File size: 2,428 Bytes
7b76145
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
from datasets import Dataset
import os
import json
import uuid
from PIL import PngImagePlugin

# Define the paths
arrow_file_path = "/home/yiyangai/stephenqs/datasets/mehrankazemi___re_mi/default/0.0.0/0bfa8c66979923a58bcca667181def86fb98c22a/re_mi-test-00001-of-00002.arrow"
image_output_dir = "/home/yiyangai/stephenqs/datasets/mehrankazemi___re_mi/default/0.0.0/images"
output_json_path = "/home/yiyangai/stephenqs/datasets/mehrankazemi___re_mi/default/0.0.0/re_mi-test-02.json"

# Ensure the image output directory exists
os.makedirs(image_output_dir, exist_ok=True)

# Load the dataset from the arrow file
dataset = Dataset.from_file(arrow_file_path)

# Function to save images and update their paths
def save_image_and_update_path(image_field, image_obj, record_id):
    # Generate a unique image filename using UUID
    image_filename = f"{record_id}_{image_field}.png"
    image_path = os.path.join(image_output_dir, image_filename)
    
    # Save the image to the local directory
    image_obj.save(image_path)
    
    # Return the relative path of the saved image
    return os.path.relpath(image_path, start=image_output_dir)

# List to hold the updated dataset
updated_dataset = []

# Iterate through the dataset records
for record in dataset:
    updated_record = record.copy()  # Copy the record to modify
    
    # Generate a unique UUID for the record if 'id' is missing
    record_id = str(uuid.uuid4())
    
    # Process each image field
    for image_field in ['image_1', 'image_2', 'image_3', 'image_4', 'image_5', 'image_6']:
        if isinstance(record.get(image_field), PngImagePlugin.PngImageFile):
            # Save the image and get the relative path
            relative_image_path = save_image_and_update_path(image_field, record[image_field], record_id)
            # Update the record with the new relative path
            updated_record[image_field] = relative_image_path
        else:
            # If there's no image, set the value to None
            updated_record[image_field] = None
    
    # Add the record ID to the record
    updated_record['id'] = record_id
    
    # Append the updated record to the list
    updated_dataset.append(updated_record)

# Save the updated dataset to a JSON file
with open(output_json_path, 'w') as json_file:
    json.dump(updated_dataset, json_file, indent=4)

print(f"Images saved to {image_output_dir} and dataset saved to {output_json_path}")