Datasets:
normalize HF JSONL rows for parquet conversion
Browse files- data/masks/train.jsonl +0 -0
- scripts/prepare_dataset.py +12 -2
data/masks/train.jsonl
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
scripts/prepare_dataset.py
CHANGED
|
@@ -94,14 +94,24 @@ def strip_image_data(json_path: str) -> None:
|
|
| 94 |
|
| 95 |
def build_hf_annotation_row(annotation: dict, json_path: str) -> dict:
|
| 96 |
"""Return a JSON-serializable row for HF streaming from raw annotations."""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 97 |
return {
|
| 98 |
"source_file": os.path.basename(json_path),
|
| 99 |
"version": annotation.get("version"),
|
| 100 |
-
"
|
| 101 |
-
"shapes": annotation.get("shapes", []),
|
| 102 |
"imagePath": annotation.get("imagePath"),
|
| 103 |
"imageHeight": annotation.get("imageHeight"),
|
| 104 |
"imageWidth": annotation.get("imageWidth"),
|
|
|
|
| 105 |
}
|
| 106 |
|
| 107 |
|
|
|
|
| 94 |
|
| 95 |
def build_hf_annotation_row(annotation: dict, json_path: str) -> dict:
|
| 96 |
"""Return a JSON-serializable row for HF streaming from raw annotations."""
|
| 97 |
+
shapes = []
|
| 98 |
+
for shape in annotation.get("shapes", []):
|
| 99 |
+
shapes.append({
|
| 100 |
+
"label": shape.get("label"),
|
| 101 |
+
"shape_type": shape.get("shape_type"),
|
| 102 |
+
"points": shape.get("points", []),
|
| 103 |
+
"description": shape.get("description"),
|
| 104 |
+
"group_id": shape.get("group_id"),
|
| 105 |
+
})
|
| 106 |
+
|
| 107 |
return {
|
| 108 |
"source_file": os.path.basename(json_path),
|
| 109 |
"version": annotation.get("version"),
|
| 110 |
+
"shapes": shapes,
|
|
|
|
| 111 |
"imagePath": annotation.get("imagePath"),
|
| 112 |
"imageHeight": annotation.get("imageHeight"),
|
| 113 |
"imageWidth": annotation.get("imageWidth"),
|
| 114 |
+
"num_shapes": len(shapes),
|
| 115 |
}
|
| 116 |
|
| 117 |
|