lenagibee commited on
Commit
f1d001d
·
verified ·
1 Parent(s): 4cdada8

Upload GenDocVQA2024.py with huggingface_hub

Browse files
Files changed (1) hide show
  1. GenDocVQA2024.py +140 -0
GenDocVQA2024.py ADDED
@@ -0,0 +1,140 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
2
+ #
3
+ # Licensed under the Apache License, Version 2.0 (the "License");
4
+ # you may not use this file except in compliance with the License.
5
+ # You may obtain a copy of the License at
6
+ #
7
+ # http://www.apache.org/licenses/LICENSE-2.0
8
+ #
9
+ # Unless required by applicable law or agreed to in writing, software
10
+ # distributed under the License is distributed on an "AS IS" BASIS,
11
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12
+ # See the License for the specific language governing permissions and
13
+ # limitations under the License.
14
+ """This is a data loader for the GenDocVQA-2024 Dataset."""
15
+
16
+
17
+ import csv
18
+ import json
19
+ import os
20
+ import ast
21
+ import pandas as pd
22
+
23
+ import datasets
24
+
25
+ _DESCRIPTION = """\
26
+ This dataset is dedicated to the non-extractive document visual question challenge GenDocVQA-2024.
27
+ """
28
+
29
+ _URLS = {
30
+ 'img_tar': 'https://huggingface.co/datasets/lenagibee/GenDocVQA2024/resolve/main/archives/gendocvqa2024_imgs.tar.gz?download=true',
31
+ 'ocr_tar': 'https://huggingface.co/datasets/lenagibee/GenDocVQA2024/resolve/main/archives/gendocvqa2024_ocr.tar.gz?download=true',
32
+ 'annotations_tar': 'https://huggingface.co/datasets/lenagibee/GenDocVQA2024/resolve/main/archives/gendocvqa2024_annotations.tar.gz?download=true'
33
+ }
34
+
35
+ _LICENSE = "Other"
36
+
37
+ class GenDocVQA2024Small(datasets.GeneratorBasedBuilder):
38
+
39
+ VERSION = datasets.Version("1.0.0")
40
+
41
+ BUILDER_CONFIGS = [
42
+ datasets.BuilderConfig(name="default", version=VERSION, description="Whole dataset config"),
43
+ ]
44
+
45
+ DEFAULT_CONFIG_NAME = "default"
46
+
47
+ def _info(self):
48
+ features = datasets.Features(
49
+ {
50
+ "unique_id": datasets.Value("int64"),
51
+ "image_path": datasets.Value("string"),
52
+ "ocr": datasets.Sequence(feature={"text": datasets.Value("string"), "bbox": datasets.Sequence(datasets.Value("int64")),
53
+ 'block_id': datasets.Value("int64"),
54
+ 'text_id': datasets.Value("int64"),
55
+ 'par_id': datasets.Value("int64"),
56
+ 'line_id': datasets.Value("int64"),
57
+ 'word_id': datasets.Value("int64")
58
+ }),
59
+ "question": datasets.Value("string"),
60
+ "answer": datasets.Sequence(datasets.Value("string")),
61
+
62
+ }
63
+ )
64
+
65
+ return datasets.DatasetInfo(
66
+
67
+ features=features,
68
+ description=_DESCRIPTION,
69
+ license=_LICENSE
70
+ )
71
+
72
+ def _split_generators(self, dl_manager):
73
+ imgs_dir = dl_manager.download_and_extract(_URLS["img_tar"])
74
+ ocr_dir = dl_manager.download_and_extract(_URLS["ocr_tar"])
75
+ annotations_dir = dl_manager.download_and_extract(_URLS["annotations_tar"])
76
+
77
+ return [
78
+ datasets.SplitGenerator(
79
+ name=datasets.Split.TRAIN,
80
+ gen_kwargs={
81
+ "annot_path": annotations_dir,
82
+ "imgs_dir": imgs_dir,
83
+ "ocr_dir": ocr_dir,
84
+ "split": "train",
85
+ },
86
+ ),
87
+ datasets.SplitGenerator(
88
+ name=datasets.Split.VALIDATION,
89
+ gen_kwargs={
90
+ "annot_path": annotations_dir,
91
+ "imgs_dir": imgs_dir,
92
+ "ocr_dir": ocr_dir,
93
+ "split": "dev",
94
+ },
95
+ )
96
+ ]
97
+
98
+
99
+ def _generate_examples(self, annot_path, imgs_dir, ocr_dir, split):
100
+ df = pd.read_csv(os.path.join(annot_path, 'gendocvqa2024_annotations', f'{split}_v1.csv'))
101
+ for _, row in df.iterrows():
102
+ img_path = os.path.join(imgs_dir, 'gendocvqa2024_imgs', split, row['image_filename'])
103
+ q_id = row['unique_id']
104
+ ocr_path = os.path.join(ocr_dir, 'gendocvqa2024_ocr', split, row['ocr_filename'])
105
+ question = row['question']
106
+ answer = row['answer']
107
+ with open(ocr_path, 'r') as f:
108
+ ocr = json.load(f)
109
+ ocr_list = []
110
+ for item in ocr:
111
+ ocr_dict = {
112
+ 'block_id': item[0],
113
+ 'text_id': item[1],
114
+ 'par_id': item[2],
115
+ 'line_id': item[3],
116
+ 'word_id': item[4],
117
+ 'bbox': item[5],
118
+ 'text': item[6]
119
+ }
120
+ ocr_list.append(ocr_dict)
121
+ if split != "test":
122
+ answer = ast.literal_eval(answer)
123
+ else:
124
+ answer = []
125
+
126
+ yield q_id, {
127
+ "unique_id": q_id,
128
+ "image_path": img_path,
129
+ "ocr": ocr_list,
130
+ "answer": answer,
131
+ "question": question,
132
+
133
+ }
134
+
135
+
136
+
137
+ def read_image(img_path):
138
+ with Image.open(img_path) as f:
139
+ original_image = f.convert("RGB")
140
+ return original_image