add loading script for local files

cdip-annotations-formnet.py

@@ -0,0 +1,109 @@
+import json
+import os
+from pathlib import Path
+
+from PIL import Image
+
+import datasets
+
+logger = datasets.logging.get_logger(__name__)
+def load_json(json_path):
+    with open(json_path, "r") as f:
+        data = json.load(f)
+    return data
+
+
+
+_CITATION = """\
+@article{Jaume2019FUNSDAD,
+  title={FUNSD: A Dataset for Form Understanding in Noisy Scanned Documents},
+  author={Guillaume Jaume and H. K. Ekenel and J. Thiran},
+  journal={2019 International Conference on Document Analysis and Recognition Workshops (ICDARW)},
+  year={2019},
+  volume={2},
+  pages={1-6}
+}
+"""
+
+_DESCRIPTION = """\
+https://guillaumejaume.github.io/FUNSD/
+"""
+
+
+class CDIPOcrConfig(datasets.BuilderConfig):
+    """BuilderConfig for FUNSD"""
+
+    def __init__(self, **kwargs):
+        """BuilderConfig for FUNSD.
+        Args:
+          **kwargs: keyword arguments forwarded to super.
+        """
+        super(CDIPOcrConfig, self).__init__(**kwargs)
+
+
+class CDIPOcr(datasets.GeneratorBasedBuilder):
+    """Conll2003 dataset."""
+
+    BUILDER_CONFIGS = [
+        CDIPOcrConfig(name="CDIP_OCR", version=datasets.Version("1.0.0"), description="CDIP OCR"),
+    ]
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(
+                {
+                    "id": datasets.Value("string"),
+                    "words": datasets.Sequence(datasets.Value("string")),
+                    "bboxes": datasets.Sequence(datasets.Sequence(datasets.Value("int64"))),
+                }
+            ),
+        )
+
+    def _split_generators(self, dl_manager):
+        """Returns SplitGenerators."""
+        # downloaded_file = dl_manager.download_and_extract("https://guillaumejaume.github.io/FUNSD/dataset.zip")
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN, gen_kwargs={"annot_root": f"/home/jinho/datasets/IIT-CDIP-annotations"}
+            ),
+        ]
+
+    def _generate_examples(self, annot_root):
+        logger.info("⏳ Generating examples from = %s", annot_root)
+
+        cnt = 0
+        start_from = 'o.a.f'
+        start_flag = False
+        for subfolder in sorted(list(os.listdir(annot_root))):  # b.a, b.b
+            subfolder_path = os.path.join(annot_root, subfolder)
+            logger.info(f"{subfolder = }")
+
+            for json_folder in os.listdir(subfolder_path):  # b.a.p, b.a.c
+                json_folder_path = os.path.join(subfolder_path, json_folder)
+
+
+                if not start_flag:
+                    if start_from != json_folder:
+                        continue
+                    else:
+                        start_flag = True
+
+                json_paths = list(Path(json_folder_path).glob("**/*.json"))
+                logger.info(f"{json_folder, len(json_paths) = }")
+
+                for idx, json_path in enumerate(json_paths):
+                    words = []
+                    bboxes = []
+                    data = load_json(json_path)
+                    for block in data["blocks"]:
+                        for paragraph in block["paragraphs"]:
+                            for line in paragraph["lines"]:
+                                line_words = line["words"]
+                                line_texts = [w["text"] for w in line_words]
+                                line_bboxes = [w["box"] for w in line_words]
+                                words.extend(line_texts)
+                                bboxes.extend(line_bboxes)
+                    cnt += 1
+
+                    yield cnt, {"id": str(json_path.name), "words": words, "bboxes": bboxes }