KamisatoAyaka0928 commited on
Commit
573ba04
·
verified ·
1 Parent(s): df64bf2

Update vlm-projects-multi-lang-final.py

Browse files
Files changed (1) hide show
  1. vlm-projects-multi-lang-final.py +58 -14
vlm-projects-multi-lang-final.py CHANGED
@@ -1,14 +1,18 @@
1
-
2
- import datasets
3
  import os
4
  import pandas as pd
 
5
 
6
  _DESCRIPTION = "A multilingual medical imaging dataset with questions and answers, structured by language."
7
  _HOMEPAGE = "https://huggingface.co/datasets/tungvu3196/vlm-projects-multi-lang-final"
8
  _LICENSE = "apache-2.0"
9
- _CITATION = "" # Add your citation here
10
 
11
- LANGUAGES = ['English', 'Vietnamese', 'French', 'German', 'Spanish', 'Russian', 'Korean', 'Mandarin', 'Japanese', 'Thai', 'Indonesian', 'Malay', 'Arabic', 'Hindi', 'Turkish', 'Portuguese']
 
 
 
 
12
 
13
  class VlmProjectsMultiLangFinal(datasets.GeneratorBasedBuilder):
14
  BUILDER_CONFIGS = [
@@ -21,33 +25,73 @@ class VlmProjectsMultiLangFinal(datasets.GeneratorBasedBuilder):
21
  ]
22
 
23
  def _info(self):
24
- # Use the exact features we extracted from the Parquet file
25
  return datasets.DatasetInfo(
26
  description=_DESCRIPTION,
27
- features={'A1': Value('string'), 'A2': Value('string'), 'A3': Value('string'), 'A4': Value('string'), 'Bbox coordinates normalized (X, Y, W, H)': Value('string'), 'Column 9': Value('float64'), 'Deliverable': Value('string'), 'Doctor': Value('string'), 'Google Drive Link': Value('string'), 'No.': Value('int64'), 'Notes': Value('string'), 'Original': Value('string'), 'Patient ID': Value('string'), 'Q1': Value('string'), 'Q2': Value('string'), 'Q3': Value('string'), 'Q4': Value('string'), 'Remove Status': Value('string'), 'Slide': Value('string'), 'Start date': Value('float64'), 'Status': Value('string'), '__index_level_0__': Value('int64'), 'image': Image(mode=None, decode=True), 'image_with_bboxes': Image(mode=None, decode=True), 'rotated_link': Value('string')},
28
  homepage=_HOMEPAGE,
29
  license=_LICENSE,
30
  citation=_CITATION,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31
  )
32
 
33
  def _split_generators(self, dl_manager):
34
- # The data files are already in the repository, so we don't download.
35
- # We just point to their relative paths.
36
- lang_dir = self.config.name.lower() # 'English' -> 'english'
37
-
38
  return [
39
  datasets.SplitGenerator(
40
  name=datasets.Split.TRAIN,
41
- gen_kwargs={"filepath": os.path.join("data", lang_dir, "train.parquet")},
 
42
  ),
43
  datasets.SplitGenerator(
44
  name=datasets.Split.TEST,
45
- gen_kwargs={"filepath": os.path.join("data", lang_dir, "test.parquet")},
 
46
  ),
47
  ]
48
 
49
- def _generate_examples(self, filepath):
 
50
  df = pd.read_parquet(filepath)
 
51
  for i, row in df.iterrows():
52
- yield i, row.to_dict()
 
 
 
 
 
 
 
 
 
53
 
 
 
1
+ # dataset.py
 
2
  import os
3
  import pandas as pd
4
+ import datasets
5
 
6
  _DESCRIPTION = "A multilingual medical imaging dataset with questions and answers, structured by language."
7
  _HOMEPAGE = "https://huggingface.co/datasets/tungvu3196/vlm-projects-multi-lang-final"
8
  _LICENSE = "apache-2.0"
9
+ _CITATION = ""
10
 
11
+ LANGUAGES = [
12
+ "English","Vietnamese","French","German","Spanish","Russian","Korean",
13
+ "Mandarin","Japanese","Thai","Indonesian","Malay","Arabic","Hindi",
14
+ "Turkish","Portuguese"
15
+ ]
16
 
17
  class VlmProjectsMultiLangFinal(datasets.GeneratorBasedBuilder):
18
  BUILDER_CONFIGS = [
 
25
  ]
26
 
27
  def _info(self):
 
28
  return datasets.DatasetInfo(
29
  description=_DESCRIPTION,
 
30
  homepage=_HOMEPAGE,
31
  license=_LICENSE,
32
  citation=_CITATION,
33
+ features=datasets.Features({
34
+ "A1": datasets.Value("string"),
35
+ "A2": datasets.Value("string"),
36
+ "A3": datasets.Value("string"),
37
+ "A4": datasets.Value("string"),
38
+ "Bbox coordinates normalized (X, Y, W, H)": datasets.Value("string"),
39
+ "Column 9": datasets.Value("float64"),
40
+ "Deliverable": datasets.Value("string"),
41
+ "Doctor": datasets.Value("string"),
42
+ "Google Drive Link": datasets.Value("string"),
43
+ "No.": datasets.Value("int64"),
44
+ "Notes": datasets.Value("string"),
45
+ "Original": datasets.Value("string"),
46
+ "Patient ID": datasets.Value("string"),
47
+ "Q1": datasets.Value("string"),
48
+ "Q2": datasets.Value("string"),
49
+ "Q3": datasets.Value("string"),
50
+ "Q4": datasets.Value("string"),
51
+ "Remove Status": datasets.Value("string"),
52
+ "Slide": datasets.Value("string"),
53
+ "Start date": datasets.Value("float64"),
54
+ "Status": datasets.Value("string"),
55
+ "__index_level_0__": datasets.Value("int64"),
56
+ # These two will render in the Viewer if the underlying files exist in the repo:
57
+ "image": datasets.Image(), # path or dict -> file in repo
58
+ "image_with_bboxes": datasets.Image(),
59
+ # keep as string/URL if it's not a local file:
60
+ "rotated_link": datasets.Value("string"),
61
+ }),
62
  )
63
 
64
  def _split_generators(self, dl_manager):
65
+ # Map config name ("English") to folder ("english")
66
+ lang_dir = self.config.name.lower()
67
+ base = os.path.join(self.config.data_dir or "data", lang_dir)
 
68
  return [
69
  datasets.SplitGenerator(
70
  name=datasets.Split.TRAIN,
71
+ gen_kwargs={"filepath": os.path.join(base, "train.parquet"),
72
+ "base_dir": base},
73
  ),
74
  datasets.SplitGenerator(
75
  name=datasets.Split.TEST,
76
+ gen_kwargs={"filepath": os.path.join(base, "test.parquet"),
77
+ "base_dir": base},
78
  ),
79
  ]
80
 
81
+ def _generate_examples(self, filepath, base_dir):
82
+ # Read parquet produced by your pipeline
83
  df = pd.read_parquet(filepath)
84
+
85
  for i, row in df.iterrows():
86
+ ex = row.to_dict()
87
+
88
+ # If parquet stored relative paths like "images/xyz.png", keep them relative to repo:
89
+ for col in ("image", "image_with_bboxes"):
90
+ p = ex.get(col)
91
+ if isinstance(p, str) and len(p):
92
+ # If the path isn't an URL, make it relative to the dataset files
93
+ if not (p.startswith("http://") or p.startswith("https://")):
94
+ ex[col] = os.path.join(base_dir, p).replace("\\", "/")
95
+ # if it *is* a URL, leave as-is (Image will try to download)
96
 
97
+ yield i, ex