Update stats.py

stats.py

@@ -1,8 +1,9 @@
+from typing import List, Literal
+
 import numpy as np
-from requests import get
 import torchaudio
-from datasets import load_dataset
-from pandas import read_csv
+from datasets import concatenate_datasets, load_dataset
+from pandas import concat, read_csv
 from tqdm import tqdm
 
 
@@ -14,14 +15,28 @@ def get_audio_length(file_path: str) -> float:
     return metadata.num_frames / metadata.sample_rate
 
 
-def get_info(subset):
-    dataset = load_dataset('audiofolder', data_dir=f'./opus/{subset}')
+def get_info(subsets: List[str] | Literal["saamgwokjinji", "seoiwuzyun"]) -> None:
+    """
+    計音頻長度、字數、平均字數、中位數字數、覆蓋字數、平均語速
+    """
+    if not isinstance(subsets, list):
+        subsets = [subsets]
+
+    datasets = []
+    for subset in subsets:
+        dataset = load_dataset('audiofolder', data_dir=f'./opus/{subset}', split='train')
+        datasets.append(dataset)
+
+    if len(datasets) > 1:
+        dataset = concatenate_datasets(datasets)
+    else:
+        dataset = datasets[0]
 
     # List to store individual audio durations
     durations = []
     total_duration = 0
 
-    for item in tqdm(dataset['train']):
+    for item in tqdm(dataset):
         file_path = item['audio']['path']
         duration = get_audio_length(file_path)
         durations.append(duration)
@@ -34,6 +49,7 @@ def get_info(subset):
     median_duration = np.median(durations)
 
     # Print results
+    print(f"Statistics for: {' & '.join(subsets)}")
     print(f"Total audio duration: {total_duration / 3600:.2f} hours")
     print(f"Total audio duration: {total_duration / 60:.2f} minutes")
     print(f"Minimum audio duration: {min_duration:.3f} seconds")
@@ -41,8 +57,16 @@ def get_info(subset):
     print(f"Average audio duration: {avg_duration:.3f} seconds")
     print(f"Median audio duration: {median_duration:.3f} seconds")
 
-    metadata = read_csv(f'./opus/{subset}/metadata.csv')
-    # transcriptions = metadata['transcription']
+    # Concatenate metadata DataFrames
+    metadata_list = []
+    for subset in subsets:
+        metadata = read_csv(f'./opus/{subset}/metadata.csv')
+        metadata_list.append(metadata)
+
+    if len(metadata_list) > 1:
+        metadata = concat(metadata_list)
+    else:
+        metadata = metadata_list[0]
 
     # Calculate total number of characters
     total_characters = metadata['transcription'].str.len().sum()
@@ -63,8 +87,10 @@ def get_info(subset):
     print(f"Number of unique characters: {len(unique_characters)}")
     print(f"Unique characters: {''.join(sorted(unique_characters))}")
     print(f"Average speech rate: {total_characters / total_duration:.2f} characters per second")
+    print("-" * 20)
 
 
 if __name__ == '__main__':
-    # get_info('saamgwokjinji')
+    get_info('saamgwokjinji')
     get_info('seoiwuzyun')
+    get_info(['saamgwokjinji', 'seoiwuzyun'])