agresdominik/predictify
1
0
Fork 0
mirror of https://github.com/agresdominik/predictify.git synced 2026-04-21 17:55:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

further analisys

Browse Source
This commit is contained in:
agres
2025-05-05 16:16:24 +02:00
parent acb7717d22
commit fd341c284a
10 changed files with 330550 additions and 64904 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/ai_analysis/data_combined_dominik_chris_kaggle.csv
+111743
View File
File diff suppressed because it is too large Load Diff
src/ai_analysis/downloader.ipynb
-61045
View File
File diff suppressed because it is too large Load Diff
src/ai_analysis/feature_extaction_extension.ipynb
+1006
View File
File diff suppressed because it is too large Load Diff
src/ai_analysis/feature_extraction.ipynb
+98 -6
View File
File diff suppressed because one or more lines are too long
src/ai_analysis/final_dataset.csv
+140873
View File
File diff suppressed because it is too large Load Diff
src/ai_analysis/genre_prediction.ipynb
+45151 -3849
View File
File diff suppressed because one or more lines are too long
src/ai_analysis/spotify_preview.py
-4
View File
@@ -26,7 +26,3 @@ def get_spotify_preview_url(spotify_track_id: str) -> Optional[str]:
except Exception as e: except Exception as e:
print(f"Failed to fetch Spotify preview URL: {e}") print(f"Failed to fetch Spotify preview URL: {e}")
return None return None
# example usage:
preview_url = get_spotify_preview_url('1301WleyT98MSxVHPZCA6M')
src/ai_analysis/track_downloader.ipynb
+391
View File
@@ -0,0 +1,391 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 7814.41track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 8865.11track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 8410.16track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 10286.20track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 6751.92track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 7016.85track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 9608.71track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 569.98track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 8934.23track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 3487.43track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 8381.08track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 3057.72track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 6150.47track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 6555.71track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 2342.34track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 9073.67track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 6341.27track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 4801.47track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 4224.31track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 7571.09track/s]\n",
"Downloading previews: 100%|██████████| 91/91 [00:00<00:00, 6534.41track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 7016.58track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 7011.93track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:00<00:00, 7224.25track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 5970.09track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 1830.87track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 7771.45track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 3839.22track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 8010.83track/s]\n",
"Downloading previews: 100%|██████████| 7/7 [00:00<00:00, 1725.85track/s]\n",
"Downloading previews: 100%|██████████| 80/80 [00:00<00:00, 3127.45track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 5919.12track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 2211.42track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 5711.20track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 5389.72track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 5007.79track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 5448.83track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 1677.91track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 5254.51track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 5087.50track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 6186.85track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 1513.61track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 6105.52track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 4209.85track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 1611.84track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 127.48track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 200.62track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:00<00:00, 5717.10track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 3484.29track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 177.04track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 5664.96track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 239.08track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 223.04track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 5842.92track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 7040.71track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 7355.77track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 292.89track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 8041.64track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 420.54track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 6490.87track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 5549.89track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 5031.36track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 1444.37track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 5870.31track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 4974.82track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 4823.21track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 6310.05track/s]\n",
"Downloading previews: 100%|██████████| 196/196 [00:00<00:00, 312.44track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 5850.47track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 4904.72track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 5343.90track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 4764.65track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 4891.16track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 280.38track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 4945.14track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 4609.60track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 1155.63track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:00<00:00, 3454.36track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 4191.60track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 4414.67track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 4393.90track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:00<00:00, 2788.99track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 6180.40track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 260.50track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 4974.38track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 204.43track/s]\n",
"Downloading previews: 100%|██████████| 189/189 [00:00<00:00, 433.69track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 4620.28track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 5229.06track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:00<00:00, 6571.83track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 252.47track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 7138.69track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 4936.31track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 5408.81track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 6418.59track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 6733.21track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 6277.22track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 168.85track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 5975.06track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 7002.79track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 6256.22track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 6033.96track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:00<00:00, 283.78track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 6277.83track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 5573.59track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:00<00:00, 6510.58track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 6384.23track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 6124.12track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 6541.53track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:00<00:00, 857.85track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:00<00:00, 375.59track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 10254.22track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 6399.47track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 6457.48track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:00<00:00, 237.51track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 6714.17track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 287.82track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 6351.42track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:00<00:00, 7704.99track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:00<00:00, 449.76track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 6541.76track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:00<00:00, 7323.53track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:00<00:00, 465.08track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:15<00:00, 6.16track/s] \n",
"Downloading previews: 100%|██████████| 97/97 [00:26<00:00, 3.60track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:25<00:00, 3.85track/s]\n",
"Downloading previews: 100%|██████████| 191/191 [00:57<00:00, 3.34track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:27<00:00, 3.49track/s]\n",
"Downloading previews: 100%|██████████| 194/194 [00:53<00:00, 3.63track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:27<00:00, 3.58track/s]\n",
"Downloading previews: 100%|██████████| 187/187 [00:55<00:00, 3.35track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:29<00:00, 3.19track/s]\n",
"Downloading previews: 100%|██████████| 196/196 [00:57<00:00, 3.41track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:25<00:00, 3.63track/s]\n",
"Downloading previews: 100%|██████████| 197/197 [00:52<00:00, 3.75track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:51<00:00, 3.71track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:26<00:00, 3.69track/s]\n",
"Downloading previews: 100%|██████████| 194/194 [00:55<00:00, 3.50track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:28<00:00, 3.46track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:25<00:00, 3.69track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:25<00:00, 3.65track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:55<00:00, 3.46track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:26<00:00, 3.59track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:50<00:00, 3.74track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:25<00:00, 3.86track/s]\n",
"Downloading previews: 100%|██████████| 191/191 [00:52<00:00, 3.63track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:28<00:00, 3.40track/s]\n",
"Downloading previews: 100%|██████████| 195/195 [00:55<00:00, 3.54track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:27<00:00, 3.44track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:27<00:00, 3.57track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:34<00:00, 2.81track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:27<00:00, 3.55track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:27<00:00, 3.38track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:52<00:00, 3.64track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:32<00:00, 3.01track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:28<00:00, 3.36track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:24<00:00, 3.92track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:27<00:00, 3.40track/s]\n",
"Downloading previews: 100%|██████████| 188/188 [00:49<00:00, 3.79track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:26<00:00, 3.53track/s]\n",
"Downloading previews: 100%|██████████| 191/191 [00:55<00:00, 3.45track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:27<00:00, 3.30track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:29<00:00, 3.23track/s]\n",
"Downloading previews: 100%|██████████| 90/90 [00:22<00:00, 3.93track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:25<00:00, 3.63track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:26<00:00, 3.60track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:26<00:00, 3.72track/s]\n",
"Downloading previews: 100%|██████████| 90/90 [00:24<00:00, 3.66track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:28<00:00, 3.38track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:27<00:00, 3.59track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:25<00:00, 3.74track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:25<00:00, 3.80track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:25<00:00, 3.69track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:27<00:00, 3.62track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:25<00:00, 3.71track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:26<00:00, 3.55track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:50<00:00, 3.83track/s]\n",
"Downloading previews: 100%|██████████| 197/197 [00:53<00:00, 3.67track/s]\n",
"Downloading previews: 100%|██████████| 185/185 [00:46<00:00, 4.01track/s]\n",
"Downloading previews: 100%|██████████| 195/195 [00:48<00:00, 4.03track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:51<00:00, 3.68track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:26<00:00, 3.64track/s]\n",
"Downloading previews: 100%|██████████| 197/197 [00:52<00:00, 3.72track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:24<00:00, 3.87track/s]\n",
"Downloading previews: 100%|██████████| 195/195 [01:04<00:00, 3.01track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:54<00:00, 3.57track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:28<00:00, 3.35track/s]\n",
"Downloading previews: 100%|██████████| 194/194 [00:55<00:00, 3.47track/s]\n",
"Downloading previews: 100%|██████████| 192/192 [00:59<00:00, 3.23track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:27<00:00, 3.36track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:51<00:00, 3.67track/s]\n",
"Downloading previews: 100%|██████████| 189/189 [01:02<00:00, 3.01track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:28<00:00, 3.51track/s]\n",
"Downloading previews: 100%|██████████| 188/188 [00:55<00:00, 3.40track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:29<00:00, 3.19track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:38<00:00, 2.45track/s]\n",
"Downloading previews: 100%|██████████| 194/194 [00:55<00:00, 3.50track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:30<00:00, 3.13track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:27<00:00, 3.35track/s]\n",
"Downloading previews: 100%|██████████| 186/186 [00:56<00:00, 3.31track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:54<00:00, 3.52track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:40<00:00, 2.39track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:56<00:00, 1.64track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:54<00:00, 3.57track/s]\n",
"Downloading previews: 100%|██████████| 195/195 [01:04<00:00, 3.03track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:32<00:00, 2.93track/s]\n",
"Downloading previews: 100%|██████████| 192/192 [01:05<00:00, 2.92track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:31<00:00, 3.12track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:34<00:00, 2.82track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [00:55<00:00, 3.40track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:26<00:00, 3.49track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:53<00:00, 3.58track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:31<00:00, 3.03track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:28<00:00, 3.42track/s]\n",
"Downloading previews: 100%|██████████| 184/184 [00:50<00:00, 3.61track/s]\n",
"Downloading previews: 100%|██████████| 99/99 [00:31<00:00, 3.15track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:27<00:00, 3.42track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:27<00:00, 3.33track/s]\n",
"Downloading previews: 100%|██████████| 189/189 [00:52<00:00, 3.60track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:27<00:00, 3.54track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:25<00:00, 3.72track/s]\n",
"Downloading previews: 100%|██████████| 91/91 [00:26<00:00, 3.47track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:27<00:00, 3.50track/s]\n",
"Downloading previews: 100%|██████████| 88/88 [00:23<00:00, 3.78track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:29<00:00, 3.35track/s]\n",
"Downloading previews: 100%|██████████| 186/186 [00:53<00:00, 3.46track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:29<00:00, 3.22track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:30<00:00, 3.13track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:32<00:00, 2.91track/s]\n",
"Downloading previews: 100%|██████████| 186/186 [00:56<00:00, 3.27track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:27<00:00, 3.34track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:24<00:00, 3.72track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:26<00:00, 3.56track/s]\n",
"Downloading previews: 100%|██████████| 186/186 [00:53<00:00, 3.46track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:29<00:00, 3.18track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:27<00:00, 3.43track/s]\n",
"Downloading previews: 100%|██████████| 190/190 [01:01<00:00, 3.08track/s]\n",
"Downloading previews: 100%|██████████| 93/93 [00:28<00:00, 3.29track/s]\n",
"Downloading previews: 100%|██████████| 197/197 [00:59<00:00, 3.31track/s]\n",
"Downloading previews: 100%|██████████| 192/192 [00:59<00:00, 3.22track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:37<00:00, 2.59track/s]\n",
"Downloading previews: 100%|██████████| 192/192 [00:55<00:00, 3.48track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:26<00:00, 3.62track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:27<00:00, 3.48track/s]\n",
"Downloading previews: 100%|██████████| 188/188 [00:54<00:00, 3.44track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:28<00:00, 3.39track/s]\n",
"Downloading previews: 100%|██████████| 92/92 [00:28<00:00, 3.22track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:29<00:00, 3.30track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:27<00:00, 3.48track/s]\n",
"Downloading previews: 100%|██████████| 97/97 [00:29<00:00, 3.34track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:26<00:00, 3.66track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:55<00:00, 3.49track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:53<00:00, 3.62track/s]\n",
"Downloading previews: 100%|██████████| 94/94 [00:27<00:00, 3.41track/s]\n",
"Downloading previews: 100%|██████████| 188/188 [00:51<00:00, 3.62track/s]\n",
"Downloading previews: 100%|██████████| 191/191 [00:56<00:00, 3.41track/s]\n",
"Downloading previews: 100%|██████████| 193/193 [00:54<00:00, 3.52track/s]\n",
"Downloading previews: 100%|██████████| 191/191 [00:57<00:00, 3.30track/s]\n",
"Downloading previews: 100%|██████████| 196/196 [00:57<00:00, 3.43track/s]\n",
"Downloading previews: 100%|██████████| 95/95 [00:25<00:00, 3.67track/s]\n",
"Downloading previews: 100%|██████████| 98/98 [00:34<00:00, 2.82track/s]\n",
"Downloading previews: 100%|██████████| 188/188 [00:56<00:00, 3.35track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:28<00:00, 3.34track/s]\n",
"Downloading previews: 100%|██████████| 96/96 [00:29<00:00, 3.22track/s]\n",
"Downloading previews: 100%|██████████| 191/191 [00:58<00:00, 3.29track/s]\n",
"Downloading previews: 100%|██████████| 82/82 [00:25<00:00, 3.27track/s]\n",
"Downloading previews: 100%|██████████| 5/5 [00:00<00:00, 649.53track/s]\n",
"Downloading previews: 100%|██████████| 16/16 [00:00<00:00, 2081.48track/s]\n",
"Downloading previews: 100%|██████████| 4/4 [00:00<00:00, 1143.17track/s]\n",
"Downloading previews: 100%|██████████| 16/16 [00:00<00:00, 2154.59track/s]\n",
"Downloading previews: 100%|██████████| 49/49 [00:10<00:00, 4.51track/s]\n",
"Downloading previews: 100%|██████████| 36/36 [00:10<00:00, 3.49track/s]\n",
"Downloading previews: 100%|██████████| 19/19 [00:06<00:00, 2.76track/s]\n",
"Downloading previews: 100%|██████████| 24/24 [00:07<00:00, 3.36track/s]\n",
"Downloading previews: 100%|██████████| 20/20 [00:06<00:00, 2.99track/s]\n",
"Downloading previews: 100%|██████████| 33/33 [00:09<00:00, 3.33track/s]\n",
"Downloading previews: 100%|██████████| 30/30 [00:08<00:00, 3.72track/s]\n",
"Downloading previews: 100%|██████████| 10/10 [00:02<00:00, 3.87track/s]\n",
"Downloading previews: 100%|██████████| 2/2 [00:00<00:00, 439.26track/s]\n",
"Downloading previews: 100%|██████████| 1/1 [00:00<00:00, 5.52track/s]\n"
]
},
{
"ename": "KeyboardInterrupt",
"evalue": "",
"output_type": "error",
"traceback": [
"\u001b[31m---------------------------------------------------------------------------\u001b[39m",
"\u001b[31mKeyboardInterrupt\u001b[39m Traceback (most recent call last)",
"\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[1]\u001b[39m\u001b[32m, line 29\u001b[39m\n\u001b[32m 26\u001b[39m df_new = df[~df[\u001b[33m'\u001b[39m\u001b[33mtrack_id\u001b[39m\u001b[33m'\u001b[39m].isin(processed)].copy()\n\u001b[32m 27\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m df_new.empty:\n\u001b[32m 28\u001b[39m \u001b[38;5;66;03m# nothing new → wait and retry\u001b[39;00m\n\u001b[32m---> \u001b[39m\u001b[32m29\u001b[39m time.sleep(SLEEP_INTERVAL)\n\u001b[32m 30\u001b[39m \u001b[38;5;28;01mcontinue\u001b[39;00m\n\u001b[32m 32\u001b[39m \u001b[38;5;66;03m# 3) Download each new preview with a progress bar\u001b[39;00m\n",
"\u001b[31mKeyboardInterrupt\u001b[39m: "
]
}
],
"source": [
"import os\n",
"import time\n",
"import requests\n",
"import pandas as pd\n",
"from tqdm import tqdm\n",
"\n",
"CSV_PATH = './track_genre_balanced_url.csv'\n",
"DOWNLOAD_DIR = 'audio_previews'\n",
"SLEEP_INTERVAL = 60 # seconds to wait between checks\n",
"\n",
"os.makedirs(DOWNLOAD_DIR, exist_ok=True)\n",
"\n",
"# Keep track of which track_ids we've already attempted\n",
"processed = set()\n",
"\n",
"while True:\n",
" # 1) Load current CSV\n",
" try:\n",
" df = pd.read_csv(CSV_PATH)\n",
" except FileNotFoundError:\n",
" print(f\"{CSV_PATH} not found, waiting...\")\n",
" time.sleep(SLEEP_INTERVAL)\n",
" continue\n",
"\n",
" # 2) Identify new tracks we haven't processed yet\n",
" df_new = df[~df['track_id'].isin(processed)].copy()\n",
" if df_new.empty:\n",
" # nothing new → wait and retry\n",
" time.sleep(SLEEP_INTERVAL)\n",
" continue\n",
"\n",
" # 3) Download each new preview with a progress bar\n",
" for _, row in tqdm(df_new.iterrows(),\n",
" total=len(df_new),\n",
" desc=\"Downloading previews\",\n",
" unit=\"track\"):\n",
" track_id = row['track_id']\n",
" preview_url = row['preview']\n",
" out_path = os.path.join(DOWNLOAD_DIR, f\"{track_id}.mp3\")\n",
"\n",
" # mark as processed so we don't retry on crashes\n",
" processed.add(track_id)\n",
"\n",
" # skip if file already exists\n",
" if os.path.exists(out_path):\n",
" continue\n",
"\n",
" # attempt download\n",
" try:\n",
" resp = requests.get(preview_url, timeout=30)\n",
" if resp.status_code == 200:\n",
" with open(out_path, 'wb') as f:\n",
" f.write(resp.content)\n",
" else:\n",
" print(f\"HTTP {resp.status_code} for {track_id}\")\n",
" except Exception as e:\n",
" print(f\"Error downloading {track_id}: {e}\")\n",
"\n",
" # 4) Pause before next check\n",
" time.sleep(SLEEP_INTERVAL)\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": "3.11.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
src/ai_analysis/track_genre_balanced_url.csv
+30991
View File
File diff suppressed because it is too large Load Diff
src/ai_analysis/url_extraction.ipynb
+297
View File
@@ -0,0 +1,297 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Read out Data from Kaggle Dataset, get preview URL-s and save to file\n",
"\n",
"## this should be run only once"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 6%|▋ | 76/1183 [00:35<33:39, 1.82s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/64ffsubBonytxZc5fQJhdO\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 9%|▊ | 102/1183 [00:55<34:18, 1.90s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/2Iu5wxKFiEEQDQK1Pldsis\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 9%|▉ | 111/1183 [01:03<33:10, 1.86s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/6syvS9gZzjB8b9DdKVhAJH\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 15%|█▌ | 180/1183 [01:54<53:30, 3.20s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/2qrVR11O44iJ0DVTNCExjA\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 19%|█▉ | 225/1183 [02:25<29:37, 1.86s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/3njPW0vttbjt5j1Elt6sJI\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 32%|███▏ | 381/1183 [03:26<23:39, 1.77s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/3T7zNYia3nk9d8uXhO9Xud\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 53%|█████▎ | 630/1183 [05:23<16:28, 1.79s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/41Sfs0E8hr8w2BvzUtof4O\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 54%|█████▎ | 633/1183 [05:29<20:57, 2.29s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/3H9aA6IO5gfHW72m8YU8Iv\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 57%|█████▋ | 675/1183 [05:56<15:49, 1.87s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/0lvHnw9Exl8jLV3zuRsksJ\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 67%|██████▋ | 792/1183 [07:06<12:08, 1.86s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/17sSDGIRIkB0jOKb2cBURf\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 77%|███████▋ | 911/1183 [08:03<09:15, 2.04s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/5RcZ5jbBgKDdM6BuoSeh8P\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 77%|███████▋ | 912/1183 [08:08<13:32, 3.00s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/0YQrHOpi219lZA8SDly4iG\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 90%|█████████ | 1069/1183 [09:31<03:31, 1.85s/it]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Failed to fetch Spotify preview URL: 504 Server Error: Gateway Timeout for url: https://open.spotify.com/embed/track/2iql0ydkQX1hZ375EyRFFF\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Fetching previews: 100%|██████████| 1183/1183 [10:19<00:00, 1.91it/s]\n"
]
}
],
"source": [
"import pandas as pd\n",
"import os\n",
"from spotify_preview import get_spotify_preview_url\n",
"from tqdm import tqdm\n",
"\n",
"# --- 0) Load & dedupe your balanced track/genre file ---\n",
"df = pd.read_csv('track_genres_balanced.csv')\n",
"df = df.drop_duplicates(subset=['track_id'])\n",
"df = df.dropna(subset=['genre'])\n",
"\n",
"# --- 1) Prep output CSV (header only once) ---\n",
"#output_csv = 'track_genre_balanced_url.csv'\n",
"#pd.DataFrame(columns=['track_id','genre','preview']).to_csv(output_csv, index=False)\n",
"#output_csv = pd.read_csv('track_genre_balanced_url.csv')\n",
"\n",
"output_csv = 'track_genre_balanced_url.csv'\n",
"\n",
"if os.path.exists(output_csv):\n",
" # load already-fetched track_ids and drop them from df\n",
" done = pd.read_csv(output_csv, usecols=['track_id'])\n",
" processed_ids = set(done['track_id'].astype(str))\n",
" df = df[~df['track_id'].astype(str).isin(processed_ids)]\n",
" write_header = False\n",
"else:\n",
" # new file → write header\n",
" pd.DataFrame(columns=['track_id','genre','preview']) \\\n",
" .to_csv(output_csv, index=False)\n",
" write_header = False # header is already there\n",
"\n",
"# --- 2) Parameters ---\n",
"BATCH_SIZE = 100 # how many tracks to process per “mini‐batch”\n",
"PAUSE = 0.1 # if you want a small sleep between API calls\n",
"\n",
"# --- 3) Loop with a single progress bar over all tracks ---\n",
"with tqdm(total=len(df), desc=\"Fetching previews\") as pbar:\n",
" for start in range(0, len(df), BATCH_SIZE):\n",
" chunk = df.iloc[start:start + BATCH_SIZE]\n",
" rows = []\n",
"\n",
" # 4) Per‐track lookup\n",
" for _, row in chunk.iterrows():\n",
" track_id = row['track_id']\n",
" genre = row['genre']\n",
"\n",
" preview = get_spotify_preview_url(track_id)\n",
" if preview:\n",
" rows.append({\n",
" 'track_id': track_id,\n",
" 'genre': genre,\n",
" 'preview': preview\n",
" })\n",
" # else: silently skip or print an error if you prefer\n",
"\n",
" pbar.update(1)\n",
" if PAUSE:\n",
" import time; time.sleep(PAUSE)\n",
"\n",
" # 5) Append this batch’s hits to disk\n",
" if rows:\n",
" pd.DataFrame(rows).to_csv(\n",
" output_csv,\n",
" mode='a',\n",
" header=False,\n",
" index=False\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": "3.11.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}