Update app.py

app.py

@@ -3,124 +3,47 @@ import subprocess
 import tempfile
 import os
 from pdf2image import convert_from_path
-import logging
-import time
 
-# Configure logging
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger(__name__)
-
-# ===== HEALTH CHECKS =====
-def system_check():
-    """Verify all required binaries exist."""
-    required = ["libreoffice", "pdftoppm"]
-    missing = []
-    for cmd in required:
-        try:
-            subprocess.run([cmd, "--version"], 
-                          check=True, 
-                          stdout=subprocess.PIPE, 
-                          stderr=subprocess.PIPE)
-            logger.info(f"✅ {cmd} is installed")
-        except Exception:
-            missing.append(cmd)
-    
-    if missing:
-        logger.critical(f"❌ Missing: {', '.join(missing)}")
-        raise RuntimeError("System dependencies missing. Check .dockerfile")
-
-# Run checks on startup
-system_check()
-
-# ===== CORE CONVERSION =====
-def convert_docx_to_image(docx_path: str):
-    """Convert DOCX to PNG with error recovery."""
-    start_time = time.time()
-    
+def convert_docx_to_image(docx_path: str) -> str:
+    """Convert DOCX to PNG with LibreOffice (pixel-perfect)"""
     with tempfile.TemporaryDirectory() as tmpdir:
-        try:
-            # Step 1: DOCX → PDF
-            pdf_path = os.path.join(tmpdir, "output.pdf")
-            cmd = [
-                "libreoffice",
-                "--headless",
-                "--convert-to", "pdf",
-                "--outdir", tmpdir,
-                docx_path
-            ]
-            
-            logger.info("Starting DOCX to PDF conversion...")
-            subprocess.run(cmd, check=True, timeout=60)
-            
-            if not os.path.exists(pdf_path):
-                raise FileNotFoundError("PDF not generated")
-            
-            # Step 2: PDF → PNG
-            logger.info("Converting PDF to PNG...")
-            images = convert_from_path(
-                pdf_path,
-                dpi=300,
-                fmt="png",
-                poppler_path="/usr/bin",  # Critical for Hugging Face
-                thread_count=4
-            )
-            
-            if not images:
-                raise ValueError("No images generated")
-            
-            output_path = os.path.join(tmpdir, "output.png")
-            images[0].save(output_path, "PNG")
-            
-            logger.info(f"Conversion successful in {time.time()-start_time:.2f}s")
-            return output_path
-            
-        except subprocess.TimeoutExpired:
-            logger.error("Conversion timed out after 60s")
-            raise gr.Error("Timeout: Try a smaller file or upgrade hardware")
-        except Exception as e:
-            logger.error(f"Conversion failed: {str(e)}", exc_info=True)
-            raise gr.Error(f"Error: {str(e)}")
+        # Step 1: DOCX → PDF using LibreOffice's render engine
+        pdf_path = os.path.join(tmpdir, "output.pdf")
+        cmd = [
+            "libreoffice",
+            "--headless",
+            "--convert-to", "pdf",
+            "--outdir", tmpdir,
+            docx_path
+        ]
+        subprocess.run(cmd, check=True, timeout=30)
+        
+        # Step 2: PDF → PNG (300 DPI for sharpness)
+        images = convert_from_path(
+            pdf_path,
+            dpi=300,
+            fmt="png",
+            poppler_path="/usr/bin"  # Critical for Hugging Face
+        )
+        output_path = os.path.join(tmpdir, "output.png")
+        images[0].save(output_path, "PNG")
+        return output_path
 
-# ===== GRADIO UI =====
-with gr.Blocks(title="DOCX to Image Converter") as app:
-    gr.Markdown("## 📄→🖼️ Professional DOCX to Image")
-    
+# Gradio UI
+with gr.Blocks() as app:
+    gr.Markdown("## 🖥️ DOCX to Image (1:1 Conversion)")
     with gr.Row():
         input_file = gr.File(
-            label="Upload DOCX File",
+            label="Upload DOCX",
             file_types=[".docx"],
             type="filepath"
         )
-        output_image = gr.Image(
-            label="Converted Image",
-            type="filepath",
-            interactive=False
-        )
-    
-    status = gr.Textbox(label="Conversion Status", visible=False)
+        output_image = gr.Image(label="Output", type="filepath")
     
-    convert_btn = gr.Button("Convert", variant="primary")
-    
-    # Detailed output for debugging
-    with gr.Accordion("Advanced Info", open=False):
-        gr.Markdown("""
-        **Technical Details:**
-        - Uses LibreOffice 7.5 for perfect rendering
-        - 300 DPI high-quality output
-        - Handles tables, fonts, and complex layouts
-        """)
-
-    def process(file_path: str):
-        try:
-            image_path = convert_docx_to_image(file_path)
-            return image_path, "✅ Conversion successful!"
-        except Exception as e:
-            return None, f"❌ Error: {str(e)}"
-
-    convert_btn.click(
-        fn=process,
+    gr.Button("Convert").click(
+        fn=convert_docx_to_image,
         inputs=input_file,
-        outputs=[output_image, status]
+        outputs=output_image
     )
 
-app.launch(server_name="0.0.0.0")
+app.launch()