Spaces:
Running
Running
adds MCP server
Browse filesdescriptive how-to in docstring
app.py
CHANGED
|
@@ -43,6 +43,22 @@ def image_to_base64(image):
|
|
| 43 |
|
| 44 |
@spaces.GPU()
|
| 45 |
def process_image(image, task, ocr_type=None, ocr_box=None, ocr_color=None):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
if image is None:
|
| 47 |
return "Error: No image provided", None, None
|
| 48 |
|
|
@@ -290,4 +306,4 @@ with gr.Blocks(theme=gr.themes.Base()) as demo:
|
|
| 290 |
|
| 291 |
if __name__ == "__main__":
|
| 292 |
cleanup_old_files()
|
| 293 |
-
demo.launch(ssr_mode = False)
|
|
|
|
| 43 |
|
| 44 |
@spaces.GPU()
|
| 45 |
def process_image(image, task, ocr_type=None, ocr_box=None, ocr_color=None):
|
| 46 |
+
"""
|
| 47 |
+
Process an input image using the OCR model based on the specified task.
|
| 48 |
+
|
| 49 |
+
Args:
|
| 50 |
+
image (Union[dict, np.ndarray, str, PIL.Image]): Input image in various formats
|
| 51 |
+
task (str): Type of OCR task to perform
|
| 52 |
+
ocr_type (str, optional): Type of OCR processing ('ocr' or 'format')
|
| 53 |
+
ocr_box (str, optional): Bounding box coordinates for fine-grained OCR
|
| 54 |
+
ocr_color (str, optional): Color specification for fine-grained OCR
|
| 55 |
+
|
| 56 |
+
Returns:
|
| 57 |
+
tuple: (result_text, html_content, unique_id)
|
| 58 |
+
- result_text (str): OCR processing result or error message
|
| 59 |
+
- html_content (str): HTML content for visualization if applicable
|
| 60 |
+
- unique_id (str): Unique identifier for the processed image
|
| 61 |
+
"""
|
| 62 |
if image is None:
|
| 63 |
return "Error: No image provided", None, None
|
| 64 |
|
|
|
|
| 306 |
|
| 307 |
if __name__ == "__main__":
|
| 308 |
cleanup_old_files()
|
| 309 |
+
demo.launch(ssr_mode = False, mcp_server=True)
|