use dl_manager to get the ocrs files (#5)

- use dl_manager to get the ocrs files (05f5fb2c7111a7cffc2fda58781408a85640979b)


Co-authored-by: Juan Pizarro <jpizarrom@users.noreply.huggingface.co>

rvl_cdip_easyocr.py

@@ -52,6 +52,12 @@ _METADATA_URLS = {
     "val": "https://huggingface.co/datasets/rvl_cdip/resolve/main/data/val.txt",
 }
 
+_OCR_URLS = {
+    "train": "https://huggingface.co/datasets/jordyvl/rvl_cdip_easyocr/resolve/main/data/Easy_Train_Data.npy",
+    "test": "https://huggingface.co/datasets/jordyvl/rvl_cdip_easyocr/resolve/main/data/Easy_Test_Data.npy",
+    "val": "https://huggingface.co/datasets/jordyvl/rvl_cdip_easyocr/resolve/main/data/Easy_Valid_Data.npy",
+}
+
 _CLASSES = [
     "letter",
     "form",
@@ -72,10 +78,6 @@ _CLASSES = [
 ]
 
 _IMAGES_DIR = "images/"   
-# hardcoded to not get stuck in annoying IO and LFS problems in Hub
-_OCR_DIR = "/cw/liir_data/NoCsBack/jordy/BDPC"
-_OCR_DIR = _OCR_DIR if os.path.exists(_OCR_DIR) else "data/"
-
 
 
 # class OCRConfig(datasets.BuilderConfig):
@@ -125,6 +127,7 @@ class RvlCdipEasyOcr(datasets.GeneratorBasedBuilder):
                 _URLS["rvl-cdip"]
             )  # only download images if need be
         labels_path = dl_manager.download(_METADATA_URLS)
+        ocrs_filepath = dl_manager.download(_OCR_URLS)
 
         return [
             datasets.SplitGenerator(
@@ -132,6 +135,7 @@ class RvlCdipEasyOcr(datasets.GeneratorBasedBuilder):
                 gen_kwargs={
                     "archive_iterator": dl_manager.iter_archive(archive_path),
                     "labels_filepath": labels_path["train"],
+                    "ocrs_filepath": ocrs_filepath["train"],
                     "split": "train",
                 },
             ),
@@ -140,6 +144,7 @@ class RvlCdipEasyOcr(datasets.GeneratorBasedBuilder):
                 gen_kwargs={
                     "archive_iterator": dl_manager.iter_archive(archive_path),
                     "labels_filepath": labels_path["test"],
+                    "ocrs_filepath": ocrs_filepath["test"],
                     "split": "test",
                 },
             ),
@@ -148,6 +153,7 @@ class RvlCdipEasyOcr(datasets.GeneratorBasedBuilder):
                 gen_kwargs={
                     "archive_iterator": dl_manager.iter_archive(archive_path),
                     "labels_filepath": labels_path["val"],
+                    "ocrs_filepath": ocrs_filepath["val"],
                     "split": "validation",
                 },
             ),
@@ -164,16 +170,14 @@ class RvlCdipEasyOcr(datasets.GeneratorBasedBuilder):
         return image_to_class_id
 
     @staticmethod
-    def _get_image_to_OCR(OCR_dir, split):
+    def _get_image_to_OCR(ocrs_filepath, split):
         def parse_easyOCR_box(box):
             # {'x0': 39, 'y0': 39, 'x1': 498, 'y1': 82, 'width': 459, 'height': 43}
             return (box["x0"], box["y0"], box["x1"], box["y1"])
 
-        if OCR_dir is None:
-            return {}
         image_to_OCR = {}
         data = np.load(
-            os.path.join(OCR_dir, f"Easy_{split[0].upper()+split[1:]}_Data.npy"),
+            ocrs_filepath,
             allow_pickle=True,
         )
         for ex in tqdm(data, desc="Loading OCR data"):
@@ -197,11 +201,11 @@ class RvlCdipEasyOcr(datasets.GeneratorBasedBuilder):
             words, boxes = image_to_OCR[file_path]
         return words, boxes
 
-    def _generate_examples(self, archive_iterator, labels_filepath, split):
+    def _generate_examples(self, archive_iterator, labels_filepath, ocrs_filepath, split):
         with open(labels_filepath, encoding="utf-8") as f:
             data = f.read().splitlines()
 
-        image_to_OCR = self._get_image_to_OCR(_OCR_DIR, split)
+        image_to_OCR = self._get_image_to_OCR(ocrs_filepath, split)
         image_to_class_id = self._get_image_to_class_map(data)
 
         for file_path, file_obj in archive_iterator: