chore: update files

Author: SWHL

python/demo.py

@@ -19,28 +19,19 @@
     img = f.read()
 
 # result, elapse_list = engine(img, use_det=True, use_cls=False, use_rec=False)
-result, elapse_list = engine(img, return_word_box=True)
+result = engine(img, return_word_box=True)
 print(result)
-print(elapse_list)
+print(result.elapse)
 
-# result, elapse = engine(image_path, use_det=False, use_cls=False, use_rec=True)
+boxes = result.boxes
+txts = result.txts
+scores = result.scores
 
-(boxes, txts, scores, words_boxes, words, words_scores) = list(zip(*result))
 font_path = "resources/fonts/FZYTK.TTF"
+vis_img = vis(img, result.boxes, result.txts, result.scores, font_path)
+cv2.imwrite("vis.png", vis_img)
 
-words_boxes = sum(words_boxes, [])
-words_all = sum(words, [])
-words_scores = sum(words_scores, [])
-vis_img = vis(img, words_boxes, words_all, words_scores, font_path)
+words_results = result.word_results
+words, words_scores, words_boxes = list(zip(*words_results))
+vis_img = vis(img, words_boxes, words, words_scores, font_path)
 cv2.imwrite("vis_single.png", vis_img)
-
-# (boxes, txts, scores, words_boxes, words) = list(zip(*result))
-
-# vis_img = vis(img, boxes, txts, scores, font_path)
-# cv2.imwrite("vis.png", vis_img)
-
-# words_boxes = sum(words_boxes, [])
-# words_all = sum(words, [])
-# words_scores = [1.0] * len(words_boxes)
-# vis_img = vis(img, words_boxes, words_all, words_scores, font_path)
-# cv2.imwrite("vis_single.png", vis_img)

python/rapidocr/cal_rec_boxes/main.py

@@ -28,7 +28,7 @@ def __call__(
         for idx, (img, box) in enumerate(zip(imgs, dt_boxes)):
             direction = self.get_box_direction(box)
 
-            rec_txt = rec_res.line_txts[idx]
+            rec_txt = rec_res.txts[idx]
             rec_word_info = rec_res.word_results[idx]
 
             h, w = img.shape[:2]

python/rapidocr/ch_ppocr_cls/utils.py

@@ -23,6 +23,11 @@ class TextClsOutput:
     cls_res: Optional[List[Tuple[str, float]]] = None
     elapse: Optional[float] = None
 
+    def __len__(self):
+        if self.img_list is None:
+            return 0
+        return len(self.img_list)
+
 
 class ClsPostProcess:
     def __init__(self, label_list: List[str]):

python/rapidocr/ch_ppocr_det/utils.py

@@ -16,6 +16,11 @@ class TextDetOutput:
     scores: Optional[Tuple[float]] = None
     elapse: float = 0.0
 
+    def __len__(self):
+        if self.boxes is None:
+            return 0
+        return len(self.boxes)
+
 
 class DetPreProcess:
     def __init__(

python/rapidocr/ch_ppocr_rec/main.py

@@ -90,8 +90,8 @@ def __call__(self, args: TextRecInput) -> TextRecOutput:
             elapse += time.perf_counter() - start_time
 
         all_line_results, all_word_results = list(zip(*rec_res))
-        line_txts, line_scores = list(zip(*all_line_results))
-        return TextRecOutput(line_txts, line_scores, all_word_results, elapse)
+        txts, scores = list(zip(*all_line_results))
+        return TextRecOutput(txts, scores, all_word_results, elapse)
 
     def resize_norm_img(self, img: np.ndarray, max_wh_ratio: float) -> np.ndarray:
         img_channel, img_height, img_width = self.rec_image_shape

python/rapidocr/ch_ppocr_rec/utils.py

@@ -29,13 +29,18 @@ class TextRecInput:
 
 @dataclass
 class TextRecOutput:
-    line_txts: Optional[Tuple[str]] = None
-    line_scores: Tuple[float] = (1.0,)
+    txts: Optional[Tuple[str]] = None
+    scores: Tuple[float] = (1.0,)
     word_results: Tuple[Tuple[str, float, Optional[List[List[int]]]]] = (
         ("", 1.0, None),
     )
     elapse: Optional[float] = None
 
+    def __len__(self):
+        if self.txts is None:
+            return 0
+        return len(self.txts)
+
 
 class CTCLabelDecode:
     def __init__(

python/rapidocr/main.py

@@ -64,7 +64,7 @@ def __init__(self, config_path: Optional[str] = None, **kwargs):
 
         self.cal_rec_boxes = CalRecBoxes()
 
-        self.return_paddleocr_format = True
+        self.return_paddleocr_format = False
 
     def __call__(
         self,
@@ -255,7 +255,7 @@ def get_final_res(
         self, det_res: TextDetOutput, cls_res: TextClsOutput, rec_res: TextRecOutput
     ) -> Union[TextDetOutput, TextClsOutput, TextRecOutput, RapidOCROutput]:
         dt_boxes = det_res.boxes
-        txt_res = rec_res.line_txts
+        txt_res = rec_res.txts
 
         if dt_boxes is None and txt_res is None and cls_res.cls_res is not None:
             return cls_res
@@ -266,13 +266,13 @@ def get_final_res(
         if dt_boxes is None and txt_res is not None:
             return rec_res
 
-        if dt_boxes is not None and rec_res is None:
+        if dt_boxes is not None and txt_res is None:
             return det_res
 
         ocr_res = RapidOCROutput(
             boxes=det_res.boxes,
-            txts=rec_res.line_txts,
-            scores=rec_res.line_scores,
+            txts=rec_res.txts,
+            scores=rec_res.scores,
             word_results=rec_res.word_results,
             elapse_list=[det_res.elapse, cls_res.elapse, rec_res.elapse],
         )

python/rapidocr/utils/typings.py

@@ -17,13 +17,19 @@ class RapidOCROutput:
     elapse: float = field(init=False)
 
     def __post_init__(self):
-        self.elapse = sum(self.elapse_list)
+        self.elapse = sum(v for v in self.elapse_list if isinstance(v, float))
+
+    def __len__(self):
+        if self.txts is None:
+            return 0
+        return len(self.txts)
 
     def to_json(self):
         pass
 
     def to_paddleocr_format(self):
-        """Return format like:
+        """Refer: https://pypi.org/project/paddleocr/2.7.3/
+        Return format like:
         [
           [[[6.0, 2.0], [322.0, 9.0], [320.0, 104.0], [4.0, 97.0]], ['正品促销', 0.99893]],
           [[[70.0, 98.0], [252.0, 98.0], [252.0, 125.0], [70.0, 125.0]], ['大桶装更划算', 0.9843]]