{
"cells": [
{
"cell_type": "code",
"execution_count": 166,
"id": "33a80790",
"metadata": {},
"outputs": [],
"source": [
"from datasets import load_dataset, Audio\n",
"import IPython.display as ipd\n",
"\n",
"ds = load_dataset(\"znhoughton/TeleSpoof\", streaming=True)"
]
},
{
"cell_type": "code",
"execution_count": 178,
"id": "f7ef3dab",
"metadata": {},
"outputs": [],
"source": [
"from datasets import load_dataset, Audio\n",
"import tarfile\n",
"\n",
"\n",
"# Step 2: Load the dataset with the Audio feature\n",
"dataset = load_dataset(\n",
" \"znhoughton/TeleSpoof\"\n",
")\n",
"\n",
"# Step 3: Access and play the audio\n",
"audio_sample = dataset['train'][\"hf_path\"] # Access the first audio sample\n"
]
},
{
"cell_type": "code",
"execution_count": 179,
"id": "86a9df96",
"metadata": {},
"outputs": [],
"source": [
"audio_sample = audio_sample[0]\n"
]
},
{
"cell_type": "code",
"execution_count": 180,
"id": "c4a87a76",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 180,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import tarfile\n",
"import io\n",
"import soundfile as sf # pip install soundfile\n",
"import IPython.display as ipd\n",
"\n",
"\n",
"tar_path, wav_inside_tar = audio_sample.split(\"::\")\n",
"\n",
"with tarfile.open(tar_path, \"r:gz\") as tar:\n",
" wav_fileobj = tar.extractfile(wav_inside_tar)\n",
" data, samplerate = sf.read(io.BytesIO(wav_fileobj.read()))\n",
"\n",
"# Play the audio\n",
"ipd.Audio(data, rate=samplerate)\n"
]
},
{
"cell_type": "markdown",
"id": "60513b98",
"metadata": {},
"source": [
"Some code to organize the dataset. \n",
"Will delete later."
]
},
{
"cell_type": "markdown",
"id": "65089641",
"metadata": {},
"source": [
"import pandas as pd"
]
},
{
"cell_type": "markdown",
"id": "93a78868",
"metadata": {},
"source": [
"train_manifest = pd.read_csv('manifests/train_manifest.csv')"
]
},
{
"cell_type": "markdown",
"id": "42af883e",
"metadata": {},
"source": [
"train_manifest.columns"
]
},
{
"cell_type": "markdown",
"id": "9a0be068",
"metadata": {},
"source": [
"train_manifest = train_manifest.drop(columns=['Unnamed: 0', 'new_path'])"
]
},
{
"cell_type": "markdown",
"id": "270ac655",
"metadata": {},
"source": [
"test_manifest = pd.read_csv('manifests/test_manifest.csv')\n",
"test_manifest = test_manifest.drop(columns=['Unnamed: 0', 'new_path'])"
]
},
{
"cell_type": "markdown",
"id": "c632cfb8",
"metadata": {},
"source": [
"validation_manifest = pd.read_csv('manifests/validation_manifest.csv')\n",
"validation_manifest = validation_manifest.drop(columns=['Unnamed: 0', 'new_path'])"
]
},
{
"cell_type": "markdown",
"id": "30b3951c",
"metadata": {},
"source": [
"train_manifest[\"file\"] = train_manifest[\"file\"].apply(lambda x: x if x.endswith(\".wav\") else x + \".wav\")\n",
"validation_manifest[\"file\"] = validation_manifest[\"file\"].apply(lambda x: x if x.endswith(\".wav\") else x + \".wav\")\n",
"test_manifest[\"file\"] = test_manifest[\"file\"].apply(lambda x: x if x.endswith(\".wav\") else x + \".wav\")"
]
},
{
"cell_type": "markdown",
"id": "156c5db7",
"metadata": {},
"source": [
"import tarfile\n",
"import pandas as pd\n",
"from pathlib import Path\n",
"\n",
"rows = []\n",
"\n",
"# loop over all tar.gz files in train/validation/test\n",
"for split in [\"train\", \"validation\", \"test\"]:\n",
" for tar_path in Path(f\"audio_data/{split}\").glob(\"*.tar.gz\"):\n",
" with tarfile.open(tar_path, \"r:gz\") as tar:\n",
" for member in tar.getmembers():\n",
" if member.isfile():\n",
" inner_path = member.name.lstrip(\"./\")\n",
" rows.append({\n",
" \"hf_path\": f\"audio_data/{split}/{tar_path.name}::{inner_path}\", # tar::inner path\n",
" \"file\": Path(inner_path).name, # just the filename\n",
" \"split\": split\n",
" })\n",
"\n",
"df = pd.DataFrame(rows)\n"
]
},
{
"cell_type": "markdown",
"id": "a40df0d5",
"metadata": {},
"source": [
"df.sample(10)"
]
},
{
"cell_type": "markdown",
"id": "d3d0caf6",
"metadata": {},
"source": [
"train_manifest[train_manifest['dataset']=='call home'].head(5)"
]
},
{
"cell_type": "markdown",
"id": "818da652",
"metadata": {},
"source": [
"train_manifest_merged = train_manifest.merge(df, how='left')"
]
},
{
"cell_type": "markdown",
"id": "e0dbaac5",
"metadata": {},
"source": [
"train_manifest_merged.columns"
]
},
{
"cell_type": "markdown",
"id": "ce48bacd",
"metadata": {},
"source": [
"train_manifest_merged[\"hf_path\"].isna().any()"
]
},
{
"cell_type": "markdown",
"id": "ce5389ab",
"metadata": {},
"source": [
"validation_manifest_merged = validation_manifest.merge(df, how='left')\n",
"validation_manifest_merged.columns"
]
},
{
"cell_type": "markdown",
"id": "5ef3b0ef",
"metadata": {},
"source": [
"validation_manifest_merged[\"hf_path\"].isna().any()"
]
},
{
"cell_type": "markdown",
"id": "f48557af",
"metadata": {},
"source": [
"test_manifest_merged = test_manifest.merge(df, how='left')\n",
"test_manifest_merged[\"hf_path\"].isna().any()"
]
},
{
"cell_type": "markdown",
"id": "c3b45230",
"metadata": {},
"source": [
"train_manifest_merged['dataset'].value_counts()"
]
},
{
"cell_type": "markdown",
"id": "d86188aa",
"metadata": {},
"source": [
"validation_manifest_merged['dataset'].value_counts()"
]
},
{
"cell_type": "markdown",
"id": "eb4a456e",
"metadata": {},
"source": [
"test_manifest_merged['dataset'].value_counts()"
]
},
{
"cell_type": "markdown",
"id": "4ec5a193",
"metadata": {},
"source": [
"len(test_manifest_merged)"
]
},
{
"cell_type": "markdown",
"id": "c5080606",
"metadata": {},
"source": [
"test_manifest_merged = test_manifest_merged.drop(columns = ['split'])\n",
"validation_manifest_merged = validation_manifest_merged.drop(columns = ['split'])\n",
"train_manifest_merged = train_manifest_merged.drop(columns = ['Unnamed: 0', 'new_path', 'split'])"
]
},
{
"cell_type": "markdown",
"id": "9f033083",
"metadata": {},
"source": [
"test_manifest_merged.columns"
]
},
{
"cell_type": "markdown",
"id": "7500e99a",
"metadata": {},
"source": [
"train_manifest_merged['path'] = \"audio_data/train/\" + train_manifest_merged[\"hf_path\"].str.split(\"::\").str[1]\n",
"validation_manifest_merged['path'] = \"audio_data/validation/\" + train_manifest_merged[\"hf_path\"].str.split(\"::\").str[1]\n",
"test_manifest_merged['path'] = \"audio_data/test/\" + train_manifest_merged[\"hf_path\"].str.split(\"::\").str[1]"
]
},
{
"cell_type": "markdown",
"id": "f9d08a15",
"metadata": {},
"source": [
"train_manifest_merged.head(5)"
]
},
{
"cell_type": "markdown",
"id": "9ddabb8e",
"metadata": {},
"source": [
"train_manifest_merged.to_csv('manifests/train.csv', index = False)\n",
"validation_manifest_merged.to_csv('manifests/validation.csv', index = False)\n",
"test_manifest_merged.to_csv('manifests/test.csv', index = False)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.10"
}
},
"nbformat": 4,
"nbformat_minor": 5
}