Add to support visisual result under CLI

.github/workflows/gen_whl_to_pypi_rapidocr_ort.yml

@@ -86,7 +86,7 @@ jobs:
           mv rapidocr_onnxruntime rapidocr_onnxruntime_t
           mv rapidocr_onnxruntime_t rapidocr_onnxruntime
           cd rapidocr_onnxruntime
-          echo "from .rapidocr_onnxruntime.main import RapidOCR" > __init__.py
+          echo "from .rapidocr_onnxruntime.main import RapidOCR, VisRes" > __init__.py
 
           cd ..
           python -m pip install --upgrade pip

.github/workflows/gen_whl_to_pypi_rapidocr_vino.yml

@@ -86,7 +86,7 @@ jobs:
           mv rapidocr_openvino rapidocr_openvino_t
           mv rapidocr_openvino_t rapidocr_openvino
           cd rapidocr_openvino
-          echo "from .rapidocr_openvino.main import RapidOCR" > __init__.py
+          echo "from .rapidocr_openvino.main import RapidOCR, VisRes" > __init__.py
 
           cd ..
           python -m pip install --upgrade pip

python/demo.py

@@ -8,17 +8,17 @@
 # from rapidocr_openvino import RapidOCR, VisRes
 
 
-rapid_ocr = RapidOCR()
+engine = RapidOCR()
 vis = VisRes(font_path="resources/fonts/FZYTK.TTF")
 
 image_path = "tests/test_files/ch_en_num.jpg"
 with open(image_path, "rb") as f:
     img = f.read()
 
-result, elapse_list = rapid_ocr(img)
+result, elapse_list = engine(img)
 print(result)
 print(elapse_list)
 
 boxes, txts, scores = list(zip(*result))
-res = vis(img, boxes, txts, scores)
-cv2.imwrite("vis.png", res)
+vis_img = vis(img, boxes, txts, scores)
+cv2.imwrite("vis.png", vis_img)

python/rapidocr_onnxruntime/main.py

@@ -11,7 +11,14 @@
 from .ch_ppocr_v2_cls import TextClassifier
 from .ch_ppocr_v3_det import TextDetector
 from .ch_ppocr_v3_rec import TextRecognizer
-from .utils import LoadImage, UpdateParameters, concat_model_path, init_args, read_yaml
+from .utils import (
+    LoadImage,
+    UpdateParameters,
+    VisRes,
+    concat_model_path,
+    init_args,
+    read_yaml,
+)
 
 root_dir = Path(__file__).resolve().parent
 
@@ -257,9 +264,27 @@ def main():
         args.img_path, use_det=use_det, use_cls=use_cls, use_rec=use_rec
     )
     print(result)
+
     if args.print_cost:
         print(elapse_list)
 
+    if args.vis_res and args.vis_font_path:
+        vis = VisRes(font_path=args.vis_font_path)
+        Path(args.vis_save_path).mkdir(parents=True, exist_ok=True)
+
+        save_path = Path(args.vis_save_path) / f"{Path(args.img_path).stem}_vis.png"
+
+        if use_det and not use_cls and not use_rec:
+            boxes, *_ = list(zip(*result))
+            vis_img = vis(args.img_path, boxes, None, None)
+            cv2.imwrite(str(save_path), vis_img)
+            print(f"The vis result has saved in {save_path}")
+        elif use_det and use_rec:
+            boxes, txts, scores = list(zip(*result))
+            vis_img = vis(args.img_path, boxes, txts, scores)
+            cv2.imwrite(str(save_path), vis_img)
+            print(f"The vis result has saved in {save_path}")
+
 
 if __name__ == "__main__":
     main()

python/rapidocr_onnxruntime/utils.py

@@ -265,6 +265,21 @@ def init_args():
     rec_group.add_argument("--rec_img_shape", type=list, default=[3, 48, 320])
     rec_group.add_argument("--rec_batch_num", type=int, default=6)
 
+    vis_group = parser.add_argument_group(title="Visual Result")
+    vis_group.add_argument("-vis", "--vis_res", action="store_true", default=False)
+    vis_group.add_argument(
+        "--vis_font_path",
+        type=str,
+        default=None,
+        help="When -vis is True, the font_path must have value.",
+    )
+    vis_group.add_argument(
+        "--vis_save_path",
+        type=str,
+        default=".",
+        help="The directory of saving the vis image.",
+    )
+
     args = parser.parse_args()
     return args
 
@@ -352,7 +367,7 @@ class VisRes:
     def __init__(
         self, font_path: Optional[Union[str, Path]] = None, text_score: float = 0.5
     ):
-        if font_path is None:
+        if font_path is None or not Path(font_path).exists():
             raise FileNotFoundError(
                 f"The {font_path} does not exists! \n"
                 f"You could download the file in the https://drive.google.com/file/d/1evWVX38EFNwTq_n5gTFgnlv8tdaNcyIA/view?usp=sharing"

python/rapidocr_openvino/main.py

@@ -11,7 +11,14 @@
 from .ch_ppocr_v2_cls import TextClassifier
 from .ch_ppocr_v3_det import TextDetector
 from .ch_ppocr_v3_rec import TextRecognizer
-from .utils import LoadImage, UpdateParameters, concat_model_path, init_args, read_yaml
+from .utils import (
+    LoadImage,
+    UpdateParameters,
+    VisRes,
+    concat_model_path,
+    init_args,
+    read_yaml,
+)
 
 root_dir = Path(__file__).resolve().parent
 
@@ -257,9 +264,27 @@ def main():
         args.img_path, use_det=use_det, use_cls=use_cls, use_rec=use_rec
     )
     print(result)
+
     if args.print_cost:
         print(elapse_list)
 
+    if args.vis_res and args.vis_font_path:
+        vis = VisRes(font_path=args.vis_font_path)
+        Path(args.vis_save_path).mkdir(parents=True, exist_ok=True)
+
+        save_path = Path(args.vis_save_path) / f"{Path(args.img_path).stem}_vis.png"
+
+        if use_det and not use_cls and not use_rec:
+            boxes, *_ = list(zip(*result))
+            vis_img = vis(args.img_path, boxes, None, None)
+            cv2.imwrite(str(save_path), vis_img)
+            print(f"The vis result has saved in {save_path}")
+        elif use_det and use_rec:
+            boxes, txts, scores = list(zip(*result))
+            vis_img = vis(args.img_path, boxes, txts, scores)
+            cv2.imwrite(str(save_path), vis_img)
+            print(f"The vis result has saved in {save_path}")
+
 
 if __name__ == "__main__":
     main()

python/rapidocr_openvino/utils.py

@@ -194,6 +194,21 @@ def init_args():
     rec_group.add_argument("--rec_img_shape", type=list, default=[3, 48, 320])
     rec_group.add_argument("--rec_batch_num", type=int, default=6)
 
+    vis_group = parser.add_argument_group(title="Visual Result")
+    vis_group.add_argument("-vis", "--vis_res", action="store_true", default=False)
+    vis_group.add_argument(
+        "--vis_font_path",
+        type=str,
+        default=None,
+        help="When -vis is True, the font_path must have value.",
+    )
+    vis_group.add_argument(
+        "--vis_save_path",
+        type=str,
+        default=".",
+        help="The directory of saving the vis image.",
+    )
+
     args = parser.parse_args()
     return args
 
@@ -281,7 +296,7 @@ class VisRes:
     def __init__(
         self, font_path: Optional[Union[str, Path]] = None, text_score: float = 0.5
     ):
-        if font_path is None:
+        if font_path is None or not Path(font_path).exists():
             raise FileNotFoundError(
                 f"The {font_path} does not exists! \n"
                 f"You could download the file in the https://drive.google.com/file/d/1evWVX38EFNwTq_n5gTFgnlv8tdaNcyIA/view?usp=sharing"

python/setup_onnxruntime.py

@@ -38,7 +38,7 @@ def get_readme():
 sys.argv = sys.argv[:2]
 
 project_urls = {
-    "Documentation": "https://rapidai.github.io/RapidOCRDocs/docs/install_usage/rapidocr/rapidocr_onnxruntime/",
+    "Documentation": "https://rapidai.github.io/RapidOCRDocs/docs/install_usage/rapidocr/usage/",
     "Changelog": "https://rapidai.github.io/RapidOCRDocs/docs/changelog/rapidocr/",
 }
 

python/setup_openvino.py

@@ -38,7 +38,7 @@ def get_readme():
 sys.argv = sys.argv[:2]
 
 project_urls = {
-    "Documentation": "https://rapidai.github.io/RapidOCRDocs/docs/install_usage/rapidocr/rapidocr_openvino/",
+    "Documentation": "https://rapidai.github.io/RapidOCRDocs/docs/install_usage/rapidocr/usage/",
     "Changelog": "https://rapidai.github.io/RapidOCRDocs/docs/changelog/rapidocr/",
 }