Datasets:
Delete restructure.py
Browse files- restructure.py +0 -37
restructure.py
DELETED
|
@@ -1,37 +0,0 @@
|
|
| 1 |
-
import os
|
| 2 |
-
import shutil
|
| 3 |
-
import pandas as pd
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
def restructure_images(df, image_dir, output_dir):
|
| 7 |
-
"""
|
| 8 |
-
Moves images from a flat folder to a nested structure: /01/ab/image_id.jpg
|
| 9 |
-
and updates the DataFrame with the new paths.
|
| 10 |
-
"""
|
| 11 |
-
new_paths = []
|
| 12 |
-
for img_id in df["image_path"]:
|
| 13 |
-
# Create nested folders based on first 4 chars of ID
|
| 14 |
-
sub_dir_1 = img_id[0:2]
|
| 15 |
-
sub_dir_2 = img_id[2:4]
|
| 16 |
-
|
| 17 |
-
target_folder = os.path.join(output_dir, sub_dir_1, sub_dir_2)
|
| 18 |
-
os.makedirs(target_folder, exist_ok=True)
|
| 19 |
-
|
| 20 |
-
src = os.path.join(image_dir, f"{img_id}.jpg")
|
| 21 |
-
dst = os.path.join(target_folder, f"{img_id}.jpg")
|
| 22 |
-
|
| 23 |
-
if os.path.exists(src):
|
| 24 |
-
shutil.copy(src, dst)
|
| 25 |
-
# Store relative path for the metadata
|
| 26 |
-
new_paths.append(f"{sub_dir_1}/{sub_dir_2}/{img_id}.jpg")
|
| 27 |
-
else:
|
| 28 |
-
new_paths.append(None)
|
| 29 |
-
|
| 30 |
-
df["relative_path"] = new_paths
|
| 31 |
-
return df
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
# Usage
|
| 35 |
-
df = pd.read_parquet("products.parquet")
|
| 36 |
-
df = restructure_images(df, "./images_flat", "./images_nested")
|
| 37 |
-
df.to_parquet("products.parquet", index=False)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|