Upload 49 files

Dockerfile

@@ -46,6 +46,7 @@ RUN fc-cache -fv
 # Pre-initialize LibreOffice to avoid first-run errors
 # Run LibreOffice once to complete initial setup
 RUN mkdir -p /tmp/.config/libreoffice && \
+    chmod -R 777 /tmp/.config && \
     HOME=/tmp libreoffice --headless --version || true
 
 # Set working directory
@@ -56,7 +57,7 @@ COPY requirements.txt .
 RUN pip3 install --no-cache-dir -r requirements.txt
 
 # Create necessary directories with proper permissions
-RUN mkdir -p /tmp/libreoffice_conversion /tmp/.config && \
+RUN mkdir -p /tmp/libreoffice_conversion /tmp/.config /tmp/fonts/truetype && \
     chmod -R 777 /tmp
 
 # Create static directory
@@ -68,12 +69,19 @@ COPY . .
 # Setup additional Arabic fonts using Python script with proper permissions
 # Use /tmp for font installation to avoid permission issues
 RUN sed -i 's|/usr/share/fonts/truetype|/tmp/fonts/truetype|g' setup_fonts.py && \
-    mkdir -p /tmp/fonts/truetype && \
+    mkdir -p /tmp/fonts/truetype/arabic-enhanced && \
+    chmod -R 777 /tmp/fonts/truetype && \
     python3 setup_fonts.py || echo "Font setup failed, continuing with default fonts..."
 
 # Update font cache after installing additional fonts
 RUN fc-cache -fv
 
+# Fix LibreOffice Java integration issues
+RUN mkdir -p /tmp/.config/libreoffice/4/user && \
+    chmod -R 777 /tmp/.config/libreoffice && \
+    touch /tmp/.config/libreoffice/4/user/registrymodifications.xcu && \
+    chmod 666 /tmp/.config/libreoffice/4/user/registrymodifications.xcu
+
 # Expose port (Hugging Face Spaces requires port 7860)
 EXPOSE 7860
 

app.py

@@ -122,13 +122,23 @@ def setup_local_arial_font():
         # Create system fonts directory for local Arial in /tmp to avoid permission issues
         system_fonts_dir = Path("/tmp/fonts/truetype/local-arial")
         system_fonts_dir.mkdir(parents=True, exist_ok=True)
+        # Ensure proper permissions
+        try:
+            system_fonts_dir.chmod(0o777)
+        except PermissionError:
+            # If we can't change permissions, continue anyway
+            pass
 
         # Copy Arial font to system directory
         system_arial_path = system_fonts_dir / "arial.ttf"
         if not system_arial_path.exists():
             print("📥 Installing local Arial font...")
             shutil.copy2(arial_font_path, system_arial_path)
-            os.chmod(system_arial_path, 0o644)
+            try:
+                system_arial_path.chmod(0o644)
+            except PermissionError:
+                # If we can't change permissions, continue anyway
+                pass
             print("✅ Local Arial font installed successfully")
         else:
             print("✅ Local Arial font already installed")
@@ -150,6 +160,12 @@ def install_arabic_fonts():
         # Create fonts directory in /tmp to avoid permission issues
         fonts_dir = Path("/tmp/fonts/truetype/arabic-custom")
         fonts_dir.mkdir(parents=True, exist_ok=True)
+        # Ensure proper permissions
+        try:
+            fonts_dir.chmod(0o777)
+        except PermissionError:
+            # If we can't change permissions, continue anyway
+            pass
 
         print("🔤 Installing Arabic fonts for RTL support...")
 
@@ -172,7 +188,11 @@ def install_arabic_fonts():
                             src = os.path.join(amiri_dir, file)
                             dst = fonts_dir / file
                             shutil.copy2(src, dst)
-                            os.chmod(dst, 0o644)
+                            try:
+                                dst.chmod(0o644)
+                            except PermissionError:
+                                # If we can't change permissions, continue anyway
+                                pass
                     print("✅ Amiri font installed successfully")
                 else:
                     print("❌ Amiri font directory not found")
@@ -198,7 +218,11 @@ def install_arabic_fonts():
                             src = os.path.join(scheherazade_dir, file)
                             dst = fonts_dir / file
                             shutil.copy2(src, dst)
-                            os.chmod(dst, 0o644)
+                            try:
+                                dst.chmod(0o644)
+                            except PermissionError:
+                                # If we can't change permissions, continue anyway
+                                pass
                     print("✅ Scheherazade New font installed successfully")
                 else:
                     print("❌ Scheherazade New font directory not found")
@@ -216,7 +240,11 @@ def install_arabic_fonts():
 
                 dst = fonts_dir / "NotoSansArabic-Regular.ttf"
                 shutil.copy2(noto_file, dst)
-                os.chmod(dst, 0o644)
+                try:
+                    dst.chmod(0o644)
+                except PermissionError:
+                    # If we can't change permissions, continue anyway
+                    pass
                 print("✅ Noto Sans Arabic font installed successfully")
         except Exception as e:
             print(f"❌ Noto Sans Arabic font installation failed: {e}")
@@ -232,14 +260,22 @@ def install_arabic_fonts():
 
                 dst = fonts_dir / "Cairo-Regular.ttf"
                 shutil.copy2(cairo_file, dst)
-                os.chmod(dst, 0o644)
+                try:
+                    dst.chmod(0o644)
+                except PermissionError:
+                    # If we can't change permissions, continue anyway
+                    pass
                 print("✅ Cairo font installed successfully")
         except Exception as e:
             print(f"❌ Cairo font installation failed: {e}")
 
         # Update font cache after installation
         print("🔄 Updating font cache...")
-        subprocess.run(["fc-cache", "-f"], capture_output=True, timeout=30)
+        try:
+            subprocess.run(["fc-cache", "-f"], capture_output=True, timeout=30)
+            print("✅ Font cache updated successfully")
+        except Exception as e:
+            print(f"❌ Font cache update failed: {e}")
         print("🎯 Enhanced Arabic fonts setup completed!")
 
     except Exception as e:
@@ -1198,14 +1234,31 @@ def post_process_pdf_for_perfect_formatting(pdf_path, docx_info):
     Uses PyMuPDF to verify and correct any layout issues
     """
     try:
-        # Try to import PyMuPDF (fitz)
+        # Try to import PyMuPDF (fitz) - multiple import methods for compatibility
+        fitz = None
         try:
             import fitz  # PyMuPDF
         except ImportError:
             try:
                 from pymupdf import fitz
             except ImportError:
-                import pymupdf as fitz
+                try:
+                    import pymupdf as fitz
+                except ImportError:
+                    fitz = None
+
+        # If PyMuPDF is not available, skip advanced post-processing
+        if fitz is None:
+            print("⚠️ PyMuPDF not available - skipping advanced post-processing")
+            return {
+                'pages_processed': 0,
+                'placeholders_verified': 0,
+                'tables_verified': 0,
+                'arabic_text_verified': 0,
+                'layout_issues_fixed': 0,
+                'warnings': ['PyMuPDF not available for advanced verification'],
+                'success_metrics': ['Basic PDF validation completed']
+            }
 
         print("🔍 Post-processing PDF for perfect formatting...")
 
@@ -1326,17 +1379,6 @@ def post_process_pdf_for_perfect_formatting(pdf_path, docx_info):
 
         return post_process_results
 
-    except ImportError:
-        print("⚠️ PyMuPDF not available - skipping advanced post-processing")
-        return {
-            'pages_processed': 0,
-            'placeholders_verified': 0,
-            'tables_verified': 0,
-            'arabic_text_verified': 0,
-            'layout_issues_fixed': 0,
-            'warnings': ['PyMuPDF not available for advanced verification'],
-            'success_metrics': ['Basic PDF validation completed']
-        }
     except Exception as e:
         print(f"❌ PDF post-processing error: {e}")
         return {
@@ -1692,9 +1734,24 @@ def create_libreoffice_config(temp_path):
     """Create comprehensive LibreOffice configuration for PERFECT Arabic RTL formatting preservation"""
     config_dir = temp_path / ".config" / "libreoffice" / "4" / "user"
     config_dir.mkdir(parents=True, exist_ok=True)
+    # Ensure proper permissions
+    try:
+        config_dir.chmod(0o777)
+    except PermissionError:
+        # If we can't change permissions, continue anyway
+        pass
 
     # Create comprehensive registrymodifications.xcu for maximum formatting preservation
     registry_config = config_dir / "registrymodifications.xcu"
+    
+    # Create empty file first to ensure it exists
+    try:
+        registry_config.touch()
+        registry_config.chmod(0o666)
+    except PermissionError:
+        # If we can't change permissions, continue anyway
+        pass
+
     config_content = '''<?xml version="1.0" encoding="UTF-8"?>
 <oor:items xmlns:oor="http://openoffice.org/2001/registry" xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
   <!-- PDF Export Settings for Maximum Quality with Arabic Support -->
@@ -1975,8 +2032,11 @@ def create_libreoffice_config(temp_path):
   </item>
 </oor:items>'''
 
-    with open(registry_config, 'w', encoding='utf-8') as f:
-        f.write(config_content)
+    try:
+        with open(registry_config, 'w', encoding='utf-8') as f:
+            f.write(config_content)
+    except Exception as e:
+        print(f"❌ Failed to write LibreOffice config: {e}")
 
     return str(config_dir.parent.parent.parent)
 

main.py

@@ -111,7 +111,13 @@ async def startup_event():
     
     # Setup LibreOffice and fonts
     libreoffice_available = setup_libreoffice()
-    setup_font_environment()
+    
+    # Setup font environment with error handling
+    try:
+        setup_font_environment()
+    except Exception as e:
+        logger.warning(f"Font environment setup failed: {e}")
+        logger.warning("Continuing with default system fonts...")
     
     # Create static directory if it doesn't exist
     Path("static").mkdir(exist_ok=True)

setup_fonts.py

@@ -43,7 +43,14 @@ def setup_arabic_fonts():
     
     # Use /tmp for font installation to avoid permission issues
     fonts_dir = "/tmp/fonts/truetype/arabic-enhanced"
-    os.makedirs(fonts_dir, exist_ok=True)
+    try:
+        os.makedirs(fonts_dir, exist_ok=True)
+        # Ensure proper permissions
+        os.chmod(fonts_dir, 0o777)
+    except PermissionError:
+        # If we can't change permissions, continue anyway
+        print(f"⚠️ Permission denied when setting up {fonts_dir}, continuing anyway...")
+        pass
     
     # Try to download and install Amiri font
     print("📥 Installing Amiri font...")
@@ -59,7 +66,11 @@ def setup_arabic_fonts():
                             src = os.path.join(amiri_dir, file)
                             dst = os.path.join(fonts_dir, file)
                             shutil.copy2(src, dst)
-                            os.chmod(dst, 0o644)
+                            try:
+                                os.chmod(dst, 0o644)
+                            except PermissionError:
+                                # If we can't change permissions, continue anyway
+                                pass
                     print("✅ Amiri font installed successfully")
                     amiri_installed = True
                 else:
@@ -83,7 +94,11 @@ def setup_arabic_fonts():
                             src = os.path.join(scheherazade_dir, file)
                             dst = os.path.join(fonts_dir, file)
                             shutil.copy2(src, dst)
-                            os.chmod(dst, 0o644)
+                            try:
+                                os.chmod(dst, 0o644)
+                            except PermissionError:
+                                # If we can't change permissions, continue anyway
+                                pass
                     print("✅ Scheherazade New font installed successfully")
                     scheherazade_installed = True
                 else:
@@ -100,7 +115,11 @@ def setup_arabic_fonts():
             if os.path.exists("arial.ttf"):
                 dst = os.path.join(fonts_dir, "arial.ttf")
                 shutil.copy2("arial.ttf", dst)
-                os.chmod(dst, 0o644)
+                try:
+                    os.chmod(dst, 0o644)
+                except PermissionError:
+                    # If we can't change permissions, continue anyway
+                    pass
                 print("✅ Arial font copied successfully")
             else:
                 print("❌ Local Arial font not found")