app.py

# https://huggingface.co/spaces/asigalov61/LAKH-MIDI-Dataset-Search

import os

import time as reqtime
import datetime
from pytz import timezone

from sentence_transformers import SentenceTransformer
from sentence_transformers import util

import numpy as np

from datasets import load_dataset

import gradio as gr

import copy
import random
import pickle

import zlib

from midi_to_colab_audio import midi_to_colab_audio

import TMIDIX

import matplotlib.pyplot as plt

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

def find_midi(input_search_string):

    print('=' * 70)
    print('Req start time: {:%Y-%m-%d %H:%M:%S}'.format(datetime.datetime.now(PDT)))
    start_time = reqtime.time()

    print('-' * 70)
    print('Req search str:', input_search_string)
    print('-' * 70)
    
    print('Searching...')

    query_embedding = model.encode([input_search_string])
    
    # Compute cosine similarity between query and each sentence in the corpus
    similarities = util.cos_sim(query_embedding, corpus_embeddings)

    top_ten_matches_idxs = np.argsort(-similarities)[0][:10].tolist()
   
    # Find the index of the most similar sentence
    closest_index = np.argmax(similarities)
    closest_index_match_ratio = max(similarities[0].tolist())

    best_corpus_match = mc_dataset['train'][closest_index.tolist()]

    print('Done!')
    print('=' * 70)
    
    print('Match corpus index', closest_index)
    print('Match corpus ratio', closest_index_match_ratio)
    
    print('=' * 70)
    print('Done!')
    print('=' * 70)

    LAKH_hash = best_corpus_match['location'].split('/')[-1].split('.mid')[0]
    LAKH_caption = str(best_corpus_match['caption'])
    zlib_file_name = all_MIDI_files_names[MIDI_files_names.index(LAKH_hash)][1]

    print('Fetching MIDI score...')

    with open(zlib_file_name, 'rb') as f:
        compressed_data = f.read()
    
    # Decompress the data
    decompressed_data = zlib.decompress(compressed_data)
    
    # Convert the bytes back to a list using pickle
    scores_data = pickle.loads(decompressed_data)
    
    fnames = [f[0] for f in scores_data]
    
    fnameidx = fnames.index(LAKH_hash)

    MIDI_score_metadata = scores_data[fnameidx][1]
    MIDI_score_data = scores_data[fnameidx][2]
    
    print('Rendering results...')
    print('=' * 70)
    print('MIDi Title:', LAKH_hash)
    print('Sample INTs', MIDI_score_data[:12])
    print('=' * 70)
    
    if len(MIDI_score_data) != 0:
    
        song = MIDI_score_data
        song_f = []
        
        time = 0
        dur = 0
        vel = 90
        pitch = 0
        channel = 0
        
        patches = [-1] * 16
        
        channels = [0] * 16
        channels[9] = 1
        
        for ss in song:
    
            if 0 <= ss < 256:
            
              time += ss * 16
            
            if 256 <= ss < 512:
            
              dur = (ss-256) * 16
            
            if 512 <= ss <= 640:

                patch = (ss-512)
                
                if patch < 128:
            
                  if patch not in patches:
                    if 0 in channels:
                        cha = channels.index(0)
                        channels[cha] = 1
                    else:
                        cha = 15
            
                    patches[cha] = patch
                    channel = patches.index(patch)
                  else:
                    channel = patches.index(patch)
            
                if patch == 128:
                  channel = 9
                
            if 640 < ss < 768:

                ptc = (ss-640)

            if 768 < ss < 896:

                vel = (ss - 768)
            
                song_f.append(['note', time, dur, channel, ptc, vel, patch ])
    
    patches = [0 if x==-1 else x for x in patches]

    song_f = song_f[:3000]

    print('=' * 70)
    
    #===============================================================================
    
    output_score, patches, overflow_patches = TMIDIX.patch_enhanced_score_notes(song_f)
    
    detailed_stats = TMIDIX.Tegridy_ms_SONG_to_MIDI_Converter(output_score,
                                                          output_signature = 'LAKH MIDI Dataset Search',
                                                          output_file_name = LAKH_hash,
                                                          track_name='Project Los Angeles',
                                                          list_of_MIDI_patches=patches
                                                          )
    
    new_fn = LAKH_hash + '.mid'
        
    audio = midi_to_colab_audio(new_fn, 
                    soundfont_path=soundfont,
                    sample_rate=16000,
                    volume_scale=10,
                    output_for_gradio=True
                    )
    
    print('Done!')
    print('=' * 70)
    
    #========================================================
    
    output_midi_title = str(LAKH_hash)
    output_midi_caption = str(best_corpus_match['caption'])
    output_midi_summary = str(MIDI_score_metadata)
    output_midi_caps = str(best_corpus_match)
    output_midi = str(new_fn)
    output_audio = (16000, audio)
    
    output_plot = TMIDIX.plot_ms_SONG(output_score, plot_title=output_midi_title, return_plt=True)
    
    print('Output MIDI file name:', output_midi)
    print('Output MIDI caption string:', output_midi_caption)
    print('Output MIDI title:', output_midi_title)
    print('Output MIDI summary:', output_midi_summary)
    print('Output MidiCaps information:', output_midi_caps)
    print('=' * 70) 
    
    #========================================================
    
    print('-' * 70)
    print('Req end time: {:%Y-%m-%d %H:%M:%S}'.format(datetime.datetime.now(PDT)))
    print('-' * 70)
    print('Req execution time:', (reqtime.time() - start_time), 'sec')
    
    return output_midi_title, output_midi_caption, output_midi_summary, output_midi_caps, output_midi, output_audio, output_plot        
    
#==========================================================================================================

if __name__ == "__main__":

    PDT = timezone('US/Pacific')
    
    print('=' * 70)
    print('App start time: {:%Y-%m-%d %H:%M:%S}'.format(datetime.datetime.now(PDT)))
    print('=' * 70)

    soundfont = "SGM-v2.01-YamahaGrand-Guit-Bass-v2.7.sf2"

    print('Loading MidiCaps dataset...')
    
    mc_dataset = load_dataset("amaai-lab/MidiCaps")
    print('=' * 70)
    
    print('Loading files list...')
    
    all_MIDI_files_names = TMIDIX.Tegridy_Any_Pickle_File_Reader('LAKH_all_files_names')
    MIDI_files_names = [f[0] for f in all_MIDI_files_names]
    print('=' * 70)

    print('Loading MIDI corpus embeddings...')
    
    corpus_embeddings = np.load('MIDI_corpus_embeddings_all-mpnet-base-v2.npz')['data']
    print('Done!')
    print('=' * 70)

    print('Loading Sentence Transformer model...')
    model = SentenceTransformer('all-mpnet-base-v2')
    print('Done!')
    print('=' * 70)
    
    app = gr.Blocks()
    
    with app:
        gr.Markdown("<h1 style='text-align: center; margin-bottom: 1rem'>LAKH MIDI Dataset Search</h1>")
        gr.Markdown("<h1 style='text-align: center; margin-bottom: 1rem'>Search and explore LAKH MIDI dataset with MidiCaps dataset and sentence transformer</h1>")
        
        gr.Markdown("![Visitors](https://api.visitorbadge.io/api/visitors?path=asigalov61.LAKH-MIDI-Dataset-Search&style=flat)\n\n"
                    "This is a demo for MidiCaps dataset\n\n"
                    "Check out [MidiCaps Dataset](https://huggingface.co/datasets/amaai-lab/MidiCaps) on Hugging Face!\n\n"
                   )

        gr.Markdown("# Enter any desired song description\n\n")
        
        input_search_string = gr.Textbox(label="Search string", value="Cheery pop song about love and happiness")
        submit = gr.Button(value='Search')
        gr.ClearButton(components=[input_search_string])

        gr.Markdown("# Search results")

        output_midi_title = gr.Textbox(label="Output MIDI title")
        output_midi_caption = gr.Textbox(label="MIDI caption string")
        output_midi_summary = gr.Textbox(label="Aggregated MIDI file text metadata")
        output_midi_caps = gr.Textbox(label="MidiCaps dataset information")
        output_audio = gr.Audio(label="Output MIDI audio", format="mp3", elem_id="midi_audio")
        output_plot = gr.Plot(label="Output MIDI score plot")
        output_midi = gr.File(label="Output MIDI file", file_types=[".mid"])
        
        run_event = submit.click(find_midi, [input_search_string],
                                                  [output_midi_title, output_midi_caption, output_midi_summary, output_midi_caps, output_midi, output_audio, output_plot ])
        
    app.launch()