

print('Resulting Stats:')
print('=' * 70)
print('Total good processed MIDI files:', files_count)
print('=' * 70)

"""# (BUILD FINAL METADATA FILE)"""

#@title Build final metadata file
full_path_to_metadata_pickle_files = "/content/drive/MyDrive" #@param {type:"string"}

print('=' * 70)
print('Los Angeles MIDI Dataset Metadata File Builder')
print('=' * 70)
print('Searching for files...')

filez = list()
for (dirpath, dirnames, filenames) in os.walk(full_path_to_metadata_pickle_files):
    filez += [os.path.join(dirpath, file) for file in filenames if file.split('.')[-1] == 'pickle']
print('=' * 70)

filez.sort()

print('Loading metadata files... Please wait...')
print('=' * 70)

metadata = []

for f in tqdm(filez):

    metadata.extend(pickle.load(open(f, 'rb')))
    print('Done!')
    print('=' * 70)
    print('Loaded file:', f)
    print('=' * 70)

print('Done!')
print('=' * 70)
print('Randomizing metadata entries order...')
random.shuffle(metadata)
print('=' * 70)
print('Writing final metadata pickle file...Please wait...')

with open('/content/LAMDa_META_DATA.pickle', 'wb') as handle:
  pickle.dump(metadata, handle, protocol=pickle.HIGHEST_PROTOCOL)

print('=' * 70)
print('Done!')
print('=' * 70)

#@title Zip final metadata file
print('=' * 70)
print('Zipping... Please wait...')
print('=' * 70)
!zip LAMDa_META_DATA.zip LAMDa_META_DATA.pickle
print('=' * 70)
print('Done!')
print('=' * 70)

"""# Congrats! You did it! :)"""