data_processing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import librosa\n",
    "import librosa.display\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from tqdm import tqdm\n",
    "from PIL import Image\n",
    "import resampy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "# old method for loading audio\n",
    "def audio_to_spectrogram(audio_file, output_dir):\n",
    "    # Load the audio file\n",
    "    audio, sr = librosa.load(\n",
    "        audio_file, sr=None\n",
    "    )  # sr=None preserves the original sampling rate\n",
    "\n",
    "    # Create the spectrogram\n",
    "    S = librosa.feature.melspectrogram(y=audio, sr=sr)\n",
    "    S_dB = librosa.power_to_db(S, ref=np.max)\n",
    "\n",
    "    # Set the file name for the spectrogram\n",
    "    fname = os.path.basename(audio_file).rsplit(\".\")[0] + \".png\"\n",
    "    spectrogram_file = os.path.join(output_dir, fname)\n",
    "\n",
    "    # Plot and save the spectrogram as an image\n",
    "    plt.figure(figsize=(8, 8))  # Adjust the figure size as needed\n",
    "    librosa.display.specshow(S_dB)\n",
    "    plt.savefig(spectrogram_file, bbox_inches=\"tight\", pad_inches=0)\n",
    "    plt.close()\n",
    "\n",
    "    return spectrogram_file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import WhisperFeatureExtractor\n",
    "\n",
    "\n",
    "# https://towardsdatascience.com/spoken-language-recognition-on-mozilla-common-voice-part-i-3f5400bbbcd8\n",
    "# reasoning behind feature size of 13 (or 20)\n",
    "\n",
    "N_MELS = 20\n",
    "\n",
    "class CustomWhisperFeatureExtractor(WhisperFeatureExtractor):\n",
    "    def __init__(self, n_mels=N_MELS, **kwargs):\n",
    "        super().__init__(feature_size=n_mels, **kwargs)\n",
    "\n",
    "\n",
    "feature_extractor = CustomWhisperFeatureExtractor(\n",
    "    n_mels=N_MELS, sampling_rate=16000, padding_value=0.0, return_attention_mask=False\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "# based on https://github.com/pietz/language-recognition/blob/master/Language%20Classifier.ipynb\n",
    "# see https://towardsdatascience.com/spoken-language-recognition-on-mozilla-common-voice-part-i-3f5400bbbcd8\n",
    "# for choice of features (13 or 20)\n",
    "def mp3_to_img(path, min_length=0.0, max_length=20.0, n_mels=N_MELS):\n",
    "    signal, sr = librosa.load(path)\n",
    "    signal = resampy.resample(signal, sr_orig=sr, sr_new=16000, res_type=\"kaiser_fast\")\n",
    "    # hl = signal.shape[0] // (width * 1.1)  # Cut away 5% from start and end\n",
    "\n",
    "    duration = librosa.get_duration(y=signal, sr=16000)\n",
    "    if duration < min_length or duration > max_length:\n",
    "        # print(f\"Skipping {path} due to length {duration}\")\n",
    "        return None, None\n",
    "\n",
    "    f = feature_extractor(\n",
    "        signal,\n",
    "        sampling_rate=16000,\n",
    "        padding=\"max_length\", # pads to 30 seconds\n",
    "        do_normalize=True,\n",
    "        feature_size=n_mels,\n",
    "        return_attention_mask=False,\n",
    "        # hop_length=hl,\n",
    "    )\n",
    "    # [print(k, v[0].shape) for k, v in f.items()]\n",
    "\n",
    "    # spec = librosa.feature.melspectrogram(y=signal, n_mels=height, hop_length=int(hl))\n",
    "    input_features = np.array(f[\"input_features\"])[0]\n",
    "    # truncate to width of 1000\n",
    "    # print(input_features.shape)\n",
    "    input_features = input_features[:, :1000]\n",
    "\n",
    "    fname = os.path.basename(path).rsplit(\".\")[0]\n",
    "    return input_features, fname\n",
    "\n",
    "\n",
    "# https://stackoverflow.com/questions/56719138/how-can-i-save-a-librosa-spectrogram-plot-as-a-specific-sized-image/57204349#57204349\n",
    "def scale_minmax(X, min=0.0, max=1.0):\n",
    "    X_std = (X - X.min()) / (X.max() - X.min())\n",
    "    X_scaled = X_std * (max - min) + min\n",
    "    return X_scaled\n",
    "\n",
    "\n",
    "# https://stackoverflow.com/questions/56719138/how-can-i-save-a-librosa-spectrogram-plot-as-a-specific-sized-image/57204349#57204349\n",
    "def save_spec(spec, out):\n",
    "    # min-max scale to fit inside 8-bit range\n",
    "    spec = scale_minmax(spec, 0, 255).astype(np.uint8)\n",
    "    img = np.flip(spec, axis=0)  # put low frequencies at the bottom in image\n",
    "    # img = 255 - img  # invert. make black==more energy\n",
    "\n",
    "    # save with PIL\n",
    "    # display(Image.fromarray(img))\n",
    "    Image.fromarray(img).save(out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/m9/9vx0zm4904qdn2jzwc0shtpm0000gn/T/ipykernel_25354/12447177.py:5: UserWarning: PySoundFile failed. Trying audioread instead.\n",
      "  signal, sr = librosa.load(path)\n",
      "/Users/thabnir/maisproj/.venv/lib/python3.10/site-packages/librosa/core/audio.py:183: FutureWarning: librosa.core.audio.__audioread_load\n",
      "\tDeprecated as of librosa version 0.10.0.\n",
      "\tIt will be removed in librosa version 1.0.\n",
      "  y, sr_native = __audioread_load(path, offset, duration, dtype)\n",
      "/Users/thabnir/maisproj/audiotospect.py:48: UserWarning: PySoundFile failed. Trying audioread instead.\n",
      "  signal, sr = librosa.load(audiopath, res_type=\"kaiser_fast\")\n",
      "/Users/thabnir/maisproj/.venv/lib/python3.10/site-packages/librosa/core/audio.py:183: FutureWarning: librosa.core.audio.__audioread_load\n",
      "\tDeprecated as of librosa version 0.10.0.\n",
      "\tIt will be removed in librosa version 1.0.\n",
      "  y, sr_native = __audioread_load(path, offset, duration, dtype)\n",
      "/Users/thabnir/maisproj/audiotospect.py:48: UserWarning: PySoundFile failed. Trying audioread instead.\n",
      "  signal, sr = librosa.load(audiopath, res_type=\"kaiser_fast\")\n",
      "/Users/thabnir/maisproj/.venv/lib/python3.10/site-packages/librosa/core/audio.py:183: FutureWarning: librosa.core.audio.__audioread_load\n",
      "\tDeprecated as of librosa version 0.10.0.\n",
      "\tIt will be removed in librosa version 1.0.\n",
      "  y, sr_native = __audioread_load(path, offset, duration, dtype)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.QuadMesh at 0x2b57377c0>"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import audiotospect\n",
    "\n",
    "file = \"data/mp3/en/common_voice_en_604929.mp3\"\n",
    "file2 = \"data/mp3/fr/common_voice_fr_17301086.mp3\"\n",
    "file3 = \"data/mp3/en/common_voice_en_25282.mp3\"\n",
    "\n",
    "old_version, fname = mp3_to_img(\"./flask/temp/recording.webm\")\n",
    "new_version = audiotospect.audio_to_spect(\n",
    "    \"./flask/temp/recording.webm\",\n",
    "    # file,\n",
    "    target_size=(13, 500),\n",
    "    img_dimensions=(13, 500, 1),\n",
    "    save=False,\n",
    ")\n",
    "new_version_2 = audiotospect.audio_to_spect(\n",
    "    \"./flask/temp/recording.webm\",\n",
    "    # file,\n",
    "    target_size=(20, 500),\n",
    "    img_dimensions=(20, 500, 1),\n",
    "    save=False,\n",
    ")\n",
    "\n",
    "librosa.display.specshow(\n",
    "    new_version_2, x_axis=\"time\", y_axis=\"mel\", sr=16000, hop_length=160\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.QuadMesh at 0x2b5793cd0>"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "librosa.display.specshow(new_version, x_axis=\"time\", y_axis=\"mel\", sr=16000, hop_length=160) \n",
    "# librosa.display.specshow(\n",
    "#     old_version,\n",
    "#     x_axis=\"time\",\n",
    "#     y_axis=\"mel\",\n",
    "#     sr=feature_extractor.sampling_rate,\n",
    "#     hop_length=feature_extractor.hop_length,\n",
    "# )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "MIN_DURATION_IN_SECONDS = 4.0\n",
    "MAX_DURATION_IN_SECONDS = 9.0\n",
    "\n",
    "import random\n",
    "\n",
    "\n",
    "def get_valid_lengths(input_dir):\n",
    "    if not os.path.exists(input_dir):\n",
    "        print(\"Input directory does not exist.\")\n",
    "        return\n",
    "\n",
    "    files = os.listdir(input_dir)\n",
    "    lengths = []\n",
    "    random.shuffle(files)\n",
    "    # files = files[:1000]\n",
    "    num_removed = 0\n",
    "    num_files = len(files)\n",
    "    for file in tqdm(files):\n",
    "        if file.endswith(\".mp3\"):\n",
    "            audio_file = os.path.join(input_dir, file)\n",
    "            length = librosa.get_duration(filename=audio_file)\n",
    "            if length > MIN_DURATION_IN_SECONDS and length < MAX_DURATION_IN_SECONDS:\n",
    "                lengths.append(length)\n",
    "            else:\n",
    "                print(f\"File {file} has length {length}.\")\n",
    "                num_removed += 1\n",
    "\n",
    "    print(f\"Removed {num_removed} files out of {num_files} files.\")\n",
    "    # equitvalent in pct:\n",
    "    print(f\"Removed {num_removed/num_files*100:.2f}% of files.\")\n",
    "    return lengths"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_spects(\n",
    "    input_dir,\n",
    "    output_dir,\n",
    "    num_files=100,\n",
    "    shuffle=True,\n",
    "    offset=0,\n",
    "    min_length=MIN_DURATION_IN_SECONDS,\n",
    "    max_length=MAX_DURATION_IN_SECONDS,\n",
    "):\n",
    "    # TODO: make it auto-detect files that have already been converted and skip them\n",
    "    # so this script can be run multiple times to add more files\n",
    "    if not os.path.exists(input_dir):\n",
    "        print(\"Input directory does not exist.\")\n",
    "        return\n",
    "\n",
    "    os.makedirs(output_dir, exist_ok=True)\n",
    "\n",
    "    files = os.listdir(input_dir)\n",
    "    if shuffle:\n",
    "        np.random.shuffle(files)\n",
    "\n",
    "    already_converted = os.listdir(output_dir)\n",
    "    # take intersection of files and already_converted\n",
    "    already_converted = set([f.rsplit(\".\")[0] for f in already_converted])\n",
    "    print(len(already_converted))\n",
    "    # subtract\n",
    "    files = [f for f in files if f.rsplit(\".\")[0] not in already_converted]\n",
    "    print(f\"Converting {len(files)} files.\")\n",
    "    # note: this doesn't work for files that have been read and found to be too long or too short\n",
    "    # and thus skipped\n",
    "    # those will be re-measured, re-determined to be too long or too short, and skipped again\n",
    "    # there are ~7000 of those per language...\n",
    "    # oh well\n",
    "    # could keep track of which files have been skipped previously and skip them again but whatever\n",
    "\n",
    "\n",
    "    files = files[offset : num_files + offset]\n",
    "    # files = files[:num_files]\n",
    "    print(f\"Converting {len(files)} files.\")\n",
    "\n",
    "    lengths = []\n",
    "    for file in tqdm(files):\n",
    "        if file.endswith(\".mp3\"):\n",
    "            audio_file = os.path.join(input_dir, file)\n",
    "            img, fname = mp3_to_img(path=audio_file, min_length=min_length, max_length=max_length)\n",
    "            if img is not None:\n",
    "                lengths.append(img.shape[1])\n",
    "                save_spec(img, os.path.join(output_dir, fname) + \".png\")\n",
    "            # print(img.shape)\n",
    "            # lengths.append(img.shape[1])\n",
    "            # convert to rgb using PIL\n",
    "            # img = Image.fromarray(img[0])\n",
    "            # display(img.convert(\"RGB\"))\n",
    "            # img.save(os.path.join(output_dir, fname) + \".png\")\n",
    "    return lengths # measured in number of timesteps (width of image)\n",
    "\n",
    "# current data used: <lang>_train_0, <lang>_train_1, <lang>_train_10\n",
    "# don't know why they skip 2-9 but whatever, they're not on huggingface\n",
    "# except for italian and spanish, which have <lang>_train_0, <lang>_train_1, <lang>_train_2\n",
    "\n",
    "# langs = [\"ar\", \"de\", \"en\", \"es\", \"fr\", \"hi\", \"it\", \"ja\", \"ru\", \"zh-CN\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting mp3 files in `data/mp3/de/` to spectrograms in `data/spectrogram/de_train_20_mels/`\n",
      "0\n",
      "Converting 120000 files.\n",
      "Converting 120000 files.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 120000/120000 [3:26:18<00:00,  9.69it/s]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting mp3 files in `data/mp3/en/` to spectrograms in `data/spectrogram/en_train_20_mels/`\n",
      "0\n",
      "Converting 120000 files.\n",
      "Converting 120000 files.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 120000/120000 [2:25:17<00:00, 13.77it/s] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting mp3 files in `data/mp3/es/` to spectrograms in `data/spectrogram/es_train_20_mels/`\n",
      "0\n",
      "Converting 120000 files.\n",
      "Converting 120000 files.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 120000/120000 [1:59:01<00:00, 16.80it/s] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting mp3 files in `data/mp3/fr/` to spectrograms in `data/spectrogram/fr_train_20_mels/`\n",
      "0\n",
      "Converting 120000 files.\n",
      "Converting 120000 files.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 120000/120000 [2:09:22<00:00, 15.46it/s] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting mp3 files in `data/mp3/it/` to spectrograms in `data/spectrogram/it_train_20_mels/`\n",
      "0\n",
      "Converting 120000 files.\n",
      "Converting 120000 files.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 120000/120000 [2:11:45<00:00, 15.18it/s] \n"
     ]
    }
   ],
   "source": [
    "\n",
    "langs = [\"de\", \"en\", \"es\", \"fr\", \"it\"]\n",
    "\n",
    "for lang in langs:\n",
    "    input_dir = f\"data/mp3/{lang}/\"\n",
    "    output_dir = f\"data/spectrogram/{lang}_train_{N_MELS}_mels/\"\n",
    "    print(f\"Converting mp3 files in `{input_dir}` to spectrograms in `{output_dir}`\")\n",
    "    lens = make_spects(\n",
    "        input_dir,\n",
    "        output_dir,\n",
    "        num_files=120_000,\n",
    "        shuffle=False,\n",
    "        offset=0, # adding new files to the previously converted 20,000 ignoring previously converted files\n",
    "        min_length=MIN_DURATION_IN_SECONDS,\n",
    "        max_length=MAX_DURATION_IN_SECONDS,\n",
    "    )\n",
    "    # lens = get_valid_lengths(input_dir)\n",
    "    lens = np.array(lens)\n",
    "    # print(f\"Average length: {np.mean(lens)}\")\n",
    "    # print(f\"Max length: {np.max(lens)}\")\n",
    "    # print(f\"Min length: {np.min(lens)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # measure length of spectrograms\n",
    "# def get_spect_width(path):\n",
    "#     img = Image.open(path)\n",
    "#     return img.size[0]    \n",
    "    \n",
    "\n",
    "# # langs2 = [\"en\", \"de\"]\n",
    "# lengths = []\n",
    "# i = 0\n",
    "# for lang in langs:\n",
    "#     input_dir = f\"data/spectrogram/{lang}_train_new/\"\n",
    "#     lengths.append([])\n",
    "#     for file in tqdm(os.listdir(input_dir)):\n",
    "#         if file.endswith(\".png\"):\n",
    "#             path = os.path.join(input_dir, file)\n",
    "#             width = get_spect_width(path)\n",
    "#             lengths[i].append(width)\n",
    "    \n",
    "#     lens = np.array(lengths[i])\n",
    "#     print(f\"Average length: {np.mean(lens)}\")\n",
    "#     print(f\"Max length: {np.max(lens)}\")\n",
    "#     print(f\"Min length: {np.min(lens)}\")\n",
    "#     i += 1\n",
    "# # get average length\n",
    "# # get max length\n",
    "# # get min length\n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.10.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}