add pptx convert

apps/common/handle/impl/mineru/converter.py

@@ -131,15 +131,47 @@ class DocumentConverter:
             True if conversion successful, False otherwise
         """
         try:
-            cmd = [
-                self.config.libreoffice_path,
-                '--headless',
-                '--convert-to', 'pdf',
-                '--outdir', temp_dir,
-                ppt_path
-            ]
-            
-            self.logger.info(f"mineru-converter:  executing LibreOffice conversion")
+            # Try direct conversion first, fallback to xvfb if needed
+            use_xvfb = os.getenv('USE_XVFB', 'auto').lower()
+
+            if use_xvfb == 'auto':
+                # Try direct conversion first
+                cmd = [
+                    self.config.libreoffice_path,
+                    '--headless',
+                    '--convert-to', 'pdf',
+                    '--outdir', temp_dir,
+                    ppt_path
+                ]
+                use_xvfb_now = False
+                fallback_to_xvfb = True
+            elif use_xvfb == 'true':
+                # Force use xvfb
+                cmd = [
+                    'xvfb-run',
+                    '--auto-servernum',
+                    '--server-args=-screen 0, 1024x768x24',
+                    self.config.libreoffice_path,
+                    '--headless',
+                    '--convert-to', 'pdf',
+                    '--outdir', temp_dir,
+                    ppt_path
+                ]
+                use_xvfb_now = True
+                fallback_to_xvfb = False
+            else:
+                # No xvfb
+                cmd = [
+                    self.config.libreoffice_path,
+                    '--headless',
+                    '--convert-to', 'pdf',
+                    '--outdir', temp_dir,
+                    ppt_path
+                ]
+                use_xvfb_now = False
+                fallback_to_xvfb = False
+
+            self.logger.info(f"mineru-converter:  executing LibreOffice conversion (use_xvfb={use_xvfb}, fallback={fallback_to_xvfb})")
             
             # Run with timeout
             process = await asyncio.create_subprocess_exec(
@@ -156,8 +188,47 @@ class DocumentConverter:
                 
                 if process.returncode != 0:
                     self.logger.warning(f"mineru-converter:  LibreOffice returned non-zero: {stderr.decode()}")
-                    return False
-                
+                    # If direct conversion failed and we have xvfb fallback enabled
+                    if fallback_to_xvfb and not use_xvfb_now:
+                        self.logger.info(f"mineru-converter:  direct conversion failed, retrying with xvfb")
+                        fallback_cmd = [
+                            'xvfb-run',
+                            '--auto-servernum',
+                            '--server-args=-screen 0, 1024x768x24',
+                            self.config.libreoffice_path,
+                            '--headless',
+                            '--convert-to', 'pdf',
+                            '--outdir', temp_dir,
+                            ppt_path
+                        ]
+
+                        # Try again with xvfb
+                        process_xvfb = await asyncio.create_subprocess_exec(
+                            *fallback_cmd,
+                            stdout=asyncio.subprocess.PIPE,
+                            stderr=asyncio.subprocess.PIPE
+                        )
+
+                        try:
+                            stdout_xvfb, stderr_xvfb = await asyncio.wait_for(
+                                process_xvfb.communicate(),
+                                timeout=self.config.conversion_timeout
+                            )
+
+                            if process_xvfb.returncode == 0:
+                                self.logger.info(f"mineru-converter:  LibreOffice conversion with xvfb completed successfully")
+                                return True
+                            else:
+                                self.logger.warning(f"mineru-converter:  LibreOffice with xvfb also failed: {stderr_xvfb.decode()}")
+                                return False
+
+                        except asyncio.TimeoutError:
+                            self.logger.warning(f"mineru-converter:  LibreOffice with xvfb conversion timeout")
+                            process_xvfb.kill()
+                            return False
+                    else:
+                        return False
+
                 self.logger.info(f"mineru-converter:  LibreOffice conversion completed successfully")
                 return True
                 

installer/Dockerfile

@@ -76,24 +76,18 @@ ENV MAXKB_VERSION="${DOCKER_IMAGE_TAG} (build at ${BUILD_AT}, commit: ${GITHUB_C
     PIP_TARGET=/opt/maxkb/python-packages
 
 # Install poppler-utils for PDF processing (required by MinerU)
-# Install X11 libraries for LibreOffice (required for headless operation)
+# Install essential libraries for LibreOffice headless operation
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
         poppler-utils \
         libxinerama1 \
-        libxi6 \
-        libxrender1 \
-        libxtst6 \
-        libxrandr2 \
-        libxext6 \
-        libxfixes3 \
-        libxcursor1 \
-        libxcomposite1 \
-        libxdamage1 \
-        libxss1 \
-        libxt6 \
+        libdbus-1-3 \
         libsm6 \
-        libice6 && \
+        libice6 \
+        libxt6 \
+        libglib2.0-0 \
+        libcups2 \
+        xvfb && \
     apt-get clean all && \
     rm -rf /var/lib/apt/lists/*
 

installer/docker-compose.yml

@@ -33,6 +33,10 @@ services:
 
       # LibreOffice 路径配置
       LIBREOFFICE_PATH: "soffice"
+      # LibreOffice headless mode 环境变量
+      DISPLAY: ":99"
+      # xvfb 模式: auto(自动检测是否需要), true(强制使用), false(不使用)
+      USE_XVFB: "auto"
 
       # MINERU_API_TYPE: "cloud"
       # MINERU_API_URL: "https://mineru.net"