resample all audios to be on the sample SR

notebooks/template-audio.ipynb

@@ -10,17 +10,20 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 31,
    "metadata": {},
    "outputs": [],
    "source": [
     "from fastapi import APIRouter\n",
     "from datetime import datetime\n",
     "from datasets import load_dataset\n",
+    "import librosa\n",
     "from sklearn.metrics import accuracy_score\n",
     "import random\n",
-    "\n",
+    "import pandas as pd\n",
+    "import numpy as np\n",
     "import sys\n",
+    "import json\n",
     "sys.path.append('../tasks')\n",
     "\n",
     "from utils.evaluation import AudioEvaluationRequest\n",
@@ -53,7 +56,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -67,13 +70,68 @@
     "test_dataset = train_test[\"test\"]"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Analysis"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train = dataset[\"train\"]\n",
+    "test = dataset['test']\n",
+    "\n",
+    "train_df = pd.DataFrame(train)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_df[\"path\"] = train_df[\"audio\"].apply(lambda x: x['path'])\n",
+    "train_df[\"array\"] = train_df[\"audio\"].apply(lambda x: x['array'])\n",
+    "train_df[\"sampling_rate\"] = train_df[\"audio\"].apply(lambda x: x['sampling_rate'])"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
-    "train_test.shape"
+    "# Target sampling rate\n",
+    "target_sr = 12000\n",
+    "\n",
+    "# Function to resample the audio array\n",
+    "def resample_audio(array, orig_sr, target_sr):\n",
+    "    array = np.array(array)  # Ensure it's a numpy array\n",
+    "    if orig_sr != target_sr:\n",
+    "        array = librosa.resample(array, orig_sr=orig_sr, target_sr=target_sr)\n",
+    "    return array\n",
+    "\n",
+    "# Apply resampling to each row\n",
+    "train_df[\"resampled_array\"] = train_df.apply(\n",
+    "    lambda row: resample_audio(row[\"array\"], row[\"sampling_rate\"], target_sr), axis=1\n",
+    ")\n",
+    "\n",
+    "# Update the sampling rate column to reflect the target rate\n",
+    "train_df[\"sampling_rate\"] = target_sr\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_df.sampling_rate.describe()"
    ]
   },
   {
@@ -108,6 +166,8 @@
     "\n",
     "# Make random predictions (placeholder for actual model inference)\n",
     "true_labels = test_dataset[\"label\"]\n",
+    "\n",
+    "\n",
     "predictions = [random.randint(0, 1) for _ in range(len(true_labels))]\n",
     "\n",
     "predictions\n",