tsbpp commited on
Commit
8377992
·
verified ·
1 Parent(s): 038c5ad

Upload load.py

Browse files
Files changed (1) hide show
  1. load.py +48 -94
load.py CHANGED
@@ -1,100 +1,54 @@
1
  import json
2
- import sys
3
  import os
 
4
  import urllib.request as ureq
5
- import pdb
6
 
7
- download=1 # 0 if images are already downloaded
8
-
9
- ###############################################################
10
- ######################### load dataset json file ###############
11
- ################################################################
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
12
  with open('dataset.json', 'r') as fp:
13
- data = json.load(fp)
14
-
15
- ## dictionary data contains image URL, questions and answers ##
16
-
17
-
18
-
19
-
20
- ################################################################
21
- ############### Script for downloading images ##################
22
- ################################################################
23
- ## Make a directory images to store all images there ##########
24
- if download == 1:
25
- os.makedirs('./images', exist_ok=True)
26
- for k in data.keys():
27
- ext=os.path.splitext(data[k]['imageURL'])[1]
28
- outputFile='images/%s%s'%(k,ext)
29
- #pdb.set_trace()
30
- ureq.urlretrieve(data[k]['imageURL'],outputFile)
31
-
32
-
33
-
34
- #################################################################
35
- ################### Example of data access #####################
36
- ################################################################
37
- for k in data.keys():
38
- ext=os.path.splitext(data[k]['imageURL'])[1]
39
- imageFile='images/%s%s'%(k,ext)
40
-
41
- print('************************')
42
- print('Image file: %s'%(imageFile))
43
- print('List of questions:')
44
- print(data[k]['questions'])
45
- print('List of corresponding answers:')
46
- print(data[k]['answers'])
47
- print('Use this image as training (1), validation (2) or testing (3): %s'%(data[k]['split']))
48
- print('*************************')
49
-
50
-
51
-
52
-
53
-
54
- ######################################################################
55
- ########################### Get dataset stats ########################
56
- ######################################################################
57
- genSet=set()
58
- for k in data.keys():
59
- genSet.add(data[k]['genre'])
60
-
61
-
62
-
63
- numImages=len(data.keys())
64
- numQApairs=0
65
- numWordsInQuestions=0
66
- numWordsInAnswers=0
67
- numQuestionsPerImage=0
68
- ANS=set() # Set of unique answers
69
- authorSet=set()
70
- bookSet=set()
71
-
72
-
73
- for imgId in data.keys():
74
- numQApairs = numQApairs+len(data[imgId]['questions'])
75
- numQuestionsPerImage = numQuestionsPerImage + len(data[imgId]['questions'])
76
- authorSet.add(data[imgId]['authorName'])
77
- bookSet.add(data[imgId]['title'])
78
-
79
- for qno in range(len(data[imgId]['questions'])):
80
- ques=data[imgId]['questions'][qno]
81
- numWordsInQuestions = numWordsInQuestions+len(ques.split())
82
- for ano in range(len(data[imgId]['answers'])):
83
- ans=data[imgId]['answers'][ano]
84
- ANS.add(ans)
85
- numWordsInAnswers = numWordsInAnswers+len(str(ans).split())
86
-
87
-
88
-
89
- print("--------------------------------")
90
- print("Number of Images: %d" %(numImages))
91
- print("Number of QA pairs: %d" %(numQApairs))
92
- print("Number of unique author: %d" %(len(authorSet)))
93
- print("Number of unique title: %d" %(len(bookSet)))
94
- print("Number of unique answers: %d" %(len(ANS)))
95
- print("Number of unique genre: %d" %(len(genSet)))
96
- print("Average question length (in words): %.2f" %(float(numWordsInQuestions)/float(numQApairs)))
97
- print("Average answer length (in words): %.2f" %(float(numWordsInAnswers)/float(numQApairs)))
98
- print("Average number of questions per image: %.2f" %(float(numQuestionsPerImage)/float(numImages)))
99
- print("--------------------------------")
100
 
 
 
 
1
  import json
 
2
  import os
3
+ import time
4
  import urllib.request as ureq
 
5
 
6
+ def download_image_with_retries(url, output_file, max_retries=3, backoff_factor=1):
7
+ """Attempt to download an image with retries upon failure.
8
+
9
+ Args:
10
+ url (str): The URL of the image to download.
11
+ output_file (str): The local file path to save the downloaded image.
12
+ max_retries (int): The maximum number of retry attempts.
13
+ backoff_factor (float): The factor to calculate the delay between retries.
14
+ """
15
+ attempt = 0
16
+ while attempt < max_retries:
17
+ try:
18
+ ureq.urlretrieve(url, output_file)
19
+ print(f"Successfully downloaded: {output_file}")
20
+ return True
21
+ except Exception as e:
22
+ print(f"Attempt {attempt + 1} failed for {url}: {e}")
23
+ time.sleep(backoff_factor * (2 ** attempt))
24
+ attempt += 1
25
+ print(f"Failed to download {url} after {max_retries} attempts.")
26
+ return False
27
+
28
+ def verify_and_download_images(data):
29
+ """Verify if images are downloaded; if not, download them.
30
+
31
+ Args:
32
+ data (dict): The dataset containing image URLs and additional information.
33
+ """
34
+ images_directory = './images'
35
+ os.makedirs(images_directory, exist_ok=True)
36
+
37
+ for key, value in data.items():
38
+ image_url = value['imageURL']
39
+ ext = os.path.splitext(image_url)[1]
40
+ output_file = f'{images_directory}/{key}{ext}'
41
+
42
+ if not os.path.exists(output_file):
43
+ print(f"Image {key}{ext} not found, attempting to download...")
44
+ if not download_image_with_retries(image_url, output_file):
45
+ print(f"Warning: Could not download image {image_url}")
46
+ else:
47
+ print(f"Image {key}{ext} already exists. No download needed.")
48
+
49
+ # Load the dataset JSON file
50
  with open('dataset.json', 'r') as fp:
51
+ data = json.load(fp)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
52
 
53
+ # Check and download images as necessary
54
+ verify_and_download_images(data)