filter_images.ipynb -> manually filter videos by id

_dev/filter_images.ipynb

@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -21,17 +21,17 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [],
    "source": [
-    "df = pd.read_parquet(DATAFRAME_PATH)\n",
+    "df = pd.read_parquet(\"../data/dataset-unfiltered.parquet\")\n",
     "dim_columns = df.filter(regex=\"^dim_\").columns"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
+   "execution_count": 4,
    "metadata": {},
    "outputs": [
     {
@@ -108,9 +108,39 @@
     "print(len(df_))"
    ]
   },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Manually filter videos"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "69692\n"
+     ]
+    }
+   ],
+   "source": [
+    "IDS_TO_FILTER = [\n",
+    "    \"p7FCgw_GlWc\",\n",
+    "]\n",
+    "# remove videos with ids in IDS_TO_FILTER\n",
+    "df_ = df_[~df_[\"video_id\"].isin(IDS_TO_FILTER)]\n",
+    "print(len(df_))"
+   ]
+  },
   {
    "cell_type": "code",
-   "execution_count": 28,
+   "execution_count": 11,
    "metadata": {},
    "outputs": [],
    "source": [