MCP Development Framework

MIT License
OverviewInspectSchema Related Servers Reviews Score
{
    "sourceFile": "mcp_simple_tool/tools/pdf_tool.py",
    "activeCommit": 0,
    "commits": [
        {
            "activePatchIndex": 3,
            "patches": [
                {
                    "date": 1741332266981,
                    "content": "Index: \n===================================================================\n--- \n+++ \n"
                },
                {
                    "date": 1741333308024,
                    "content": "Index: \n===================================================================\n--- \n+++ \n@@ -1,74 +1,39 @@\n-\"\"\"\n-PDF文档解析工具，提供完整的PDF解析功能，包括文本和图像提取\n-\"\"\"\n-\n import os\n import tempfile\n import shutil\n import PyPDF2\n+from pdf2image import convert_from_path\n+from PIL import Image\n import pymupdf4llm\n-import traceback\n-from typing import Dict, List, Any\n import mcp.types as types\n-from .base import BaseTool\n-from .utils.pdf_helpers import extract_images_from_pdf\n+from . import BaseTool, ToolRegistry\n \n-\n+@ToolRegistry.register\n class PdfTool(BaseTool):\n-    \"\"\"\n-    用于解析PDF文档的工具，提供完整的文本和图像提取功能\n-    \"\"\"\n+    \"\"\"PDF解析工具，用于解析PDF文件并提取文本和图片\"\"\"\n+    name = \"file\"\n+    description = \"解析PDF文件并提取文本和图片内容\"\n+    input_schema = {\n+        \"type\": \"object\",\n+        \"required\": [\"file_path\"],\n+        \"properties\": {\n+            \"file_path\": {\n+                \"type\": \"string\",\n+                \"description\": \"PDF文件的本地路径，例如'/path/to/document.pdf'\",\n+            }\n+        },\n+    }\n     \n-    @property\n-    def name(self) -> str:\n-        return \"file\"\n-    \n-    @property\n-    def description(self) -> str:\n-        return \"解析PDF文档并提取文本和图片内容\"\n-    \n-    @property\n-    def input_schema(self) -> Dict[str, Any]:\n-        return {\n-            \"type\": \"object\",\n-            \"required\": [\"file_path\"],\n-            \"properties\": {\n-                \"file_path\": {\n-                    \"type\": \"string\",\n-                    \"description\": \"PDF文件的本地路径，例如'/path/to/document.pdf'\",\n-                }\n-            },\n-        }\n-    \n-    async def execute(self, arguments: Dict[str, Any]) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n-        \"\"\"\n-        解析PDF文档\n-        \n-        Args:\n-            arguments: 参数字典，必须包含'file_path'键\n-            \n-        Returns:\n-            PDF内容列表，包括文本和图像信息\n-        \"\"\"\n+    async def execute(self, arguments: dict) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n+        \"\"\"解析PDF文件并提取文本和图片\"\"\"\n         if \"file_path\" not in arguments:\n             return [types.TextContent(\n                 type=\"text\",\n-                text=\"错误: 缺少必要参数 'file_path'\"\n+                text=\"Error: Missing required argument 'file_path'\"\n             )]\n-        \n-        return await self._parse_pdf(arguments[\"file_path\"])\n-    \n-    async def _parse_pdf(self, file_path: str) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n-        \"\"\"\n-        解析PDF文档，提取文本和图像\n-        \n-        Args:\n-            file_path: PDF文件路径\n             \n-        Returns:\n-            PDF内容列表，包括文本和图像信息\n-        \"\"\"\n+        file_path = arguments[\"file_path\"]\n         results = []\n         \n         # 添加初始状态提示\n         results.append(types.TextContent(\n@@ -136,56 +101,144 @@\n                     type=\"text\",\n                     text=md_content\n                 ))\n                 \n-                # 使用PyMuPDF提取图片（更高级的图片提取方法）\n-                results.append(types.TextContent(\n-                    type=\"text\",\n-                    text=\"正在使用高级方法提取图片，这可能需要一些时间...\"\n-                ))\n+                # 处理提取的图像\n+                image_files = []\n+                for root, dirs, files in os.walk(image_path):\n+                    for file in files:\n+                        if file.lower().endswith(('.jpg', '.jpeg', '.png')):\n+                            image_files.append(os.path.join(root, file))\n                 \n-                # 提取图片\n-                image_info = await extract_images_from_pdf(file_path, image_path)\n-                \n                 # 添加图像信息\n-                if image_info:\n-                    image_markdown = \"\\n## 提取的图像信息\\n\\n\"\n-                    image_markdown += f\"共提取了 {len(image_info)} 张图像。\\n\\n\"\n+                if image_files:\n+                    image_info = \"\\n## 提取的图像信息\\n\\n\"\n+                    image_info += f\"共提取了 {len(image_files)} 张图像。\\n\\n\"\n                     \n-                    for i, img_data in enumerate(image_info):\n-                        image_markdown += f\"### 图像 {i+1}\\n\\n\"\n-                        image_markdown += f\"- 文件名: {img_data['filename']}\\n\"\n-                        image_markdown += f\"- 页码: {img_data['page']}\\n\"\n-                        image_markdown += f\"- 尺寸: {img_data['width']}x{img_data['height']} 像素\\n\"\n-                        image_markdown += f\"- 格式: {img_data['format']}\\n\"\n-                        image_markdown += f\"- 大小: {img_data['size_bytes'] / 1024:.2f} KB\\n\"\n-                        \n-                        if 'type' in img_data and img_data['type'] == 'rendered_page':\n-                            image_markdown += f\"- 类型: 渲染页面\\n\"\n-                        else:\n-                            image_markdown += f\"- 类型: 嵌入图像\\n\"\n-                        \n-                        image_markdown += \"\\n---\\n\\n\"\n+                    for i, img_file in enumerate(image_files):\n+                        try:\n+                            with Image.open(img_file) as img:\n+                                width, height = img.size\n+                                format_name = img.format\n+                                \n+                                image_info += f\"### 图像 {i+1}\\n\\n\"\n+                                image_info += f\"- 文件名: {os.path.basename(img_file)}\\n\"\n+                                image_info += f\"- 尺寸: {width}x{height} 像素\\n\"\n+                                image_info += f\"- 格式: {format_name}\\n\\n\"\n+                                image_info += \"---\\n\\n\"\n+                        except Exception as e:\n+                            image_info += f\"### 图像 {i+1}\\n\\n\"\n+                            image_info += f\"- 文件名: {os.path.basename(img_file)}\\n\"\n+                            image_info += f\"- 错误: 无法读取图像信息: {str(e)}\\n\\n\"\n+                            image_info += \"---\\n\\n\"\n                     \n                     results.append(types.TextContent(\n                         type=\"text\",\n-                        text=image_markdown\n+                        text=image_info\n                     ))\n                 else:\n                     results.append(types.TextContent(\n                         type=\"text\",\n                         text=\"\\n## 图像信息\\n\\n未从PDF中提取到任何图像。\"\n                     ))\n                     \n             except Exception as extract_error:\n-                # 如果PymuPDF4llm提取失败，记录错误并尝试使用备用方法\n+                # 如果PymuPDF4llm提取失败，回退到原来的方法\n                 results.append(types.TextContent(\n                     type=\"text\",\n                     text=f\"警告: 使用PymuPDF4llm提取内容失败: {str(extract_error)}\\n正在尝试使用备用方法...\"\n                 ))\n                 \n-                # 这里可以添加备用提取方法，如PyPDF2等\n-                # 为简化起见，这个部分省略\n+                # 使用PyPDF2提取文本\n+                text_content = \"\"\n+                with open(file_path, 'rb') as file:\n+                    reader = PyPDF2.PdfReader(file)\n+                    num_pages = len(reader.pages)\n+                    \n+                    # 添加PDF元数据\n+                    text_content += f\"# PDF文档信息\\n\\n\"\n+                    text_content += f\"- 页数: {num_pages}\\n\"\n+                    if reader.metadata:\n+                        for key, value in reader.metadata.items():\n+                            if key.startswith('/'):\n+                                key = key[1:]  # 移除前导斜杠\n+                            if value and str(value).strip():\n+                                text_content += f\"- {key}: {value}\\n\"\n+                    \n+                    # 提取文本 - 限制页数以提高性能\n+                    max_pages_to_process = min(num_pages, 30)  # 限制处理的最大页数\n+                    if max_pages_to_process < num_pages:\n+                        text_content += f\"\\n> 注意: 由于文件较大，仅处理前 {max_pages_to_process} 页内容。\\n\"\n+                    \n+                    text_content += \"\\n## 内容摘要\\n\\n\"\n+                    \n+                    # 逐页提取文本\n+                    for page_num in range(max_pages_to_process):\n+                        # 添加进度提示\n+                        if page_num % 5 == 0 and page_num > 0:\n+                            progress_msg = f\"已处理 {page_num}/{max_pages_to_process} 页...\"\n+                            results.append(types.TextContent(type=\"text\", text=progress_msg))\n+                        \n+                        page = reader.pages[page_num]\n+                        page_text = page.extract_text()\n+                        if page_text:\n+                            text_content += f\"\\n### 第 {page_num + 1} 页\\n\\n\"\n+                            text_content += page_text + \"\\n\"\n+                \n+                # 添加文本内容到结果\n+                results.append(types.TextContent(type=\"text\", text=text_content))\n+                \n+                # 尝试使用pdf2image提取图片\n+                try:\n+                    # 限制处理的页数以提高性能\n+                    max_img_pages = min(num_pages, 10)  # 限制处理图片的最大页数\n+                    \n+                    results.append(types.TextContent(\n+                        type=\"text\",\n+                        text=f\"正在提取图片，这可能需要一些时间...\"\n+                    ))\n+                    \n+                    # 转换PDF页面为图片并保存\n+                    images = convert_from_path(\n+                        file_path, \n+                        dpi=150, \n+                        fmt=\"jpg\", \n+                        first_page=1, \n+                        last_page=max_img_pages,\n+                        thread_count=2  # 使用多线程加速\n+                    )\n+                    \n+                    # 处理每个页面图片\n+                    image_markdown = \"\\n## 图片内容\\n\\n\"\n+                    for i, img in enumerate(images):\n+                        # 保存图片到临时目录\n+                        img_path = os.path.join(temp_dir, f\"page_{i+1}.jpg\")\n+                        img.save(img_path, \"JPEG\", quality=80)\n+                        \n+                        # 获取图片尺寸\n+                        width, height = img.size\n+                        \n+                        # 添加图片信息到Markdown\n+                        image_markdown += f\"### 第 {i+1} 页图片\\n\\n\"\n+                        image_markdown += f\"- 尺寸: {width}x{height} 像素\\n\"\n+                        image_markdown += f\"- 格式: JPEG\\n\"\n+                        image_markdown += f\"- DPI: 150\\n\\n\"\n+                        \n+                        # 添加分隔线\n+                        image_markdown += \"---\\n\\n\"\n+                    \n+                    # 添加图片信息到结果\n+                    results.append(types.TextContent(\n+                        type=\"text\",\n+                        text=image_markdown\n+                    ))\n+                except Exception as img_error:\n+                    # 如果图片提取失败，添加错误信息但继续\n+                    results.append(types.TextContent(\n+                        type=\"text\",\n+                        text=f\"警告: 无法提取图片: {str(img_error)}\\n\"\n+                             f\"这可能是由于PDF文件结构复杂或缺少必要的系统依赖（如poppler-utils）。\"\n+                    ))\n             \n             # 清理临时目录\n             shutil.rmtree(temp_dir)\n             \n@@ -200,8 +253,9 @@\n             # 确保清理临时目录\n             if 'temp_dir' in locals() and os.path.exists(temp_dir):\n                 shutil.rmtree(temp_dir)\n                 \n+            import traceback\n             error_details = traceback.format_exc()\n             return [\n                 types.TextContent(\n                     type=\"text\",\n"
                },
                {
                    "date": 1741494795896,
                    "content": "Index: \n===================================================================\n--- \n+++ \n@@ -1,13 +1,18 @@\n+\"\"\"\n+PDF快速预览工具，仅提取文本内容，适用于大型PDF文件\n+\"\"\"\n+\n import os\n import tempfile\n import shutil\n import PyPDF2\n-from pdf2image import convert_from_path\n-from PIL import Image\n import pymupdf4llm\n+import traceback\n+from typing import Dict, List, Any, Optional\n import mcp.types as types\n from . import BaseTool, ToolRegistry\n+from .image_recognition_tool import ImageRecognizer\n \n @ToolRegistry.register\n class PdfTool(BaseTool):\n     \"\"\"PDF解析工具，用于解析PDF文件并提取文本和图片\"\"\"\n@@ -19,29 +24,89 @@\n         \"properties\": {\n             \"file_path\": {\n                 \"type\": \"string\",\n                 \"description\": \"PDF文件的本地路径，例如'/path/to/document.pdf'\",\n+            },\n+            \"recognize_images\": {\n+                \"type\": \"boolean\",\n+                \"description\": \"是否使用大模型识别图片内容\",\n+            },\n+            \"image_save_dir\": {\n+                \"type\": \"string\",\n+                \"description\": \"图片保存目录，默认为/host_images\",\n+            },\n+            \"api_key\": {\n+                \"type\": \"string\",\n+                \"description\": \"大模型API密钥，用于图像识别\",\n+            },\n+            \"api_base_url\": {\n+                \"type\": \"string\",\n+                \"description\": \"大模型API基础URL，用于图像识别\",\n             }\n         },\n     }\n     \n-    async def execute(self, arguments: dict) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n-        \"\"\"解析PDF文件并提取文本和图片\"\"\"\n+    async def execute(self, arguments: Dict[str, Any]) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n+        \"\"\"\n+        解析PDF文档\n+        \n+        Args:\n+            arguments: 参数字典，必须包含'file_path'键\n+            \n+        Returns:\n+            PDF文本内容列表\n+        \"\"\"\n         if \"file_path\" not in arguments:\n             return [types.TextContent(\n                 type=\"text\",\n-                text=\"Error: Missing required argument 'file_path'\"\n+                text=\"错误: 缺少必要参数 'file_path'\"\n             )]\n+        \n+        # 获取参数\n+        file_path = arguments[\"file_path\"]\n+        recognize_images = arguments.get(\"recognize_images\", False)\n+        image_save_dir = arguments.get(\"image_save_dir\", \"/host_images\")\n+        api_key = arguments.get(\"api_key\", os.environ.get(\"LLM_API_KEY\", \"\"))\n+        api_base_url = arguments.get(\"api_base_url\", os.environ.get(\"LLM_API_BASE_URL\", \"api.openai.com\"))\n+        \n+        # 如果启用图像识别但没有提供API密钥，返回错误\n+        if recognize_images and not api_key:\n+            return [types.TextContent(\n+                type=\"text\",\n+                text=\"错误: 启用图像识别功能需要提供API密钥，请通过参数提供或设置环境变量 LLM_API_KEY\"\n+            )]\n+        \n+        return await self._parse_pdf(\n+            file_path, \n+            recognize_images=recognize_images,\n+            image_save_dir=image_save_dir,\n+            api_key=api_key,\n+            api_base_url=api_base_url\n+        )\n+    \n+    async def _parse_pdf(\n+        self, \n+        file_path: str, \n+        recognize_images: bool = False,\n+        image_save_dir: str = \"/host_images\",\n+        api_key: str = \"\",\n+        api_base_url: str = \"api.openai.com\"\n+    ) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n+        \"\"\"\n+        解析PDF文件并提取文本和图片\n+        \n+        Args:\n+            file_path: PDF文件路径\n+            recognize_images: 是否使用大模型识别图片内容\n+            image_save_dir: 图片保存目录\n+            api_key: 大模型API密钥\n+            api_base_url: 大模型API基础URL\n             \n-        file_path = arguments[\"file_path\"]\n+        Returns:\n+            PDF内容列表\n+        \"\"\"\n         results = []\n         \n-        # 添加初始状态提示\n-        results.append(types.TextContent(\n-            type=\"text\",\n-            text=\"开始处理PDF文件，请稍候...\"\n-        ))\n-        \n         # 检查文件是否存在\n         if not os.path.exists(file_path):\n             return [types.TextContent(\n                 type=\"text\",\n@@ -53,8 +118,16 @@\n             temp_dir = tempfile.mkdtemp()\n             image_path = os.path.join(temp_dir, \"images\")\n             os.makedirs(image_path, exist_ok=True)\n             \n+            # 确保图片保存目录存在\n+            if recognize_images:\n+                os.makedirs(image_save_dir, exist_ok=True)\n+                results.append(types.TextContent(\n+                    type=\"text\",\n+                    text=f\"已启用图像识别功能，图片将保存到 {image_save_dir} 目录\"\n+                ))\n+            \n             # 添加文件大小信息\n             file_size_mb = os.path.getsize(file_path) / (1024 * 1024)\n             results.append(types.TextContent(\n                 type=\"text\",\n@@ -113,29 +186,79 @@\n                 if image_files:\n                     image_info = \"\\n## 提取的图像信息\\n\\n\"\n                     image_info += f\"共提取了 {len(image_files)} 张图像。\\n\\n\"\n                     \n-                    for i, img_file in enumerate(image_files):\n-                        try:\n-                            with Image.open(img_file) as img:\n-                                width, height = img.size\n-                                format_name = img.format\n+                    # 如果启用了图像识别\n+                    if recognize_images and api_key:\n+                        image_info += \"正在使用大模型识别图片内容，这可能需要一些时间...\\n\\n\"\n+                        results.append(types.TextContent(\n+                            type=\"text\",\n+                            text=image_info\n+                        ))\n+                        \n+                        # 创建图像识别器\n+                        recognizer = ImageRecognizer(api_base_url, api_key)\n+                        \n+                        # 处理每张图片\n+                        for i, img_file in enumerate(image_files):\n+                            try:\n+                                # 复制图片到保存目录\n+                                img_filename = f\"pdf_image_{os.path.basename(file_path).replace('.pdf', '')}_{i+1}.jpg\"\n+                                saved_img_path = os.path.join(image_save_dir, img_filename)\n+                                shutil.copy2(img_file, saved_img_path)\n                                 \n+                                # 识别图片内容\n+                                recognition_result = await recognizer.recognize_image(\n+                                    saved_img_path, \n+                                    \"请详细描述这张图片的内容，包括图表、文字和视觉元素\"\n+                                )\n+                                \n+                                # 添加识别结果\n+                                if recognition_result:\n+                                    image_recognition_info = f\"### 图像 {i+1} 识别结果\\n\\n\"\n+                                    image_recognition_info += f\"- 文件名: {img_filename}\\n\"\n+                                    image_recognition_info += f\"- 保存路径: {saved_img_path}\\n\\n\"\n+                                    image_recognition_info += f\"**识别内容**:\\n\\n{recognition_result}\\n\\n\"\n+                                    image_recognition_info += \"---\\n\\n\"\n+                                    \n+                                    results.append(types.TextContent(\n+                                        type=\"text\",\n+                                        text=image_recognition_info\n+                                    ))\n+                                else:\n+                                    results.append(types.TextContent(\n+                                        type=\"text\",\n+                                        text=f\"### 图像 {i+1} 识别失败\\n\\n- 文件名: {img_filename}\\n- 保存路径: {saved_img_path}\\n\\n---\\n\\n\"\n+                                    ))\n+                            except Exception as img_rec_error:\n+                                results.append(types.TextContent(\n+                                    type=\"text\",\n+                                    text=f\"### 图像 {i+1} 处理错误\\n\\n错误信息: {str(img_rec_error)}\\n\\n---\\n\\n\"\n+                                ))\n+                    else:\n+                        # 不进行图像识别，只显示图片信息\n+                        for i, img_file in enumerate(image_files):\n+                            try:\n+                                from PIL import Image\n+                                with Image.open(img_file) as img:\n+                                    width, height = img.size\n+                                    format_name = img.format\n+                                    \n+                                    image_info += f\"### 图像 {i+1}\\n\\n\"\n+                                    image_info += f\"- 文件名: {os.path.basename(img_file)}\\n\"\n+                                    image_info += f\"- 尺寸: {width}x{height} 像素\\n\"\n+                                    image_info += f\"- 格式: {format_name}\\n\\n\"\n+                                    image_info += \"---\\n\\n\"\n+                            except Exception as e:\n                                 image_info += f\"### 图像 {i+1}\\n\\n\"\n                                 image_info += f\"- 文件名: {os.path.basename(img_file)}\\n\"\n-                                image_info += f\"- 尺寸: {width}x{height} 像素\\n\"\n-                                image_info += f\"- 格式: {format_name}\\n\\n\"\n+                                image_info += f\"- 错误: 无法读取图像信息: {str(e)}\\n\\n\"\n                                 image_info += \"---\\n\\n\"\n-                        except Exception as e:\n-                            image_info += f\"### 图像 {i+1}\\n\\n\"\n-                            image_info += f\"- 文件名: {os.path.basename(img_file)}\\n\"\n-                            image_info += f\"- 错误: 无法读取图像信息: {str(e)}\\n\\n\"\n-                            image_info += \"---\\n\\n\"\n-                    \n-                    results.append(types.TextContent(\n-                        type=\"text\",\n-                        text=image_info\n-                    ))\n+                        \n+                        results.append(types.TextContent(\n+                            type=\"text\",\n+                            text=image_info\n+                        ))\n                 else:\n                     results.append(types.TextContent(\n                         type=\"text\",\n                         text=\"\\n## 图像信息\\n\\n未从PDF中提取到任何图像。\"\n@@ -197,8 +320,9 @@\n                         text=f\"正在提取图片，这可能需要一些时间...\"\n                     ))\n                     \n                     # 转换PDF页面为图片并保存\n+                    from pdf2image import convert_from_path\n                     images = convert_from_path(\n                         file_path, \n                         dpi=150, \n                         fmt=\"jpg\", \n@@ -208,30 +332,88 @@\n                     )\n                     \n                     # 处理每个页面图片\n                     image_markdown = \"\\n## 图片内容\\n\\n\"\n-                    for i, img in enumerate(images):\n-                        # 保存图片到临时目录\n-                        img_path = os.path.join(temp_dir, f\"page_{i+1}.jpg\")\n-                        img.save(img_path, \"JPEG\", quality=80)\n+                    \n+                    # 如果启用了图像识别\n+                    if recognize_images and api_key:\n+                        image_markdown += \"正在使用大模型识别图片内容，这可能需要一些时间...\\n\\n\"\n+                        results.append(types.TextContent(\n+                            type=\"text\",\n+                            text=image_markdown\n+                        ))\n                         \n-                        # 获取图片尺寸\n-                        width, height = img.size\n+                        # 创建图像识别器\n+                        recognizer = ImageRecognizer(api_base_url, api_key)\n                         \n-                        # 添加图片信息到Markdown\n-                        image_markdown += f\"### 第 {i+1} 页图片\\n\\n\"\n-                        image_markdown += f\"- 尺寸: {width}x{height} 像素\\n\"\n-                        image_markdown += f\"- 格式: JPEG\\n\"\n-                        image_markdown += f\"- DPI: 150\\n\\n\"\n+                        # 处理每张图片\n+                        for i, img in enumerate(images):\n+                            try:\n+                                # 保存图片到临时目录和保存目录\n+                                temp_img_path = os.path.join(temp_dir, f\"page_{i+1}.jpg\")\n+                                img.save(temp_img_path, \"JPEG\", quality=80)\n+                                \n+                                # 复制到保存目录\n+                                img_filename = f\"pdf_image_{os.path.basename(file_path).replace('.pdf', '')}_{i+1}.jpg\"\n+                                saved_img_path = os.path.join(image_save_dir, img_filename)\n+                                shutil.copy2(temp_img_path, saved_img_path)\n+                                \n+                                # 获取图片尺寸\n+                                width, height = img.size\n+                                \n+                                # 识别图片内容\n+                                recognition_result = await recognizer.recognize_image(\n+                                    saved_img_path, \n+                                    \"请详细描述这张图片的内容，包括图表、文字和视觉元素\"\n+                                )\n+                                \n+                                # 添加识别结果\n+                                if recognition_result:\n+                                    image_recognition_info = f\"### 第 {i+1} 页图片识别结果\\n\\n\"\n+                                    image_recognition_info += f\"- 尺寸: {width}x{height} 像素\\n\"\n+                                    image_recognition_info += f\"- 格式: JPEG\\n\"\n+                                    image_recognition_info += f\"- 保存路径: {saved_img_path}\\n\\n\"\n+                                    image_recognition_info += f\"**识别内容**:\\n\\n{recognition_result}\\n\\n\"\n+                                    image_recognition_info += \"---\\n\\n\"\n+                                    \n+                                    results.append(types.TextContent(\n+                                        type=\"text\",\n+                                        text=image_recognition_info\n+                                    ))\n+                                else:\n+                                    results.append(types.TextContent(\n+                                        type=\"text\",\n+                                        text=f\"### 第 {i+1} 页图片识别失败\\n\\n- 尺寸: {width}x{height} 像素\\n- 格式: JPEG\\n- 保存路径: {saved_img_path}\\n\\n---\\n\\n\"\n+                                    ))\n+                            except Exception as img_rec_error:\n+                                results.append(types.TextContent(\n+                                    type=\"text\",\n+                                    text=f\"### 第 {i+1} 页图片处理错误\\n\\n错误信息: {str(img_rec_error)}\\n\\n---\\n\\n\"\n+                                ))\n+                    else:\n+                        # 不进行图像识别，只显示图片信息\n+                        for i, img in enumerate(images):\n+                            # 保存图片到临时目录\n+                            img_path = os.path.join(temp_dir, f\"page_{i+1}.jpg\")\n+                            img.save(img_path, \"JPEG\", quality=80)\n+                            \n+                            # 获取图片尺寸\n+                            width, height = img.size\n+                            \n+                            # 添加图片信息到Markdown\n+                            image_markdown += f\"### 第 {i+1} 页图片\\n\\n\"\n+                            image_markdown += f\"- 尺寸: {width}x{height} 像素\\n\"\n+                            image_markdown += f\"- 格式: JPEG\\n\"\n+                            image_markdown += f\"- DPI: 150\\n\\n\"\n+                            \n+                            # 添加分隔线\n+                            image_markdown += \"---\\n\\n\"\n                         \n-                        # 添加分隔线\n-                        image_markdown += \"---\\n\\n\"\n-                    \n-                    # 添加图片信息到结果\n-                    results.append(types.TextContent(\n-                        type=\"text\",\n-                        text=image_markdown\n-                    ))\n+                        # 添加图片信息到结果\n+                        results.append(types.TextContent(\n+                            type=\"text\",\n+                            text=image_markdown\n+                        ))\n                 except Exception as img_error:\n                     # 如果图片提取失败，添加错误信息但继续\n                     results.append(types.TextContent(\n                         type=\"text\",\n"
                },
                {
                    "date": 1741495204531,
                    "content": "Index: \n===================================================================\n--- \n+++ \n@@ -1,452 +1,1 @@\n-\"\"\"\n-PDF快速预览工具，仅提取文本内容，适用于大型PDF文件\n-\"\"\"\n-\n-import os\n-import tempfile\n-import shutil\n-import PyPDF2\n-import pymupdf4llm\n-import traceback\n-from typing import Dict, List, Any, Optional\n-import mcp.types as types\n-from . import BaseTool, ToolRegistry\n-from .image_recognition_tool import ImageRecognizer\n-\n-@ToolRegistry.register\n-class PdfTool(BaseTool):\n-    \"\"\"PDF解析工具，用于解析PDF文件并提取文本和图片\"\"\"\n-    name = \"file\"\n-    description = \"解析PDF文件并提取文本和图片内容\"\n-    input_schema = {\n-        \"type\": \"object\",\n-        \"required\": [\"file_path\"],\n-        \"properties\": {\n-            \"file_path\": {\n-                \"type\": \"string\",\n-                \"description\": \"PDF文件的本地路径，例如'/path/to/document.pdf'\",\n-            },\n-            \"recognize_images\": {\n-                \"type\": \"boolean\",\n-                \"description\": \"是否使用大模型识别图片内容\",\n-            },\n-            \"image_save_dir\": {\n-                \"type\": \"string\",\n-                \"description\": \"图片保存目录，默认为/host_images\",\n-            },\n-            \"api_key\": {\n-                \"type\": \"string\",\n-                \"description\": \"大模型API密钥，用于图像识别\",\n-            },\n-            \"api_base_url\": {\n-                \"type\": \"string\",\n-                \"description\": \"大模型API基础URL，用于图像识别\",\n-            }\n-        },\n-    }\n-    \n-    async def execute(self, arguments: Dict[str, Any]) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n-        \"\"\"\n-        解析PDF文档\n-        \n-        Args:\n-            arguments: 参数字典，必须包含'file_path'键\n-            \n-        Returns:\n-            PDF文本内容列表\n-        \"\"\"\n-        if \"file_path\" not in arguments:\n-            return [types.TextContent(\n-                type=\"text\",\n-                text=\"错误: 缺少必要参数 'file_path'\"\n-            )]\n-        \n-        # 获取参数\n-        file_path = arguments[\"file_path\"]\n-        recognize_images = arguments.get(\"recognize_images\", False)\n-        image_save_dir = arguments.get(\"image_save_dir\", \"/host_images\")\n-        api_key = arguments.get(\"api_key\", os.environ.get(\"LLM_API_KEY\", \"\"))\n-        api_base_url = arguments.get(\"api_base_url\", os.environ.get(\"LLM_API_BASE_URL\", \"api.openai.com\"))\n-        \n-        # 如果启用图像识别但没有提供API密钥，返回错误\n-        if recognize_images and not api_key:\n-            return [types.TextContent(\n-                type=\"text\",\n-                text=\"错误: 启用图像识别功能需要提供API密钥，请通过参数提供或设置环境变量 LLM_API_KEY\"\n-            )]\n-        \n-        return await self._parse_pdf(\n-            file_path, \n-            recognize_images=recognize_images,\n-            image_save_dir=image_save_dir,\n-            api_key=api_key,\n-            api_base_url=api_base_url\n-        )\n-    \n-    async def _parse_pdf(\n-        self, \n-        file_path: str, \n-        recognize_images: bool = False,\n-        image_save_dir: str = \"/host_images\",\n-        api_key: str = \"\",\n-        api_base_url: str = \"api.openai.com\"\n-    ) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n-        \"\"\"\n-        解析PDF文件并提取文本和图片\n-        \n-        Args:\n-            file_path: PDF文件路径\n-            recognize_images: 是否使用大模型识别图片内容\n-            image_save_dir: 图片保存目录\n-            api_key: 大模型API密钥\n-            api_base_url: 大模型API基础URL\n-            \n-        Returns:\n-            PDF内容列表\n-        \"\"\"\n-        results = []\n-        \n-        # 检查文件是否存在\n-        if not os.path.exists(file_path):\n-            return [types.TextContent(\n-                type=\"text\",\n-                text=f\"错误: 文件不存在: {file_path}\\n请检查路径是否正确，并确保文件可访问。\"\n-            )]\n-        \n-        try:\n-            # 创建临时目录用于存储图片\n-            temp_dir = tempfile.mkdtemp()\n-            image_path = os.path.join(temp_dir, \"images\")\n-            os.makedirs(image_path, exist_ok=True)\n-            \n-            # 确保图片保存目录存在\n-            if recognize_images:\n-                os.makedirs(image_save_dir, exist_ok=True)\n-                results.append(types.TextContent(\n-                    type=\"text\",\n-                    text=f\"已启用图像识别功能，图片将保存到 {image_save_dir} 目录\"\n-                ))\n-            \n-            # 添加文件大小信息\n-            file_size_mb = os.path.getsize(file_path) / (1024 * 1024)\n-            results.append(types.TextContent(\n-                type=\"text\",\n-                text=f\"文件大小: {file_size_mb:.2f} MB\"\n-            ))\n-            \n-            # 对大文件提供警告\n-            if file_size_mb > 10:\n-                results.append(types.TextContent(\n-                    type=\"text\",\n-                    text=f\"警告: 文件较大 ({file_size_mb:.2f} MB)，处理可能需要较长时间。\"\n-                ))\n-            \n-            # 使用PymuPDF4llm提取PDF内容（包括文本和图像）\n-            try:\n-                # 获取PDF页数\n-                with open(file_path, 'rb') as file:\n-                    reader = PyPDF2.PdfReader(file)\n-                    num_pages = len(reader.pages)\n-                \n-                # 限制处理的页数\n-                max_pages = min(num_pages, 30)\n-                pages_to_process = list(range(max_pages))\n-                \n-                # 使用PymuPDF4llm提取内容\n-                md_content = pymupdf4llm.to_markdown(\n-                    doc=file_path,\n-                    pages=pages_to_process,\n-                    page_chunks=True,\n-                    write_images=True,\n-                    image_path=image_path,\n-                    image_format=\"jpg\",\n-                    dpi=150\n-                )\n-                \n-                # 如果处理的页数少于总页数，添加提示\n-                if max_pages < num_pages:\n-                    md_content = f\"# PDF文档内容（前{max_pages}页）\\n\\n> 注意: 由于文件较大，仅处理前 {max_pages} 页内容。\\n\\n{md_content}\"\n-                else:\n-                    md_content = f\"# PDF文档内容\\n\\n{md_content}\"\n-                \n-                # 添加提取的内容到结果\n-                results.append(types.TextContent(\n-                    type=\"text\",\n-                    text=md_content\n-                ))\n-                \n-                # 处理提取的图像\n-                image_files = []\n-                for root, dirs, files in os.walk(image_path):\n-                    for file in files:\n-                        if file.lower().endswith(('.jpg', '.jpeg', '.png')):\n-                            image_files.append(os.path.join(root, file))\n-                \n-                # 添加图像信息\n-                if image_files:\n-                    image_info = \"\\n## 提取的图像信息\\n\\n\"\n-                    image_info += f\"共提取了 {len(image_files)} 张图像。\\n\\n\"\n-                    \n-                    # 如果启用了图像识别\n-                    if recognize_images and api_key:\n-                        image_info += \"正在使用大模型识别图片内容，这可能需要一些时间...\\n\\n\"\n-                        results.append(types.TextContent(\n-                            type=\"text\",\n-                            text=image_info\n-                        ))\n-                        \n-                        # 创建图像识别器\n-                        recognizer = ImageRecognizer(api_base_url, api_key)\n-                        \n-                        # 处理每张图片\n-                        for i, img_file in enumerate(image_files):\n-                            try:\n-                                # 复制图片到保存目录\n-                                img_filename = f\"pdf_image_{os.path.basename(file_path).replace('.pdf', '')}_{i+1}.jpg\"\n-                                saved_img_path = os.path.join(image_save_dir, img_filename)\n-                                shutil.copy2(img_file, saved_img_path)\n-                                \n-                                # 识别图片内容\n-                                recognition_result = await recognizer.recognize_image(\n-                                    saved_img_path, \n-                                    \"请详细描述这张图片的内容，包括图表、文字和视觉元素\"\n-                                )\n-                                \n-                                # 添加识别结果\n-                                if recognition_result:\n-                                    image_recognition_info = f\"### 图像 {i+1} 识别结果\\n\\n\"\n-                                    image_recognition_info += f\"- 文件名: {img_filename}\\n\"\n-                                    image_recognition_info += f\"- 保存路径: {saved_img_path}\\n\\n\"\n-                                    image_recognition_info += f\"**识别内容**:\\n\\n{recognition_result}\\n\\n\"\n-                                    image_recognition_info += \"---\\n\\n\"\n-                                    \n-                                    results.append(types.TextContent(\n-                                        type=\"text\",\n-                                        text=image_recognition_info\n-                                    ))\n-                                else:\n-                                    results.append(types.TextContent(\n-                                        type=\"text\",\n-                                        text=f\"### 图像 {i+1} 识别失败\\n\\n- 文件名: {img_filename}\\n- 保存路径: {saved_img_path}\\n\\n---\\n\\n\"\n-                                    ))\n-                            except Exception as img_rec_error:\n-                                results.append(types.TextContent(\n-                                    type=\"text\",\n-                                    text=f\"### 图像 {i+1} 处理错误\\n\\n错误信息: {str(img_rec_error)}\\n\\n---\\n\\n\"\n-                                ))\n-                    else:\n-                        # 不进行图像识别，只显示图片信息\n-                        for i, img_file in enumerate(image_files):\n-                            try:\n-                                from PIL import Image\n-                                with Image.open(img_file) as img:\n-                                    width, height = img.size\n-                                    format_name = img.format\n-                                    \n-                                    image_info += f\"### 图像 {i+1}\\n\\n\"\n-                                    image_info += f\"- 文件名: {os.path.basename(img_file)}\\n\"\n-                                    image_info += f\"- 尺寸: {width}x{height} 像素\\n\"\n-                                    image_info += f\"- 格式: {format_name}\\n\\n\"\n-                                    image_info += \"---\\n\\n\"\n-                            except Exception as e:\n-                                image_info += f\"### 图像 {i+1}\\n\\n\"\n-                                image_info += f\"- 文件名: {os.path.basename(img_file)}\\n\"\n-                                image_info += f\"- 错误: 无法读取图像信息: {str(e)}\\n\\n\"\n-                                image_info += \"---\\n\\n\"\n-                        \n-                        results.append(types.TextContent(\n-                            type=\"text\",\n-                            text=image_info\n-                        ))\n-                else:\n-                    results.append(types.TextContent(\n-                        type=\"text\",\n-                        text=\"\\n## 图像信息\\n\\n未从PDF中提取到任何图像。\"\n-                    ))\n-                    \n-            except Exception as extract_error:\n-                # 如果PymuPDF4llm提取失败，回退到原来的方法\n-                results.append(types.TextContent(\n-                    type=\"text\",\n-                    text=f\"警告: 使用PymuPDF4llm提取内容失败: {str(extract_error)}\\n正在尝试使用备用方法...\"\n-                ))\n-                \n-                # 使用PyPDF2提取文本\n-                text_content = \"\"\n-                with open(file_path, 'rb') as file:\n-                    reader = PyPDF2.PdfReader(file)\n-                    num_pages = len(reader.pages)\n-                    \n-                    # 添加PDF元数据\n-                    text_content += f\"# PDF文档信息\\n\\n\"\n-                    text_content += f\"- 页数: {num_pages}\\n\"\n-                    if reader.metadata:\n-                        for key, value in reader.metadata.items():\n-                            if key.startswith('/'):\n-                                key = key[1:]  # 移除前导斜杠\n-                            if value and str(value).strip():\n-                                text_content += f\"- {key}: {value}\\n\"\n-                    \n-                    # 提取文本 - 限制页数以提高性能\n-                    max_pages_to_process = min(num_pages, 30)  # 限制处理的最大页数\n-                    if max_pages_to_process < num_pages:\n-                        text_content += f\"\\n> 注意: 由于文件较大，仅处理前 {max_pages_to_process} 页内容。\\n\"\n-                    \n-                    text_content += \"\\n## 内容摘要\\n\\n\"\n-                    \n-                    # 逐页提取文本\n-                    for page_num in range(max_pages_to_process):\n-                        # 添加进度提示\n-                        if page_num % 5 == 0 and page_num > 0:\n-                            progress_msg = f\"已处理 {page_num}/{max_pages_to_process} 页...\"\n-                            results.append(types.TextContent(type=\"text\", text=progress_msg))\n-                        \n-                        page = reader.pages[page_num]\n-                        page_text = page.extract_text()\n-                        if page_text:\n-                            text_content += f\"\\n### 第 {page_num + 1} 页\\n\\n\"\n-                            text_content += page_text + \"\\n\"\n-                \n-                # 添加文本内容到结果\n-                results.append(types.TextContent(type=\"text\", text=text_content))\n-                \n-                # 尝试使用pdf2image提取图片\n-                try:\n-                    # 限制处理的页数以提高性能\n-                    max_img_pages = min(num_pages, 10)  # 限制处理图片的最大页数\n-                    \n-                    results.append(types.TextContent(\n-                        type=\"text\",\n-                        text=f\"正在提取图片，这可能需要一些时间...\"\n-                    ))\n-                    \n-                    # 转换PDF页面为图片并保存\n-                    from pdf2image import convert_from_path\n-                    images = convert_from_path(\n-                        file_path, \n-                        dpi=150, \n-                        fmt=\"jpg\", \n-                        first_page=1, \n-                        last_page=max_img_pages,\n-                        thread_count=2  # 使用多线程加速\n-                    )\n-                    \n-                    # 处理每个页面图片\n-                    image_markdown = \"\\n## 图片内容\\n\\n\"\n-                    \n-                    # 如果启用了图像识别\n-                    if recognize_images and api_key:\n-                        image_markdown += \"正在使用大模型识别图片内容，这可能需要一些时间...\\n\\n\"\n-                        results.append(types.TextContent(\n-                            type=\"text\",\n-                            text=image_markdown\n-                        ))\n-                        \n-                        # 创建图像识别器\n-                        recognizer = ImageRecognizer(api_base_url, api_key)\n-                        \n-                        # 处理每张图片\n-                        for i, img in enumerate(images):\n-                            try:\n-                                # 保存图片到临时目录和保存目录\n-                                temp_img_path = os.path.join(temp_dir, f\"page_{i+1}.jpg\")\n-                                img.save(temp_img_path, \"JPEG\", quality=80)\n-                                \n-                                # 复制到保存目录\n-                                img_filename = f\"pdf_image_{os.path.basename(file_path).replace('.pdf', '')}_{i+1}.jpg\"\n-                                saved_img_path = os.path.join(image_save_dir, img_filename)\n-                                shutil.copy2(temp_img_path, saved_img_path)\n-                                \n-                                # 获取图片尺寸\n-                                width, height = img.size\n-                                \n-                                # 识别图片内容\n-                                recognition_result = await recognizer.recognize_image(\n-                                    saved_img_path, \n-                                    \"请详细描述这张图片的内容，包括图表、文字和视觉元素\"\n-                                )\n-                                \n-                                # 添加识别结果\n-                                if recognition_result:\n-                                    image_recognition_info = f\"### 第 {i+1} 页图片识别结果\\n\\n\"\n-                                    image_recognition_info += f\"- 尺寸: {width}x{height} 像素\\n\"\n-                                    image_recognition_info += f\"- 格式: JPEG\\n\"\n-                                    image_recognition_info += f\"- 保存路径: {saved_img_path}\\n\\n\"\n-                                    image_recognition_info += f\"**识别内容**:\\n\\n{recognition_result}\\n\\n\"\n-                                    image_recognition_info += \"---\\n\\n\"\n-                                    \n-                                    results.append(types.TextContent(\n-                                        type=\"text\",\n-                                        text=image_recognition_info\n-                                    ))\n-                                else:\n-                                    results.append(types.TextContent(\n-                                        type=\"text\",\n-                                        text=f\"### 第 {i+1} 页图片识别失败\\n\\n- 尺寸: {width}x{height} 像素\\n- 格式: JPEG\\n- 保存路径: {saved_img_path}\\n\\n---\\n\\n\"\n-                                    ))\n-                            except Exception as img_rec_error:\n-                                results.append(types.TextContent(\n-                                    type=\"text\",\n-                                    text=f\"### 第 {i+1} 页图片处理错误\\n\\n错误信息: {str(img_rec_error)}\\n\\n---\\n\\n\"\n-                                ))\n-                    else:\n-                        # 不进行图像识别，只显示图片信息\n-                        for i, img in enumerate(images):\n-                            # 保存图片到临时目录\n-                            img_path = os.path.join(temp_dir, f\"page_{i+1}.jpg\")\n-                            img.save(img_path, \"JPEG\", quality=80)\n-                            \n-                            # 获取图片尺寸\n-                            width, height = img.size\n-                            \n-                            # 添加图片信息到Markdown\n-                            image_markdown += f\"### 第 {i+1} 页图片\\n\\n\"\n-                            image_markdown += f\"- 尺寸: {width}x{height} 像素\\n\"\n-                            image_markdown += f\"- 格式: JPEG\\n\"\n-                            image_markdown += f\"- DPI: 150\\n\\n\"\n-                            \n-                            # 添加分隔线\n-                            image_markdown += \"---\\n\\n\"\n-                        \n-                        # 添加图片信息到结果\n-                        results.append(types.TextContent(\n-                            type=\"text\",\n-                            text=image_markdown\n-                        ))\n-                except Exception as img_error:\n-                    # 如果图片提取失败，添加错误信息但继续\n-                    results.append(types.TextContent(\n-                        type=\"text\",\n-                        text=f\"警告: 无法提取图片: {str(img_error)}\\n\"\n-                             f\"这可能是由于PDF文件结构复杂或缺少必要的系统依赖（如poppler-utils）。\"\n-                    ))\n-            \n-            # 清理临时目录\n-            shutil.rmtree(temp_dir)\n-            \n-            # 添加处理完成的提示\n-            results.append(types.TextContent(\n-                type=\"text\",\n-                text=\"PDF处理完成！\"\n-            ))\n-            \n-            return results\n-        except Exception as e:\n-            # 确保清理临时目录\n-            if 'temp_dir' in locals() and os.path.exists(temp_dir):\n-                shutil.rmtree(temp_dir)\n-                \n-            import traceback\n-            error_details = traceback.format_exc()\n-            return [\n-                types.TextContent(\n-                    type=\"text\",\n-                    text=f\"错误: 解析PDF文件失败: {str(e)}\\n\"\n-                         f\"可能的原因:\\n\"\n-                         f\"1. 文件格式不兼容\\n\"\n-                         f\"2. 文件已加密或受密码保护\\n\"\n-                         f\"3. 系统缺少必要的依赖（如poppler-utils）\\n\"\n-                         f\"4. 文件太大，处理超时\\n\\n\"\n-                         f\"详细错误信息: {error_details}\"\n-                )\n-            ] \n\\ No newline at end of file\n+ \n\\ No newline at end of file\n"
                }
            ],
            "date": 1741332266981,
            "name": "Commit-0",
            "content": "\"\"\"\nPDF文档解析工具，提供完整的PDF解析功能，包括文本和图像提取\n\"\"\"\n\nimport os\nimport tempfile\nimport shutil\nimport PyPDF2\nimport pymupdf4llm\nimport traceback\nfrom typing import Dict, List, Any\nimport mcp.types as types\nfrom .base import BaseTool\nfrom .utils.pdf_helpers import extract_images_from_pdf\n\n\nclass PdfTool(BaseTool):\n    \"\"\"\n    用于解析PDF文档的工具，提供完整的文本和图像提取功能\n    \"\"\"\n    \n    @property\n    def name(self) -> str:\n        return \"file\"\n    \n    @property\n    def description(self) -> str:\n        return \"解析PDF文档并提取文本和图片内容\"\n    \n    @property\n    def input_schema(self) -> Dict[str, Any]:\n        return {\n            \"type\": \"object\",\n            \"required\": [\"file_path\"],\n            \"properties\": {\n                \"file_path\": {\n                    \"type\": \"string\",\n                    \"description\": \"PDF文件的本地路径，例如'/path/to/document.pdf'\",\n                }\n            },\n        }\n    \n    async def execute(self, arguments: Dict[str, Any]) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n        \"\"\"\n        解析PDF文档\n        \n        Args:\n            arguments: 参数字典，必须包含'file_path'键\n            \n        Returns:\n            PDF内容列表，包括文本和图像信息\n        \"\"\"\n        if \"file_path\" not in arguments:\n            return [types.TextContent(\n                type=\"text\",\n                text=\"错误: 缺少必要参数 'file_path'\"\n            )]\n        \n        return await self._parse_pdf(arguments[\"file_path\"])\n    \n    async def _parse_pdf(self, file_path: str) -> List[types.TextContent | types.ImageContent | types.EmbeddedResource]:\n        \"\"\"\n        解析PDF文档，提取文本和图像\n        \n        Args:\n            file_path: PDF文件路径\n            \n        Returns:\n            PDF内容列表，包括文本和图像信息\n        \"\"\"\n        results = []\n        \n        # 添加初始状态提示\n        results.append(types.TextContent(\n            type=\"text\",\n            text=\"开始处理PDF文件，请稍候...\"\n        ))\n        \n        # 检查文件是否存在\n        if not os.path.exists(file_path):\n            return [types.TextContent(\n                type=\"text\",\n                text=f\"错误: 文件不存在: {file_path}\\n请检查路径是否正确，并确保文件可访问。\"\n            )]\n        \n        try:\n            # 创建临时目录用于存储图片\n            temp_dir = tempfile.mkdtemp()\n            image_path = os.path.join(temp_dir, \"images\")\n            os.makedirs(image_path, exist_ok=True)\n            \n            # 添加文件大小信息\n            file_size_mb = os.path.getsize(file_path) / (1024 * 1024)\n            results.append(types.TextContent(\n                type=\"text\",\n                text=f\"文件大小: {file_size_mb:.2f} MB\"\n            ))\n            \n            # 对大文件提供警告\n            if file_size_mb > 10:\n                results.append(types.TextContent(\n                    type=\"text\",\n                    text=f\"警告: 文件较大 ({file_size_mb:.2f} MB)，处理可能需要较长时间。\"\n                ))\n            \n            # 使用PymuPDF4llm提取PDF内容（包括文本和图像）\n            try:\n                # 获取PDF页数\n                with open(file_path, 'rb') as file:\n                    reader = PyPDF2.PdfReader(file)\n                    num_pages = len(reader.pages)\n                \n                # 限制处理的页数\n                max_pages = min(num_pages, 30)\n                pages_to_process = list(range(max_pages))\n                \n                # 使用PymuPDF4llm提取内容\n                md_content = pymupdf4llm.to_markdown(\n                    doc=file_path,\n                    pages=pages_to_process,\n                    page_chunks=True,\n                    write_images=True,\n                    image_path=image_path,\n                    image_format=\"jpg\",\n                    dpi=150\n                )\n                \n                # 如果处理的页数少于总页数，添加提示\n                if max_pages < num_pages:\n                    md_content = f\"# PDF文档内容（前{max_pages}页）\\n\\n> 注意: 由于文件较大，仅处理前 {max_pages} 页内容。\\n\\n{md_content}\"\n                else:\n                    md_content = f\"# PDF文档内容\\n\\n{md_content}\"\n                \n                # 添加提取的内容到结果\n                results.append(types.TextContent(\n                    type=\"text\",\n                    text=md_content\n                ))\n                \n                # 使用PyMuPDF提取图片（更高级的图片提取方法）\n                results.append(types.TextContent(\n                    type=\"text\",\n                    text=\"正在使用高级方法提取图片，这可能需要一些时间...\"\n                ))\n                \n                # 提取图片\n                image_info = await extract_images_from_pdf(file_path, image_path)\n                \n                # 添加图像信息\n                if image_info:\n                    image_markdown = \"\\n## 提取的图像信息\\n\\n\"\n                    image_markdown += f\"共提取了 {len(image_info)} 张图像。\\n\\n\"\n                    \n                    for i, img_data in enumerate(image_info):\n                        image_markdown += f\"### 图像 {i+1}\\n\\n\"\n                        image_markdown += f\"- 文件名: {img_data['filename']}\\n\"\n                        image_markdown += f\"- 页码: {img_data['page']}\\n\"\n                        image_markdown += f\"- 尺寸: {img_data['width']}x{img_data['height']} 像素\\n\"\n                        image_markdown += f\"- 格式: {img_data['format']}\\n\"\n                        image_markdown += f\"- 大小: {img_data['size_bytes'] / 1024:.2f} KB\\n\"\n                        \n                        if 'type' in img_data and img_data['type'] == 'rendered_page':\n                            image_markdown += f\"- 类型: 渲染页面\\n\"\n                        else:\n                            image_markdown += f\"- 类型: 嵌入图像\\n\"\n                        \n                        image_markdown += \"\\n---\\n\\n\"\n                    \n                    results.append(types.TextContent(\n                        type=\"text\",\n                        text=image_markdown\n                    ))\n                else:\n                    results.append(types.TextContent(\n                        type=\"text\",\n                        text=\"\\n## 图像信息\\n\\n未从PDF中提取到任何图像。\"\n                    ))\n                    \n            except Exception as extract_error:\n                # 如果PymuPDF4llm提取失败，记录错误并尝试使用备用方法\n                results.append(types.TextContent(\n                    type=\"text\",\n                    text=f\"警告: 使用PymuPDF4llm提取内容失败: {str(extract_error)}\\n正在尝试使用备用方法...\"\n                ))\n                \n                # 这里可以添加备用提取方法，如PyPDF2等\n                # 为简化起见，这个部分省略\n            \n            # 清理临时目录\n            shutil.rmtree(temp_dir)\n            \n            # 添加处理完成的提示\n            results.append(types.TextContent(\n                type=\"text\",\n                text=\"PDF处理完成！\"\n            ))\n            \n            return results\n        except Exception as e:\n            # 确保清理临时目录\n            if 'temp_dir' in locals() and os.path.exists(temp_dir):\n                shutil.rmtree(temp_dir)\n                \n            error_details = traceback.format_exc()\n            return [\n                types.TextContent(\n                    type=\"text\",\n                    text=f\"错误: 解析PDF文件失败: {str(e)}\\n\"\n                         f\"可能的原因:\\n\"\n                         f\"1. 文件格式不兼容\\n\"\n                         f\"2. 文件已加密或受密码保护\\n\"\n                         f\"3. 系统缺少必要的依赖（如poppler-utils）\\n\"\n                         f\"4. 文件太大，处理超时\\n\\n\"\n                         f\"详细错误信息: {error_details}\"\n                )\n            ] "
        }
    ]
}