From 132ac5afb5ac07b7ad061ff6637c57736cf4ebda Mon Sep 17 00:00:00 2001
From: Diogenes Analytics <diogenesanalytics@gmail.com>
Date: Fri, 2 Feb 2024 18:18:39 -0500
Subject: [PATCH] Adding demo video_clip sampling nb

---
 notebooks/video_clip_sampling.ipynb | 156 ++++++++++++++++++++++++++++
 1 file changed, 156 insertions(+)
 create mode 100644 notebooks/video_clip_sampling.ipynb

diff --git a/notebooks/video_clip_sampling.ipynb b/notebooks/video_clip_sampling.ipynb
new file mode 100644
index 0000000..dd3486d
--- /dev/null
+++ b/notebooks/video_clip_sampling.ipynb
@@ -0,0 +1,156 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "4732842d-7bea-46ed-a397-6f3204d1992b",
+   "metadata": {},
+   "source": [
+    "# Video Clip Sampling\n",
+    "This notebook seeks to explore the *video clip sampling* strategy avaiable in the `frame_sampling` package."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a39e90a7-4d3f-4966-866b-82806f96742b",
+   "metadata": {},
+   "source": [
+    "## Setup\n",
+    "First we need to get everything setup ..."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9a10719a-a86e-4abd-b8ee-c44b0c2a911f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# check for colab\n",
+    "if \"google.colab\" in str(get_ipython()):\n",
+    "  # install colab dependencies\n",
+    "  !pip install tqdm git+https://github.com/DiogenesAnalytics/frame-sampling"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "489d687e-6e4e-4cae-8158-1eafc5b45f13",
+   "metadata": {},
+   "source": [
+    "## Get Data\n",
+    "Need to set up a few things in order to have data for the demo. We will download *10 videos* from the [WebVid-10M dataset](https://maxbain.com/webvid-dataset/) *validation subset*."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3e7f616c-e4c7-478c-a485-e20756f9aa00",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "## load necessary libs\n",
+    "import pathlib\n",
+    "from urllib.request import urlretrieve\n",
+    "import pandas as pd\n",
+    "from tqdm.auto import tqdm\n",
+    "\n",
+    "# get csv data from WebVid dataset: https://maxbain.com/webvid-dataset/\n",
+    "webvid_csv = pd.read_csv(\"http://www.robots.ox.ac.uk/~maxbain/webvid/results_2M_val.csv\")\n",
+    "\n",
+    "# videos for testing purposes\n",
+    "DWNLD_LNKS = webvid_csv[\"contentUrl\"].to_list()[:10]\n",
+    "\n",
+    "# set path to data\n",
+    "WEBVID_VAL_DATA = pathlib.Path(\"/usr/local/src/frame-sampling/tests/data/video/demo\")\n",
+    "\n",
+    "# create directory\n",
+    "WEBVID_VAL_DATA.mkdir(parents=True, exist_ok=True)\n",
+    "\n",
+    "# notify of downloading\n",
+    "print(\"Downloading videos ...\")\n",
+    "\n",
+    "# download test videos\n",
+    "for video_url in tqdm(DWNLD_LNKS):\n",
+    "    # get video file name\n",
+    "    vid_file_name = pathlib.Path(video_url).name\n",
+    "    \n",
+    "    # create new output path for video file\n",
+    "    video_output_path = WEBVID_VAL_DATA / vid_file_name\n",
+    "    \n",
+    "    # download to demo_data path\n",
+    "    if not video_output_path.exists():\n",
+    "        _ = urlretrieve(video_url, video_output_path)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e0648368-7091-4cc9-99cd-d33fe01398c6",
+   "metadata": {},
+   "source": [
+    "## Frame Sampling\n",
+    "Now we are finally ready to start using the `VideoClipSampler` class to *sample frames* from the WebVid-10M video dataset. Here we will implement a function to get *frames* that have an *entropy* above a certain threshold."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "52606ab8-f8e9-4c30-a870-abd2be620e67",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# get necessary libs\n",
+    "import secrets\n",
+    "from PIL.Image import Image\n",
+    "from frame_sampling.dataset import VideoDataset\n",
+    "from frame_sampling.strategy import VideoClipSampler\n",
+    "\n",
+    "# get instance object\n",
+    "data = VideoDataset(WEBVID_VAL_DATA)\n",
+    "\n",
+    "# set frame sample output directory\n",
+    "FRAME_SAMPLE_OUTPUT = pathlib.Path(f\"./frame_samples/{secrets.token_hex(8)}\")\n",
+    "\n",
+    "# set chosen frame sample interval\n",
+    "FRAME_SAMPLE_INTERVAL = 30\n",
+    "\n",
+    "# custom function\n",
+    "def calculate_entropy_threshold(image: Image) -> bool:\n",
+    "    \"\"\"Check PIL image entropy is above threshold.\"\"\"\n",
+    "    # convert the image to grayscale\n",
+    "    image_gray = image.convert('L')\n",
+    "\n",
+    "    # calculate entropy\n",
+    "    entropy_value = image_gray.entropy()\n",
+    "\n",
+    "    # check threshold\n",
+    "    return entropy_value > 7.0\n",
+    "\n",
+    "# get instance\n",
+    "frame_sampler = VideoClipSampler(FRAME_SAMPLE_INTERVAL, calculate_entropy_threshold) \n",
+    "\n",
+    "# ... now sample\n",
+    "frame_sampler.sample(data, FRAME_SAMPLE_OUTPUT)"
+   ]
+  }
+ ],
+ "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": "3.11.7"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}