paddleocr-lite 0.1.0__tar.gz

paddleocr_lite-0.1.0/PKG-INFO

@@ -0,0 +1,9 @@
+Metadata-Version: 2.3
+Name: paddleocr-lite
+Version: 0.1.0
+Summary: Paddleocr wrapper package for simple usage.
+Requires-Dist: paddleocr>=3.7.0
+Requires-Dist: paddlepaddle>=3.3.1
+Requires-Python: >=3.12
+Description-Content-Type: text/markdown
+

paddleocr_lite-0.1.0/pyproject.toml

@@ -0,0 +1,17 @@
+[project]
+name = "paddleocr-lite"
+version = "0.1.0"
+description = "Paddleocr wrapper package for simple usage."
+readme = "README.md"
+requires-python = ">=3.12"
+dependencies = [
+    "paddleocr>=3.7.0",
+    "paddlepaddle>=3.3.1",
+]
+
+[project.scripts]
+paddleocr-lite = "paddleocr_lite.cli:main"
+
+[build-system]
+requires = ["uv_build>=0.12.6,<0.13.0"]
+build-backend = "uv_build"

paddleocr_lite-0.1.0/pyproject.toml.orig

@@ -0,0 +1,17 @@
+[project]
+name = "paddleocr-lite"
+version = "0.1.0"
+description = "Paddleocr wrapper package for simple usage."
+readme = "README.md"
+requires-python = ">=3.12"
+dependencies = [
+    "paddleocr>=3.7.0",
+    "paddlepaddle>=3.3.1",
+]
+
+[project.scripts]
+paddleocr-lite = "paddleocr_lite.cli:main"
+
+[build-system]
+requires = ["uv_build>=0.12.6,<0.13.0"]
+build-backend = "uv_build"

paddleocr_lite-0.1.0/src/paddleocr_lite/__init__.py

@@ -0,0 +1,3 @@
+from .core import SimpleOCR
+
+__all__ = ["SimpleOCR"]

paddleocr_lite-0.1.0/src/paddleocr_lite/cli.py

@@ -0,0 +1,29 @@
+import argparse
+from .core import SimpleOCR
+
+def main():
+    parser = argparse.ArgumentParser(
+        prog="paddleocr-lite",
+        description="Simple CLI wrapper around PaddleOCR"
+    )
+    parser.add_argument("image", help="Path to the image file to OCR")
+    parser.add_argument("--lang", default="en", help="Language for OCR (default: en)")
+    parser.add_argument("--detailed", action="store_true", help="Show confidence scores and bounding boxes")
+
+    args = parser.parse_args()
+
+    ocr = SimpleOCR(lang=args.lang)
+
+    if args.detailed:
+        results = ocr.read_with_confidence(args.image)
+        for item in results:
+            print(f"[{item['confidence']:.2f}] {item['text']}")
+    else:
+        text = ocr.read_text(args.image)
+        for line in text:
+            print(line)
+
+if __name__ == '__main__':
+    main()
+
+    

paddleocr_lite-0.1.0/src/paddleocr_lite/core.py

@@ -0,0 +1,29 @@
+from paddleocr import PaddleOCR
+
+class SimpleOCR:
+    def __init__(self, lang: str = "en"):
+        self._ocr = PaddleOCR(lang = lang, use_angle_cls = True)
+
+    def read_text(self, image_path: str) -> list[str]:
+        """Returns the recognized text lines from an image"""
+        result = self._ocr.ocr(image_path)
+        lines = []
+        for block in result:
+            for _, (text, confidence) in block:
+                lines.append(text)
+        return lines
+    
+    def read_with_confidence(self, image_path: str) -> list[dict]:
+        """Returns text with bounding box + confidence for more detail."""
+        result = self._ocr.ocr(image_path)
+        output = []
+        for block in result:
+            for box, (text, confidence) in block:
+                output.append(
+                    {
+                        "text": text,
+                        "confidence": confidence,
+                        "box": box,
+                    }
+                )
+        return output