Update code_to_download_datasets_in_wav_or_mp3.ipynb

code_to_download_datasets_in_wav_or_mp3.ipynb

@@ -1,47 +1,34 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from datasets import load_dataset\n",
-    "import soundfile as sf, os, pandas as pd, re\n",
-    "from tqdm import tqdm\n",
-    "\n",
-    "dataset = load_dataset(\"Sin2pi/JA_audio_JA_text_180k_samples\", trust_remote_code=True, streaming=True, token=\"your_token\")\n",
-    "\n",
-    "name = \"JA_audio_JA_text_180k_samples\"\n",
-    "output_file = 'metadata.csv'\n",
-    "os.makedirs(\"./datasets/\" + name, exist_ok=True)\n",
-    "folder_path = \"./datasets/\" + name # Create a folder to store the audio and transcription files\n",
-    "\n",
-    "char = '[　０１２３４５６７８９ａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺabcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890]'\n",
-    "special_characters = '[♬「」 ？！;:“%‘” ~♪… ?!゛＃＄％＆（）＊＋：；〈＝〉？＠＾＿‘｛｜｝～\"█♩♫♩ ♩』『.;:<>_()*&^%$#@?!`~, \"]'\n",
-    "dsa = dataset.filter(lambda batch: bool(batch[\"sentence\"])).filter(lambda sample: not re.search(char, sample[\"sentence\"]))\n",
-    "\n",
-    "for i, sample in tqdm(enumerate(dsa)): # Process each sample in the filtered dataset\n",
-    "    audio_sample = name + f'_{i}.mp3' # or wav\n",
-    "    audio_path = os.path.join(folder_path, audio_sample)\n",
-    "    transcription_path = os.path.join(folder_path, output_file)  # Path to save transcription file \n",
-    "    if not os.path.exists(audio_path):\n",
-    "        sf.write(audio_path, sample['audio']['array'], sample['audio']['sampling_rate'])\n",
-    "        sample[\"audio_length\"] = len(sample[\"audio\"][\"array\"]) / sample[\"audio\"][\"sampling_rate\"]  # Get audio length, remove if not needed\n",
-    "        with open(transcription_path, 'a', encoding='utf-8') as transcription_file: # Save transcription file    \n",
-    "            transcription_file.write(audio_sample+\",\") # Save transcription file name  \n",
-    "            sample[\"sentence\"] = re.sub(special_characters,'', sample[\"sentence\"])\n",
-    "            transcription_file.write(sample['sentence']) # Save transcription \n",
-    "            transcription_file.write(str(\",\"+str(sample['audio_length']))) # Save audio length, remove if not needed\n",
-    "            transcription_file.write('\\n') "
-   ]
-  }
- ],
- "metadata": {
-  "language_info": {
-   "name": "python"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}
+from datasets import load_dataset
+import soundfile as sf, os, pandas as pd, re
+from tqdm import tqdm
+
+dataset = load_dataset("mozilla-foundation/common_voice_17_0", "ja", split="other", trust_remote_code=True, streaming=True, token="your_token")
+
+name = "other"
+ouput_dir = "./datasets/CV17/"
+output_file = 'metadata.csv'
+os.makedirs(ouput_dir + name, exist_ok=True)
+folder_path = ouput_dir + name # Create a folder to store the audio and transcription files
+
+char = '[　０１２３４５６７８９ａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺabcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890]'
+special_characters = '[♬「」？！“%‘”~♪…?!゛＃＄％＆（）＊＋：；〈＝〉＠＾＿｛｜｝～"█♩♫』『.;:<>_()*&^$#@`, ]'
+
+dsa = (dataset
+       .filter(lambda sample: bool(sample["sentence"]))
+       .filter(lambda sample: not re.search(char, sample["sentence"]))
+       .filter(lambda sample: sample["down_votes"] == 0)
+       )
+
+for i, sample in tqdm(enumerate(dsa)): # Process each sample in the filtered dataset
+    audio_sample = name + f'_{i}.wav' # or wav
+    audio_path = os.path.join(folder_path, audio_sample)
+    transcription_path = os.path.join(folder_path, output_file)  # Path to save transcription file 
+    if not os.path.exists(audio_path):
+        sf.write(audio_path, sample['audio']['array'], sample['audio']['sampling_rate'])
+        sample["audio_length"] = len(sample["audio"]["array"]) / sample["audio"]["sampling_rate"]  # Get audio length, remove if not needed
+        with open(transcription_path, 'a', encoding='utf-8') as transcription_file: # Save transcription file    
+            transcription_file.write(audio_sample+",") # Save transcription file name  
+            sample["sentence"] = re.sub(special_characters,'', sample["sentence"])
+            transcription_file.write(sample['sentence']) # Save transcription 
+            transcription_file.write(str(","+str(sample['audio_length']))) # Save audio length, remove if not needed
+            transcription_file.write('\n')