Vinsingh commited on
Commit
40e254b
1 Parent(s): f94adce

Upload 2 files

Browse files
digital_green_process_data.py ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import pandas as pd
3
+ from datasets import Dataset, DatasetDict, Audio
4
+ import soundfile as sf
5
+ import numpy as np
6
+ from sklearn.model_selection import train_test_split
7
+
8
+ # Paths
9
+ audio_folder = '/home/azureuser/data2/dg_16/' # Path where your audio files are stored
10
+ csv_file = 'digital_green_recordings.csv' # Path to the CSV that contains audio paths and transcripts
11
+
12
+ # Read your CSV file (assumes it has columns: 'path' and 'transcript')
13
+ df = pd.read_csv(csv_file, sep="$")
14
+
15
+ # Create a new column for client_id (random or default if you don’t have speaker info)
16
+ df['client_id'] = ['speaker_' + str(i) for i in range(len(df))]
17
+
18
+ # If your CSV has relative paths, ensure the paths are correct
19
+ df['path'] = df['path'].apply(lambda x: os.path.join(audio_folder, x))
20
+
21
+ # Add additional columns needed for the Common Voice format (can be optional)
22
+ df['up_votes'] = 0
23
+ df['down_votes'] = 0
24
+ df['age'] = None
25
+ df['gender'] = None
26
+ df['accent'] = None
27
+
28
+ # Function to load and possibly convert audio to mono
29
+ def load_audio(file_path):
30
+ # Load audio file
31
+ audio, sr = sf.read(file_path)
32
+ # Convert to mono if stereo
33
+ if len(audio.shape) > 1:
34
+ audio = np.mean(audio, axis=1)
35
+ return {'audio': {'array': audio, 'sampling_rate': sr}}
36
+
37
+ # Apply audio loading function to DataFrame
38
+ df['audio'] = df['path'].apply(lambda x: load_audio(x))
39
+
40
+ train_df, test_df = train_test_split(df, test_size=0.2, random_state=42) # Adjust test_size as needed
41
+
42
+ # Convert DataFrames to Hugging Face Datasets
43
+ train_dataset = Dataset.from_pandas(train_df)
44
+ test_dataset = Dataset.from_pandas(test_df)
45
+
46
+ # Cast the 'audio' column to the 'audio' type
47
+ train_dataset = train_dataset.cast_column('audio', Audio())
48
+ test_dataset = test_dataset.cast_column('audio', Audio())
49
+
50
+ # Create a DatasetDict to simulate train/test/validation splits if needed
51
+ dataset_dict = DatasetDict({
52
+ 'train': train_dataset,
53
+ 'test': test_dataset # If you have separate splits, add them here (e.g., 'train', 'test', 'validation')
54
+ })
55
+
56
+ # Save the dataset (optional) for future use
57
+ dataset_dict.save_to_disk('data2/digital_green_data')
58
+
59
+ # Print a sample from the dataset
60
+ print(dataset_dict['train'][0])
61
+
62
+ print(dataset_dict['test'][0])
digital_green_recordings.csv ADDED
The diff for this file is too large to render. See raw diff