Add photo subset creator

This commit is contained in:
Andras Schmelczer 2024-08-25 22:14:53 +01:00
parent 131cc6132b
commit ead6498c1c
No known key found for this signature in database
GPG key ID: FC8F2C3D3D1A718C
2 changed files with 189 additions and 1 deletions

View file

@ -1,7 +1,9 @@
import random
from pathlib import Path
DATA = sorted(Path("/mnt/wsl/PHYSICALDRIVE1/data/unsplash").glob("*.jpg"))
# DATA = sorted(Path("/mnt/wsl/PHYSICALDRIVE0p1/downloaded-unsplash").glob("*"))
DATA = sorted(Path("/mnt/wsl/PHYSICALDRIVE0p1/featured").glob("*"))
TRAIN_SIZE = 0.9
CACHE_PATH = Path("/mnt/wsl/PHYSICALDRIVE1/data/cache2")

186
src/create_photo_set.ipynb Normal file
View file

@ -0,0 +1,186 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'Selected 294403 photos as mask'"
]
},
"execution_count": 1,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from pathlib import Path\n",
"from utils import set_up_logging\n",
"import os\n",
"import polars as pl\n",
"import shutil\n",
"from tqdm import tqdm\n",
"\n",
"set_up_logging(Path(\"../logs\"))\n",
"\n",
"SOURCE_PATH = Path(\"/mnt/wsl/PHYSICALDRIVE0p1/downloaded-unsplash\")\n",
"TARGET_PATH = Path(\"/mnt/wsl/PHYSICALDRIVE0p1/featured\")\n",
"\n",
"SOURCE_PATH = SOURCE_PATH.resolve()\n",
"assert SOURCE_PATH.exists()\n",
"shutil.rmtree(TARGET_PATH, ignore_errors=True)\n",
"TARGET_PATH.mkdir(exist_ok=True)\n",
"\n",
"photos = (\n",
" pl.scan_csv(\n",
" \"../data/unsplash-full/photos.tsv000\",\n",
" separator=\"\\t\",\n",
" infer_schema_length=100000,\n",
" )\n",
" .filter(pl.col(\"photo_featured\") == \"t\")\n",
" .sort(\"photo_id\")\n",
" .select(\"photo_id\", \"photo_image_url\")\n",
" .collect()\n",
")\n",
"\n",
"f\"Selected {len(photos)} photos as mask\""
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'1927632 downloaded photos found'"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"all_paths = list(SOURCE_PATH.glob(\"*\"))\n",
"f\"{len(all_paths)} downloaded photos found\""
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div><style>\n",
".dataframe > thead > tr,\n",
".dataframe > tbody > tr {\n",
" text-align: right;\n",
" white-space: pre-wrap;\n",
"}\n",
"</style>\n",
"<small>shape: (10, 2)</small><table border=\"1\" class=\"dataframe\"><thead><tr><th>photo_id</th><th>path</th></tr><tr><td>str</td><td>str</td></tr></thead><tbody><tr><td>&quot;HjVrVy3KTPM&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;0tQX63dH_oU&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;HFOjFLIgE64&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;1TiRQeEgzlM&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;85HrJ4N00dg&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;5ouomWc_8kY&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;1Zo2W5tW-VU&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;DG7Bv6V6-2Q&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;GLmohhX8vVQ&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr><tr><td>&quot;8Un0gXhEL0c&quot;</td><td>&quot;/mnt/wsl/PHYSICALDRIVE0p1/down…</td></tr></tbody></table></div>"
],
"text/plain": [
"shape: (10, 2)\n",
"┌─────────────┬─────────────────────────────────┐\n",
"│ photo_id ┆ path │\n",
"│ --- ┆ --- │\n",
"│ str ┆ str │\n",
"╞═════════════╪═════════════════════════════════╡\n",
"│ HjVrVy3KTPM ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ 0tQX63dH_oU ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ HFOjFLIgE64 ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ 1TiRQeEgzlM ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ 85HrJ4N00dg ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ 5ouomWc_8kY ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ 1Zo2W5tW-VU ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ DG7Bv6V6-2Q ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ GLmohhX8vVQ ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"│ 8Un0gXhEL0c ┆ /mnt/wsl/PHYSICALDRIVE0p1/down… │\n",
"└─────────────┴─────────────────────────────────┘"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"files = pl.DataFrame(\n",
" {\n",
" \"photo_id\": [path.name.split(\".\")[0] for path in all_paths],\n",
" \"path\": [str(path) for path in all_paths],\n",
" }\n",
")\n",
"files.limit(10)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'294367 photos found in the mask locally downloaded'"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"selected = files.join(photos, on=\"photo_id\", how=\"inner\")\n",
"f\"{len(selected)} photos found in the mask locally downloaded\""
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|██████████| 294367/294367 [30:04<00:00, 163.09it/s] \n"
]
}
],
"source": [
"for path in tqdm(selected[\"path\"]):\n",
" os.symlink(path, TARGET_PATH / path.split(\"/\")[-1])"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "bipolaroid",
"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": "3.12.2"
}
},
"nbformat": 4,
"nbformat_minor": 2
}