tokenizer_construct.py

import os
import json
import librosa
from tokenizers import Tokenizer
from tokenizers.models import WordPiece
from tokenizers.pre_tokenizers import Whitespace
from tokenizers.trainers import WordPieceTrainer
import numpy as np

class MalayalamDatasetTokenizer:
    def __init__(self, transcription_dir, wav_dir, vocab_size=16000):
        """
        Initialize tokenizer with directories for transcriptions and audio files
        
        :param transcription_dir: Path to folder containing text transcriptions
        :param wav_dir: Path to folder containing WAV audio files
        :param vocab_size: Size of the vocabulary for text tokenization
        """
        self.transcription_dir = transcription_dir
        self.wav_dir = wav_dir
        
        # Initialize text tokenizer
        self.text_tokenizer = self._create_text_tokenizer(vocab_size)
        
        # Audio tokenization parameters
        self.audio_tokenizer = {
            "sample_rate": 16000,  # Standard for speech models
            "n_mfcc": 13,  # Number of MFCCs to extract
            "n_fft": 2048,  # FFT window size
            "hop_length": 512  # Hop length between frames
        }
    
    def _create_text_tokenizer(self, vocab_size):
        """
        Create a WordPiece tokenizer for Malayalam text
        """
        tokenizer = Tokenizer(WordPiece(unk_token="[UNK]"))
        tokenizer.pre_tokenizer = Whitespace()
        
        special_tokens = ["[PAD]", "[UNK]", "[CLS]", "[SEP]"]
        
        trainer = WordPieceTrainer(
            vocab_size=vocab_size,
            special_tokens=special_tokens
        )
        
        return tokenizer
    
    def _get_matched_files(self):
        """
        Find matching transcription and audio files
        
        :return: List of tuples (transcription_path, audio_path)
        """
        matched_files = []
        
        # Get all transcription files
        for trans_file in os.listdir(self.transcription_dir):
            # Remove extension to match with audio file
            base_name = os.path.splitext(trans_file)[0]
            
            # Check for corresponding WAV file
            wav_path = os.path.join(self.wav_dir, base_name + '.wav')
            trans_path = os.path.join(self.transcription_dir, trans_file)
            
            if os.path.exists(wav_path):
                matched_files.append((trans_path, wav_path))
        
        return matched_files
    
    def process_dataset(self):
        """
        Process entire dataset, tokenizing text and extracting audio features
        
        :return: Processed dataset with tokenized text and audio features
        """
        dataset = []
        matched_files = self._get_matched_files()
        
        for trans_path, wav_path in matched_files:
            # Read transcription
            with open(trans_path, 'r', encoding='utf-8') as f:
                transcription = f.read().strip()
            
            # Tokenize text
            text_tokens = self.text_tokenizer.encode(transcription).ids
            
            # Extract audio features
            audio_features = self._extract_audio_features(wav_path)
            
            dataset.append({
                'transcription': transcription,
                'text_tokens': text_tokens,
                'audio_features': audio_features,
                'audio_path': wav_path,
                'transcription_path': trans_path
            })
        
        return dataset
    
    def _extract_audio_features(self, audio_path):
        """
        Extract MFCC features from audio file
        
        :param audio_path: Path to WAV file
        :return: Extracted audio features
        """
        # Load audio file
        audio, sr = librosa.load(
            audio_path, 
            sr=self.audio_tokenizer['sample_rate']
        )
        
        # Extract MFCCs
        mfccs = librosa.feature.mfcc(
            y=audio, 
            sr=sr,
            n_mfcc=self.audio_tokenizer['n_mfcc'],
            n_fft=self.audio_tokenizer['n_fft'],
            hop_length=self.audio_tokenizer['hop_length']
        )
        
        return mfccs.T.tolist()
    
    def train_text_tokenizer(self):
        """
        Train text tokenizer on all transcription files
        """
        # Collect all transcriptions
        transcriptions = []
        for trans_path, _ in self._get_matched_files():
            with open(trans_path, 'r', encoding='utf-8') as f:
                transcriptions.append(f.read().strip())
        
        # Train tokenizer
        self.text_tokenizer.train_from_iterator(transcriptions)
    
    def save_dataset(self, output_path):
        """
        Save processed dataset to JSON
        
        :param output_path: Path to save processed dataset
        """
        dataset = self.process_dataset()
        
        with open(output_path, 'w', encoding='utf-8') as f:
            json.dump(dataset, f, ensure_ascii=False, indent=2)
        
        print(f"Saved dataset to {output_path}")
    
    def save_tokenizer(self, output_dir):
        """
        Save tokenizer configurations
        
        :param output_dir: Directory to save tokenizer files
        """
        os.makedirs(output_dir, exist_ok=True)
        
        # Save text tokenizer vocabulary
        with open(os.path.join(output_dir, 'text_tokenizer.json'), 'w', encoding='utf-8') as f:
            json.dump({
                'vocab': self.text_tokenizer.get_vocab(),
                'model_type': 'WordPiece'
            }, f, ensure_ascii=False, indent=2)
        
        # Save audio tokenizer configuration
        with open(os.path.join(output_dir, 'audio_tokenizer.json'), 'w') as f:
            json.dump(self.audio_tokenizer, f, indent=2)

# Example usage
if __name__ == "__main__":
    # Initialize tokenizer
    tokenizer = MalayalamDatasetTokenizer(
        transcription_dir='transcription', 
        wav_dir='wav'
    )
    
    # Train text tokenizer
    tokenizer.train_text_tokenizer()
    
    # Process and save dataset
#    tokenizer.save_dataset('malayalam_dataset.json')
    
    # Save tokenizer configurations
    tokenizer.save_tokenizer('malayalam_tokenizer')