Update app.py

app.py

@@ -7,7 +7,6 @@ import gradio as gr
 from PIL import Image
 from io import BytesIO
 
-# Get environment variables from HF Spaces secrets
 ENDPOINT = os.environ.get("VLLM_ENDPOINT")
 MODEL = os.environ.get("VLLM_MODEL")
 
@@ -16,34 +15,18 @@ if not ENDPOINT or not MODEL:
 
 
 def image_to_base64(image):
-    """Convert PIL Image to base64 string."""
     buffered = BytesIO()
     image.save(buffered, format="PNG")
     return base64.b64encode(buffered.getvalue()).decode("utf-8")
 
 
-def convert_latex_delimiters(text):
-    """Convert [ ... ] LaTeX delimiters to $ ... $ for Gradio rendering."""
-    import re
-    
-    # Replace display math: [ ... ] -> $ ... $
-    text = re.sub(r'\[\s*(.*?)\s*\]', r'$\1$', text, flags=re.DOTALL)
-    
-    return text
-
-
 def process_image(image, temperature):
-    """
-    Send image to vLLM endpoint and stream the response.
-    """
     if image is None:
         yield "Please upload an image first.", ""
         return
     
-    # Convert image to base64
     b64_image = image_to_base64(image)
     
-    # Build the payload with only image input (no text prompt)
     payload = {
         "model": MODEL,
         "messages": [
@@ -74,7 +57,7 @@ def process_image(image, temperature):
             if line:
                 line = line.decode('utf-8')
                 if line.startswith('data: '):
-                    line = line[6:]  # Remove 'data: ' prefix
+                    line = line[6:]
                     
                 if line.strip() == '[DONE]':
                     break
@@ -86,17 +69,15 @@ def process_image(image, temperature):
                         content = delta.get('content', '')
                         if content:
                             accumulated_response += content
-                            # Convert LaTeX for rendered view, keep original for raw
-                            rendered_text = convert_latex_delimiters(accumulated_response)
-                            yield rendered_text, accumulated_response
+                            yield accumulated_response, accumulated_response
                 except json.JSONDecodeError:
                     continue
                     
     except Exception as e:
-        yield f"Error: {str(e)}", f"Error: {str(e)}"
+        error_msg = f"Error: {str(e)}"
+        yield error_msg, error_msg
 
 
-# Build the Gradio Interface
 with gr.Blocks(title="📖 Image OCR", theme=gr.themes.Soft()) as demo:
     gr.Markdown(
         """
@@ -116,7 +97,7 @@ with gr.Blocks(title="📖 Image OCR", theme=gr.themes.Soft()) as demo:
                 type="pil",
                 label="🖼️ Upload Image",
                 sources=["upload", "clipboard"],
-                height=400
+                height=600
             )
             temperature = gr.Slider(
                 minimum=0.1,
@@ -131,8 +112,8 @@ with gr.Blocks(title="📖 Image OCR", theme=gr.themes.Soft()) as demo:
         with gr.Column():
             output_text = gr.Markdown(
                 label="📄 Extracted Text (Rendered)",
-                value="<div style='min-height: 400px; padding: 10px; border: 1px solid #e0e0e0; border-radius: 4px; background-color: #f9f9f9;'><em>Extracted text will appear here...</em></div>",
-                height=500
+                value="<div style='min-height: 600px; padding: 10px; border: 1px solid #e0e0e0; border-radius: 4px; background-color: #f9f9f9;'><em>Extracted text will appear here...</em></div>",
+                height=600
             )
     
     with gr.Row():
@@ -140,11 +121,10 @@ with gr.Blocks(title="📖 Image OCR", theme=gr.themes.Soft()) as demo:
             raw_output = gr.Textbox(
                 label="Raw Markdown Output",
                 placeholder="Raw text will appear here...",
-                lines=15,
+                lines=30,
                 show_copy_button=True
             )
     
-    # Event handlers
     submit_btn.click(
         fn=process_image,
         inputs=[image_input, temperature],
@@ -155,11 +135,6 @@ with gr.Blocks(title="📖 Image OCR", theme=gr.themes.Soft()) as demo:
         fn=lambda: (None, "", ""),
         outputs=[image_input, output_text, raw_output]
     )
-    
-    gr.Markdown("""
-    ---
-    **Note:** Configure endpoint via `VLLM_ENDPOINT` and `VLLM_MODEL` environment variables.
-    """)
 
 
 if __name__ == "__main__":