feat: support to_json func (fixed issue #489)

Author: SWHL

python/rapidocr/utils/output.py

@@ -2,11 +2,12 @@
 # @Author: SWHL
 # @Contact: liekkaskono@163.com
 from dataclasses import dataclass, field
-from typing import List, Optional, Tuple, Union
+from typing import Any, Dict, List, Optional, Tuple, Union
 
 import numpy as np
 
 from .logger import Logger
+from .to_json import ToJSON
 from .to_markdown import ToMarkdown
 from .utils import save_img
 from .vis_res import VisRes
@@ -35,20 +36,23 @@ def __len__(self):
             return 0
         return len(self.txts)
 
-    def to_json(self):
-        pass
+    def to_json(self) -> Optional[List[Dict[Any, Any]]]:
+        if any(v is None for v in (self.boxes, self.txts, self.scores)):
+            logger.warning("The identified content is empty.")
+            return None
+        return ToJSON.to(self.boxes, self.txts, self.scores)
 
     def to_markdown(self) -> str:
         return ToMarkdown.to(self.boxes, self.txts)
 
-    def vis(self, save_path: Optional[str] = None):
+    def vis(self, save_path: Optional[str] = None) -> Optional[np.ndarray]:
         if self.img is None or self.boxes is None:
             logger.warning("No image or boxes to visualize.")
-            return
+            return None
 
         if self.viser is None:
             logger.error("vis instance is None")
-            return
+            return None
 
         if all(v is None for v in self.word_results):
             vis_img = self.viser(self.img, self.boxes, self.txts, self.scores)

python/rapidocr/utils/to_json.py

@@ -0,0 +1,17 @@
+# -*- encoding: utf-8 -*-
+# @Author: SWHL
+# @Contact: liekkaskono@163.com
+from typing import Any, Dict, List, Tuple
+
+import numpy as np
+
+
+class ToJSON:
+    @classmethod
+    def to(
+        cls, boxes: np.ndarray, txts: Tuple[str], scores: Tuple[float]
+    ) -> List[Dict[Any, Any]]:
+        results = []
+        for box, txt, score in zip(boxes, txts, scores):
+            results.append({"box": box.tolist(), "txt": txt, "score": score})
+        return results

python/rapidocr/utils/to_markdown.py

@@ -5,7 +5,6 @@
 class ToMarkdown:
     @classmethod
     def to(cls, boxes, txts) -> str:
-        # def to(cls, result: RapidOCROutput) -> str:
         """
         根据 OCR 结果的坐标信息，将文本还原为近似原始排版的 Markdown。