lightpdf-aipdf-mcp 0.1.136__py3-none-any.whl → 0.1.138__py3-none-any.whl

lightpdf_aipdf_mcp/common.py

@@ -2,6 +2,7 @@
 import asyncio
 import json
 import os
+import sys
 import time
 from dataclasses import dataclass
 from typing import List, Optional, Dict, Any, Tuple
@@ -39,6 +40,7 @@ class Logger:
         
         mcp_level = level_map.get(level.lower(), "info")
         
+        print(f"mcp_level: {mcp_level}, message: {message}", file=sys.stderr)
         # 直接调用session的send_log_message方法
         await self.context.session.send_log_message(mcp_level, message)
     
@@ -121,10 +123,60 @@ class BaseApiClient:
         self.logger = logger
         self.file_handler = file_handler
         self.api_key = os.getenv("API_KEY")
+        self.api_endpoint = os.getenv("API_ENDPOINT", "techsz.aoscdn.com/api")
         # 子类必须设置api_base_url
         self.api_base_url = None
 
-    async def _wait_for_task(self, client: httpx.AsyncClient, task_id: str, operation_type: str = "处理") -> str:
+    async def _create_task(self, client: httpx.AsyncClient, file_path: str, data: dict, response_action: str = "创建任务") -> str:
+        """通用任务创建方法，支持OSS、URL、本地文件三种情况
+        Args:
+            client: HTTP客户端
+            file_path: 文件路径
+            data: API参数字典
+            response_action: 日志/错误前缀
+        Returns:
+            str: 任务ID
+        """
+        await self.logger.log("info", f"正在提交{response_action}...{data}")
+        headers = {"X-API-KEY": self.api_key}
+        # 检查是否为OSS路径
+        if self.file_handler.is_oss_id(file_path):
+            data = data.copy()
+            data["resource_id"] = file_path.split("oss_id://")[1]
+            headers["Content-Type"] = "application/json"
+            response = await client.post(
+                self.api_base_url,
+                json=data,
+                headers=headers
+            )
+        elif self.file_handler.is_url(file_path):
+            file_path_mod = file_path
+            if isinstance(file_path, str) and "arxiv.org/pdf/" in file_path:
+                from urllib.parse import urlparse, urlunparse
+                url_obj = urlparse(file_path)
+                if not url_obj.path.endswith(".pdf"):
+                    new_path = url_obj.path + ".pdf"
+                    file_path_mod = urlunparse(url_obj._replace(path=new_path))
+            data = data.copy()
+            data["url"] = file_path_mod
+            headers["Content-Type"] = "application/json"
+            response = await client.post(
+                self.api_base_url,
+                json=data,
+                headers=headers
+            )
+        else:
+            with open(file_path, "rb") as f:
+                files = {"file": f}
+                response = await client.post(
+                    self.api_base_url,
+                    files=files,
+                    data=data,
+                    headers=headers
+                )
+        return await self._handle_api_response(response, response_action)
+
+    async def _wait_for_task(self, client: httpx.AsyncClient, task_id: str, operation_type: str = "处理", is_raw: bool = False) -> str | dict:
         """等待任务完成并返回下载链接
         
         Args:
@@ -158,8 +210,14 @@ class BaseApiClient:
             progress = status_result.get("progress", 0)
             
             if state == 1:  # 完成
+                if is_raw:
+                    return status_result
+                
                 download_url = status_result.get("file")
                 if not download_url:
+                    file_hash = status_result.get("file_hash")
+                    if file_hash:
+                        return file_hash
                     await self.logger.error(f"任务完成但未找到下载链接。任务状态：{json.dumps(status_result, ensure_ascii=False)}")
                 return download_url
             elif state < 0:  # 失败
@@ -189,4 +247,4 @@ class BaseApiClient:
         if "data" not in result or "task_id" not in result["data"]:
             await self.logger.error(f"无法获取任务ID。API响应：{json.dumps(result, ensure_ascii=False)}")
         
-        return result["data"]["task_id"] 
+        return result["data"]["task_id"]

lightpdf_aipdf_mcp/converter.py

@@ -98,9 +98,8 @@ class Converter(BaseApiClient):
     """PDF文档转换器"""
     def __init__(self, logger: Logger, file_handler: FileHandler):
         super().__init__(logger, file_handler)
-        api_endpoint = os.getenv("API_ENDPOINT", "techsz.aoscdn.com/api")
-        self.api_base_url = f"https://{api_endpoint}/tasks/document/conversion"
-        self.api_wkhtmltopdf_url = f"https://{api_endpoint}/tasks/document/wkhtmltopdf"
+        self.api_base_url = f"https://{self.api_endpoint}/tasks/document/conversion"
+        self.api_wkhtmltopdf_url = f"https://{self.api_endpoint}/tasks/document/wkhtmltopdf"
 
     async def add_page_numbers(self, file_path: str, start_num: int = 1, position: str = "5", margin: int = 30, password: str = None, original_name: Optional[str] = None) -> ConversionResult:
         """为PDF文档添加页码
@@ -290,7 +289,6 @@ class Converter(BaseApiClient):
                 
                 # 记录完成信息
                 await self.logger.log("info", "转换完成。可通过下载链接获取结果文件。")
-                
                 return ConversionResult(
                     success=True,
                     file_path=file_path,
@@ -311,74 +309,20 @@ class Converter(BaseApiClient):
                 )
 
     async def _create_task(self, client: httpx.AsyncClient, file_path: str, format: str, extra_params: dict = None) -> str:
-        """创建转换任务
-        
-        Args:
-            client: HTTP客户端
-            file_path: 文件路径
-            format: 目标格式，特殊格式"doc-repair"用于去除水印，"number-pdf"用于添加页码
-            extra_params: 额外API参数(可选)
-        
-        Returns:
-            str: 任务ID
-        """
-        await self.logger.log("info", "正在提交转换任务...")
-        
-        headers = {"X-API-KEY": self.api_key}
         data = {"format": format}
-        
-        # 添加额外参数
         if extra_params:
             data.update(extra_params)
-        
-        api_url = self.api_base_url
-        # html转pdf特殊接口处理
+
+        self.api_base_url = f"https://{self.api_endpoint}/tasks/document/conversion"
         if format == "pdf":
             ext = self.file_handler.get_file_extension(file_path)
-            # 支持直接转PDF的文件后缀
             direct_pdf_exts = {".docx", ".xlsx", ".pptx", ".jpg", ".jpeg", ".png", ".dwg", ".caj", ".ofd", ".html", ".txt", ".tex", ".odt"}
-            # 本地.html 或 URL 且后缀不在direct_pdf_exts都走wkhtmltopdf
             if ext == ".html" or (file_path and file_path.startswith(("http://", "https://")) and ext not in direct_pdf_exts):
-                api_url = self.api_wkhtmltopdf_url
-        
-        # 检查是否为OSS路径
-        if self.file_handler.is_oss_id(file_path):
-            # OSS路径处理方式，与URL类似，但提取resource_id
-            data["resource_id"] = file_path.split("oss_id://")[1]
-            # 使用JSON方式时添加Content-Type
-            headers["Content-Type"] = "application/json"
-            response = await client.post(
-                api_url,
-                json=data,
-                headers=headers
-            )
-        # 检查是否为URL路径
-        elif self.file_handler.is_url(file_path):
-            # arxiv.org/pdf/特殊处理
-            if isinstance(file_path, str) and "arxiv.org/pdf/" in file_path:
-                from urllib.parse import urlparse, urlunparse
-                url_obj = urlparse(file_path)
-                if not url_obj.path.endswith(".pdf"):
-                    new_path = url_obj.path + ".pdf"
-                    file_path = urlunparse(url_obj._replace(path=new_path))
-            data["url"] = file_path
-            # 使用JSON方式时添加Content-Type
-            headers["Content-Type"] = "application/json"
-            response = await client.post(
-                api_url,
-                json=data,
-                headers=headers
-            )
-        else:
-            # 对于文件上传，使用表单方式，不需要添加Content-Type
-            with open(file_path, "rb") as f:
-                files = {"file": f}
-                response = await client.post(
-                    api_url,
-                    files=files,
-                    data=data,
-                    headers=headers
-                )
-        
-        # 使用基类的方法处理API响应
-        return await self._handle_api_response(response, "创建任务") 
+                self.api_base_url = self.api_wkhtmltopdf_url
+
+        return await super()._create_task(
+            client=client,
+            file_path=file_path,
+            data=data,
+            response_action="转换任务"
+        )

lightpdf_aipdf_mcp/editor.py

@@ -31,8 +31,7 @@ class Editor(BaseApiClient):
     """PDF文档编辑器"""
     def __init__(self, logger: Logger, file_handler: FileHandler):
         super().__init__(logger, file_handler)
-        api_endpoint = os.getenv("API_ENDPOINT", "techsz.aoscdn.com/api")
-        self.api_base_url = f"https://{api_endpoint}/tasks/document/pdfedit"
+        self.api_base_url = f"https://{self.api_endpoint}/tasks/document/pdfedit"
 
     async def _validate_pdf_file(self, file_path: str) -> bool:
         """验证文件是否为PDF格式
@@ -538,68 +537,17 @@ class Editor(BaseApiClient):
                     task_id=task_id
                 )
 
-    async def _create_task(self, client: httpx.AsyncClient, file_path: str, edit_type: EditType, extra_params: Dict[str, Any] = None) -> str:
-        """创建编辑任务
-        
-        Args:
-            client: HTTP客户端
-            file_path: 文件路径
-            edit_type: 编辑操作类型
-            extra_params: 额外API参数(可选)
-        
-        Returns:
-            str: 任务ID
-        """
-        await self.logger.log("info", "正在提交PDF编辑任务...")
-        
-        headers = {"X-API-KEY": self.api_key}
+    async def _create_task(self, client: httpx.AsyncClient, file_path: str, edit_type, extra_params: dict = None) -> str:
         data = {"type": edit_type.value}
-        
-        # 添加额外参数
         if extra_params:
             data.update(extra_params)
-        
-        # 检查是否为OSS路径
-        if self.file_handler.is_oss_id(file_path):
-            # 使用JSON方式时添加Content-Type
-            headers["Content-Type"] = "application/json"
-            # OSS路径处理方式，与URL类似，但提取resource_id
-            data["resource_id"] = file_path.split("oss_id://")[1]
-            response = await client.post(
-                self.api_base_url,
-                json=data,
-                headers=headers
-            )
-        # 检查是否为URL路径
-        elif self.file_handler.is_url(file_path):
-            # arxiv.org/pdf/特殊处理
-            if isinstance(file_path, str) and "arxiv.org/pdf/" in file_path:
-                from urllib.parse import urlparse, urlunparse
-                url_obj = urlparse(file_path)
-                if not url_obj.path.endswith(".pdf"):
-                    new_path = url_obj.path + ".pdf"
-                    file_path = urlunparse(url_obj._replace(path=new_path))
-            # 使用JSON方式时添加Content-Type
-            headers["Content-Type"] = "application/json"
-            data["url"] = file_path
-            response = await client.post(
-                self.api_base_url,
-                json=data,
-                headers=headers
-            )
-        else:
-            # 对于文件上传，使用表单方式，不需要添加Content-Type
-            with open(file_path, "rb") as f:
-                files = {"file": f}
-                response = await client.post(
-                    self.api_base_url,
-                    files=files,
-                    data=data,
-                    headers=headers
-                )
-        
-        # 使用基类的方法处理API响应
-        return await self._handle_api_response(response, "创建任务")
+
+        return await super()._create_task(
+            client=client,
+            file_path=file_path,
+            data=data,
+            response_action="编辑任务"
+        )
 
     async def _create_merge_task(self, client: httpx.AsyncClient, file_paths: List[str], password: Optional[str] = None, original_name: Optional[str] = None) -> str:
         """创建PDF合并任务
@@ -612,8 +560,6 @@ class Editor(BaseApiClient):
         Returns:
             str: 任务ID
         """
-        await self.logger.log("info", "正在提交PDF合并任务...")
-        
         headers = {"X-API-KEY": self.api_key}
         data = {"type": EditType.MERGE.value}
         
@@ -652,6 +598,8 @@ class Editor(BaseApiClient):
                 # 记录本地文件，需要使用form方式
                 local_files.append(file_path)
                 
+        await self.logger.log("info", f"正在提交PDF合并任务...{data}")
+        
         # 如果全部是URL输入，使用JSON方式
         if url_inputs and not local_files:
             data["inputs"] = url_inputs

lightpdf_aipdf_mcp/ocr.py

@@ -0,0 +1,72 @@
+from dataclasses import dataclass
+import os
+import httpx
+from typing import Optional, Dict, Any
+from .common import Logger, BaseResult, FileHandler, BaseApiClient
+
+@dataclass
+class OcrResult(BaseResult):
+    """OCR结果数据类"""
+    pass
+
+class OcrClient(BaseApiClient):
+    """文档OCR识别器"""
+    def __init__(self, logger: Logger, file_handler: FileHandler):
+        super().__init__(logger, file_handler)
+        self.api_base_url = f"https://{self.api_endpoint}/tasks/document/ocr"
+
+    async def ocr_document(self, file_path: str, format: str = "pdf", language: Optional[str] = None, password: Optional[str] = None, original_name: Optional[str] = None) -> OcrResult:
+        if not self.api_key:
+            await self.logger.error("未找到API_KEY。请在客户端配置API_KEY环境变量。")
+            return OcrResult(success=False, file_path=file_path, error_message="未找到API_KEY", original_name=original_name)
+
+        # 构建API参数
+        extra_params = {
+            "format": format or "pdf"
+        }
+        if language:
+            extra_params["language"] = language
+        else:
+            extra_params["language"] = "English,Digits,ChinesePRC"
+        if password:
+            extra_params["password"] = password
+        if original_name:
+            extra_params["filename"] = os.path.splitext(original_name)[0]
+
+        async with httpx.AsyncClient(timeout=3600.0) as client:
+            task_id = None
+            try:
+                # 创建OCR任务
+                task_id = await self._create_task(client, file_path, extra_params)
+                # 等待任务完成
+                download_url = await self._wait_for_task(client, task_id, "OCR识别")
+
+                await self.logger.log("info", "OCR识别完成。可通过下载链接获取结果文件。")
+                return OcrResult(
+                    success=True,
+                    file_path=file_path,
+                    error_message=None,
+                    download_url=download_url,
+                    original_name=original_name,
+                    task_id=task_id
+                )
+            except Exception as e:
+                return OcrResult(
+                    success=False,
+                    file_path=file_path,
+                    error_message=str(e),
+                    download_url=None,
+                    original_name=original_name,
+                    task_id=task_id
+                )
+
+    async def _create_task(self, client: httpx.AsyncClient, file_path: str, extra_params: dict = None) -> str:
+        data = extra_params.copy() if extra_params else {}
+
+        # 调用基类通用方法
+        return await super()._create_task(
+            client=client,
+            file_path=file_path,
+            data=data,
+            response_action="OCR任务"
+        ) 

lightpdf_aipdf_mcp/server.py

@@ -21,6 +21,8 @@ from .common import BaseResult, Logger, FileHandler
 from .converter import Converter, ConversionResult
 from .editor import Editor, EditResult, EditType
 from .translator import Translator, TranslateResult
+from .summarizer import Summarizer
+from .ocr import OcrClient
 
 # 加载环境变量
 load_dotenv()
@@ -56,22 +58,27 @@ def generate_result_report(
     for result in results:
         if result.success:
             # 添加成功的文件信息
-            report_obj["success_files"].append({
-                "download_url": result.download_url,
+            file_info = {
                 "original_name": result.original_name,
                 "debug": {
                     "task_id": result.task_id
                 }
-            })
+            }
+            if hasattr(result, "summary") and result.summary is not None:
+                file_info["summary"] = result.summary
+            else:
+                file_info["download_url"] = result.download_url
+            report_obj["success_files"].append(file_info)
         else:
             # 添加失败的文件信息
-            report_obj["failed_files"].append({
+            file_info = {
                 "error_message": result.error_message,
                 "original_name": result.original_name,
                 "debug": {
                     "task_id": result.task_id
                 }
-            })
+            }
+            report_obj["failed_files"].append(file_info)
     
     # 返回JSON字符串
     return json.dumps(report_obj, ensure_ascii=False)
@@ -282,8 +289,44 @@ async def process_tool_call(
     editor = Editor(logger, file_handler)
     extra_params = operation_config.get("extra_params", {})
 
+    # 新增：摘要操作分支
+    if operation_config.get("is_summarize_operation"):
+        summarizer = Summarizer(logger, file_handler)
+
+        results = await process_batch_files(
+            file_objects,
+            logger,
+            lambda file_path, password, original_name: summarizer.summarize_pdf(
+                file_path=file_path,
+                prompt=extra_params.get("prompt", "Give me a summary of the document."),
+                language=extra_params.get("language", "en"),
+                password=password,
+                original_name=original_name
+            ),
+            "PDF摘要"
+        )
+        report_msg = generate_result_report(results)
+
+    # 新增：OCR操作分支
+    elif operation_config.get("is_ocr_operation"):
+        ocr_client = OcrClient(logger, file_handler)
+
+        results = await process_batch_files(
+            file_objects,
+            logger,
+            lambda file_path, password, original_name: ocr_client.ocr_document(
+                file_path=file_path,
+                format=extra_params.get("format", "pdf"),
+                language=extra_params.get("language", "English,Digits,ChinesePRC"),
+                password=password,
+                original_name=original_name
+            ),
+            "文档OCR识别"
+        )
+        report_msg = generate_result_report(results)
+
     # 新增：翻译操作分支
-    if operation_config.get("is_translate_operation"):
+    elif operation_config.get("is_translate_operation"):
         translator = Translator(logger, file_handler)
 
         results = await process_batch_files(
@@ -1076,6 +1119,64 @@ async def handle_list_tools() -> list[types.Tool]:
                 "required": ["files", "password"]
             }
         ),
+        types.Tool(
+            name="resize_pdf",
+            description="Resize PDF pages. You can specify the target page size (a0/a1/a2/a3/a4/a5/a6/letter) and/or the image resolution (dpi, e.g., 72). If not set, the corresponding property will not be changed.",
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "files": {
+                        "type": "array",
+                        "items": {
+                            "type": "object",
+                            "properties": {
+                                "path": {
+                                    "type": "string",
+                                    "description": "PDF file URL to resize, must include protocol, supports http/https/oss"
+                                },
+                                "password": {
+                                    "type": "string",
+                                    "description": "PDF document password, required if the document is password-protected"
+                                },
+                                "name": {
+                                    "type": "string",
+                                    "description": "Original filename of the document"
+                                }
+                            },
+                            "required": ["path"]
+                        },
+                        "description": "List of PDF files to resize, each containing path and optional password"
+                    },
+                    "page_size": {
+                        "type": "string",
+                        "description": "Target page size. Any valid page size name is supported (e.g., a4, letter, legal, etc.), or use width,height in points (pt, e.g., 595,842). If not set, page size will not be changed."
+                    },
+                    "resolution": {
+                        "type": "integer",
+                        "description": "Image resolution (dpi), e.g., 72. If not set, resolution will not be changed."
+                    }
+                },
+                "required": ["files"]
+            }
+        ),
+        types.Tool(
+            name="create_pdf",
+            description="Create a PDF file from LaTeX source code string only. File upload is NOT supported. If you want to convert a TEX file to PDF, please use the convert_document tool instead. This tool only accepts pure LaTeX code as input.",
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "latex_code": {
+                        "type": "string",
+                        "description": "The LaTeX source code string to be compiled into a PDF file. Only pure LaTeX code as a string is allowed; file upload, file path, or file content is NOT supported. If you have a TEX file, use the convert_document tool."
+                    },
+                    "filename": {
+                        "type": "string",
+                        "description": "The filename for the generated PDF"
+                    }
+                },
+                "required": ["latex_code", "filename"]
+            }
+        ),
         types.Tool(
             name="translate_pdf",
             description="Translate only the text in a PDF file into a specified target language and output a new PDF file. All non-text elements (such as images, tables, and layout) will remain unchanged.",
@@ -1126,8 +1227,8 @@ async def handle_list_tools() -> list[types.Tool]:
             }
         ),
         types.Tool(
-            name="resize_pdf",
-            description="Resize PDF pages. You can specify the target page size (a0/a1/a2/a3/a4/a5/a6/letter) and/or the image resolution (dpi, e.g., 72). If not set, the corresponding property will not be changed.",
+            name="ocr_document",
+            description="Perform OCR on documents. Supports PDF, DOCX, PPTX, XLSX, and TXT formats. Output as the specified format file.",
             inputSchema={
                 "type": "object",
                 "properties": {
@@ -1138,49 +1239,77 @@ async def handle_list_tools() -> list[types.Tool]:
                             "properties": {
                                 "path": {
                                     "type": "string",
-                                    "description": "PDF file URL to resize, must include protocol, supports http/https/oss"
+                                    "description": "PDF file URL, must include protocol, supports http/https/oss."
                                 },
                                 "password": {
                                     "type": "string",
-                                    "description": "PDF document password, required if the document is password-protected"
+                                    "description": "PDF document password, required if the document is password-protected."
                                 },
                                 "name": {
                                     "type": "string",
-                                    "description": "Original filename of the document"
+                                    "description": "Original filename of the document."
                                 }
                             },
                             "required": ["path"]
                         },
-                        "description": "List of PDF files to resize, each containing path and optional password"
+                        "description": "List of files to be recognized, each item contains path and optional password, name."
                     },
-                    "page_size": {
+                    "format": {
                         "type": "string",
-                        "description": "Target page size. Any valid page size name is supported (e.g., a4, letter, legal, etc.), or use width,height in points (pt, e.g., 595,842). If not set, page size will not be changed."
+                        "description": "Output format, supports pdf/docx/pptx/xlsx/txt, default is pdf.",
+                        "enum": ["pdf", "docx", "pptx", "xlsx", "txt"],
+                        "default": "pdf"
                     },
-                    "resolution": {
-                        "type": "integer",
-                        "description": "Image resolution (dpi), e.g., 72. If not set, resolution will not be changed."
+                    "language": {
+                        "type": "string",
+                        "description": "Specify the language(s) or type(s) to recognize, multiple values can be selected and separated by commas. Optional values: Abkhaz/Adyghe/Afrikaans/Agul/Albanian/Altaic/Arabic/Armenian/Awar/Aymara/Azeri/Bashkir/Basque/Belarusian/Bemba/Blackfoot/Breton/Bugotu/Bulgarian/Buryat/Catalan/Chamorro/Chechen/ChinesePRC/ChineseTaiwan/Chukcha/Chuvash/Corsican/CrimeanTatar/Croatian/Crow/Czech/Danish/Dargwa/Dungan/Dutch/English/Eskimo/Esperanto/Estonian/Even/Evenki/Faeroese/Fijian/Finnish/French/Frisian/Friulian/GaelicScottish/Gagauz/Galician/Ganda/German/Greek/Guarani/Hani/Hausa/Hawaiian/Hebrew/Hungarian/Icelandic/Ido/Indonesian/Ingush/Interlingua/Irish/Italian/Japanese/Kabardian/Kalmyk/KarachayBalkar/Karakalpak/Kasub/Kawa/Kazakh/Khakas/Khanty/Kikuyu/Kirgiz/Kongo/Korean/Koryak/Kpelle/Kumyk/Kurdish/Lak/Lappish/Latin/Latvian/LatvianGothic/Lezgin/Lithuanian/Luba/Macedonian/Malagasy/Malay/Malinke/Maltese/Mansi/Maori/Mari/Maya/Miao/Minankabaw/Mohawk/Moldavian/Mongol/Mordvin/Nahuatl/Nenets/Nivkh/Nogay/Norwegian/Nyanja/Occidental/Ojibway/Ossetic/Papiamento/PidginEnglish/Polish/PortugueseBrazilian/PortugueseStandard/Provencal/Quechua/RhaetoRomanic/Romanian/Romany/Ruanda/Rundi/Russian/Samoan/Selkup/SerbianCyrillic/SerbianLatin/Shona/Sioux/Slovak/Slovenian/Somali/Sorbian/Sotho/Spanish/Sunda/Swahili/Swazi/Swedish/Tabassaran/Tagalog/Tahitian/Tajik/Tatar/Thai/Tinpo/Tongan/Tswana/Tun/Turkish/Turkmen/Tuvin/Udmurt/UighurCyrillic/UighurLatin/Ukrainian/UzbekCyrillic/UzbekLatin/Vietnamese/Visayan/Welsh/Wolof/Xhosa/Yakut/Yiddish/Zapotec/Zulu/Basic/C++/Cobol/Fortran/Java/Pascal/Chemistry/Digits/. Default: English,Digits,ChinesePRC",
+                        "default": "English,Digits,ChinesePRC"
                     }
                 },
                 "required": ["files"]
             }
         ),
         types.Tool(
-            name="create_pdf",
-            description="Create a PDF file from LaTeX source code string only. File upload is NOT supported. If you want to convert a TEX file to PDF, please use the convert_document tool instead. This tool only accepts pure LaTeX code as input.",
+            name="summarize_document",
+            description="Summarize the content of documents and generate a concise abstract based on the user's prompt. The tool extracts and condenses the main ideas or information from the document(s) according to the user's requirements.",
             inputSchema={
                 "type": "object",
                 "properties": {
-                    "latex_code": {
+                    "files": {
+                        "type": "array",
+                        "items": {
+                            "type": "object",
+                            "properties": {
+                                "path": {
+                                    "type": "string",
+                                    "description": "PDF file URL, must include protocol, supports http/https/oss."
+                                },
+                                "password": {
+                                    "type": "string",
+                                    "description": "PDF document password, required if the document is password-protected."
+                                },
+                                "name": {
+                                    "type": "string",
+                                    "description": "Original filename of the document."
+                                }
+                            },
+                            "required": ["path"]
+                        },
+                        "description": "List of files to summarize, each containing path and optional password."
+                    },
+                    "prompt": {
                         "type": "string",
-                        "description": "The LaTeX source code string to be compiled into a PDF file. Only pure LaTeX code as a string is allowed; file upload, file path, or file content is NOT supported. If you have a TEX file, use the convert_document tool."
+                        "description": "User's requirement or instruction for the summary."
                     },
-                    "filename": {
+                    "language": {
                         "type": "string",
-                        "description": "The filename for the generated PDF"
+                        "description": "The language in which the summary should be generated. If not set, defaults to the language of the user's current query.",
+                        "enum": [
+                            "af","am","ar","as","az","ba","be","bg","bn","bo","br","bs","ca","cs","cy","da","de","el","en","es","et","eu","fa","fi","fo","fr","gl","gu","ha","haw","he","hi","hr","ht","hu","hy","id","is","it","ja","jw","ka","kk","km","kn","ko","la","lb","ln","lo","lt","lv","mg","mi","mk","ml","mn","mr","ms","mt","my","ne","nl","nn","no","oc","pa","pl","ps","pt","ro","ru","sa","sd","si","sk","sl","sn","so","sq","sr","su","sv","sw","ta","te","tg","th","tk","tl","tr","tt","uk","ur","uz","vi","yi","yo","zh"
+                        ]
                     }
                 },
-                "required": ["latex_code", "filename"]
+                "required": ["files", "prompt", "language"]
             }
         ),
     ]
@@ -1272,6 +1401,14 @@ async def handle_call_tool(name: str, arguments: dict | None) -> list[types.Text
             "is_translate_operation": True,
             "param_keys": ["source", "target", "output_type"]
         },
+        "ocr_document": {
+            "is_ocr_operation": True,
+            "param_keys": ["format", "language"]
+        },
+        "summarize_document": {
+            "is_summarize_operation": True,
+            "param_keys": ["prompt", "language"]
+        },
     }
     
     DEFAULTS = {

lightpdf_aipdf_mcp/summarizer.py

@@ -0,0 +1,125 @@
+from dataclasses import dataclass
+import os
+import httpx
+from typing import Optional
+from .common import Logger, BaseResult, FileHandler, BaseApiClient
+
+@dataclass
+class SummarizeResult(BaseResult):
+    """摘要结果数据类，结构与 TranslateResult 完全一致"""
+    summary: Optional[str] = None
+
+class Summarizer(BaseApiClient):
+    """PDF文档摘要器，结构与 Translator 完全一致"""
+    def __init__(self, logger: Logger, file_handler: FileHandler):
+        super().__init__(logger, file_handler)
+
+    async def summarize_pdf(self, file_path: str, prompt: str, language: Optional[str] = None, password: Optional[str] = None, original_name: Optional[str] = None) -> SummarizeResult:
+        if not self.api_key:
+            await self.logger.error("未找到API_KEY。请在客户端配置API_KEY环境变量。")
+            return SummarizeResult(success=False, file_path=file_path, error_message="未找到API_KEY", original_name=original_name)
+
+        # 构建API参数
+        extra_params = {
+            "po": "lightpdf"
+        }
+        if password:
+            extra_params["password"] = password
+        if original_name:
+            extra_params["filename"] = os.path.splitext(original_name)[0]
+
+        async with httpx.AsyncClient(timeout=3600.0) as client:
+            task_id = None
+            headers = {"X-API-KEY": self.api_key}
+            try:
+                # Phase 1: Embedding
+                response_action="摘要任务1"
+                self.api_base_url = f"https://{self.api_endpoint}/tasks/llm/embedding"
+
+                data = extra_params.copy() if extra_params else {}
+
+                await self.logger.log("info", f"正在提交{response_action}...{data}")
+                # 检查是否为OSS路径
+                if self.file_handler.is_oss_id(file_path):
+                    data = data.copy()
+                    data["resource_id"] = file_path.split("oss_id://")[1]
+                    headers["Content-Type"] = "application/json"
+                    response = await client.post(
+                        self.api_base_url,
+                        json=data,
+                        headers=headers
+                    )
+                elif self.file_handler.is_url(file_path):
+                    file_path_mod = file_path
+                    if isinstance(file_path, str) and "arxiv.org/pdf/" in file_path:
+                        from urllib.parse import urlparse, urlunparse
+                        url_obj = urlparse(file_path)
+                        if not url_obj.path.endswith(".pdf"):
+                            new_path = url_obj.path + ".pdf"
+                            file_path_mod = urlunparse(url_obj._replace(path=new_path))
+                    data = data.copy()
+                    data["url"] = file_path_mod
+                    headers["Content-Type"] = "application/json"
+                    response = await client.post(
+                        self.api_base_url,
+                        json=data,
+                        headers=headers
+                    )
+                else:
+                    with open(file_path, "rb") as f:
+                        files = {"file": f}
+                        response = await client.post(
+                            self.api_base_url,
+                            files=files,
+                            data=data,
+                            headers=headers
+                        )
+
+                task_id = await self._handle_api_response(response, response_action)
+                await self.logger.log("info", f"摘要任务1，task_id: {task_id}")
+
+                file_hash = await self._wait_for_task(client, task_id, "摘要1")
+
+                # Phase 2: Summarize
+                response_action="摘要任务2"
+                self.api_base_url = f"https://{self.api_endpoint}/tasks/llm/conversation"
+
+                data = extra_params.copy() if extra_params else {}
+                data["template_id"] = "63357fa3-ba37-47d5-b9c3-8b10ed0a59d6"
+                data["response_type"] = 4
+                data["file_hash"] = file_hash
+                data["prompt"] = prompt
+                data["language"] = language
+
+                await self.logger.log("info", f"正在提交{response_action}...{data}")
+                response = await client.post(
+                    self.api_base_url,
+                    json=data,
+                    headers=headers
+                )
+
+                task_id = await self._handle_api_response(response, response_action)
+                await self.logger.log("info", f"摘要任务2，task_id: {task_id}")
+
+                content = await self._wait_for_task(client, task_id, "摘要2", is_raw=True)
+
+                summary = content.get("answer", {}).get("text", "")
+
+                await self.logger.log("info", f"摘要完成。")
+                return SummarizeResult(
+                    success=True,
+                    file_path=file_path,
+                    error_message=None,
+                    summary=summary,
+                    original_name=original_name,
+                    task_id=task_id
+                )
+            except Exception as e:
+                return SummarizeResult(
+                    success=False,
+                    file_path=file_path,
+                    error_message=str(e),
+                    summary=None,
+                    original_name=original_name,
+                    task_id=task_id
+                )

lightpdf_aipdf_mcp/translator.py

@@ -13,8 +13,7 @@ class Translator(BaseApiClient):
     """PDF文档翻译器"""
     def __init__(self, logger: Logger, file_handler: FileHandler):
         super().__init__(logger, file_handler)
-        api_endpoint = os.getenv("API_ENDPOINT", "techsz.aoscdn.com/api")
-        self.api_base_url = f"https://{api_endpoint}/tasks/document/transdocument-local"
+        self.api_base_url = f"https://{self.api_endpoint}/tasks/document/transdocument-local"
 
     async def translate_pdf(self, file_path: str, source: str, target: str, output_type: str = "mono", password: Optional[str] = None, original_name: Optional[str] = None) -> TranslateResult:
         if not self.api_key:
@@ -60,42 +59,11 @@ class Translator(BaseApiClient):
                 )
 
     async def _create_task(self, client: httpx.AsyncClient, file_path: str, extra_params: dict = None) -> str:
-        await self.logger.log("info", "正在提交翻译任务...")
-        headers = {"X-API-KEY": self.api_key}
-        data = {}
-        if extra_params:
-            data.update(extra_params)
-        # 检查是否为OSS路径
-        if self.file_handler.is_oss_id(file_path):
-            data["resource_id"] = file_path.split("oss_id://")[1]
-            headers["Content-Type"] = "application/json"
-            response = await client.post(
-                self.api_base_url,
-                json=data,
-                headers=headers
-            )
-        elif self.file_handler.is_url(file_path):
-            # arxiv.org/pdf/特殊处理
-            if isinstance(file_path, str) and "arxiv.org/pdf/" in file_path:
-                from urllib.parse import urlparse, urlunparse
-                url_obj = urlparse(file_path)
-                if not url_obj.path.endswith(".pdf"):
-                    new_path = url_obj.path + ".pdf"
-                    file_path = urlunparse(url_obj._replace(path=new_path))
-            data["url"] = file_path
-            headers["Content-Type"] = "application/json"
-            response = await client.post(
-                self.api_base_url,
-                json=data,
-                headers=headers
-            )
-        else:
-            with open(file_path, "rb") as f:
-                files = {"file": f}
-                response = await client.post(
-                    self.api_base_url,
-                    files=files,
-                    data=data,
-                    headers=headers
-                )
-        return await self._handle_api_response(response, "创建翻译任务") 
+        data = extra_params.copy() if extra_params else {}
+
+        return await super()._create_task(
+            client=client,
+            file_path=file_path,
+            data=data,
+            response_action="翻译任务"
+        )

{lightpdf_aipdf_mcp-0.1.136.dist-info → lightpdf_aipdf_mcp-0.1.138.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lightpdf-aipdf-mcp
-Version: 0.1.136
+Version: 0.1.138
 Summary: MCP Server for LightPDF AI-PDF
 Author: LightPDF Team
 License: Proprietary

lightpdf_aipdf_mcp-0.1.138.dist-info/RECORD

@@ -0,0 +1,13 @@
+lightpdf_aipdf_mcp/__init__.py,sha256=PPnAgpvJLYLVOTxnHDmJAulFnHJD6wuTwS6tRGjqq6s,141
+lightpdf_aipdf_mcp/common.py,sha256=zujbjrnlz1r_VB5wi3FDwIppDBmx_z3BsJDYrovR3A0,9145
+lightpdf_aipdf_mcp/converter.py,sha256=6e-p5zh6d5ijXtTgXuBtePp4xEQVMYt6F4j29cj4Kr4,14796
+lightpdf_aipdf_mcp/create_pdf.py,sha256=oALIhOBo60D3Gu_li7d7FF0COhFfSTM-BJpc63r9iAs,2465
+lightpdf_aipdf_mcp/editor.py,sha256=BR-sWW9L7tybEPOhdc8W-uwdBoom19EPTmGDvy_2gMc,27941
+lightpdf_aipdf_mcp/ocr.py,sha256=IyzxisA6qtXcGTHZofpUYXYDdcIjUaaHcVUKpM7DH9A,2832
+lightpdf_aipdf_mcp/server.py,sha256=zuWZZcTWVIe6jpbDyqihWmyrjOjOSdydtU_EJWw8JFk,75272
+lightpdf_aipdf_mcp/summarizer.py,sha256=2QMMgo_xxlEDSd_STPh7-1lBc4VRsL4SPSTijJPyb3I,5456
+lightpdf_aipdf_mcp/translator.py,sha256=nuZa4FpsA0xeRWAEGqSPIM55aJuazJX1m32uajowo7I,2778
+lightpdf_aipdf_mcp-0.1.138.dist-info/METADATA,sha256=_q7BdfrlZznsSkdHHldpzR_-VvAOeCxqsKF_AVqkmfs,8120
+lightpdf_aipdf_mcp-0.1.138.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+lightpdf_aipdf_mcp-0.1.138.dist-info/entry_points.txt,sha256=X7TGUe52N4sYH-tYt0YUGApeJgw-efQlZA6uAZmlmr4,63
+lightpdf_aipdf_mcp-0.1.138.dist-info/RECORD,,

lightpdf_aipdf_mcp-0.1.136.dist-info/RECORD

@@ -1,11 +0,0 @@
-lightpdf_aipdf_mcp/__init__.py,sha256=PPnAgpvJLYLVOTxnHDmJAulFnHJD6wuTwS6tRGjqq6s,141
-lightpdf_aipdf_mcp/common.py,sha256=PhTf7Zg6mEgn1rTmJDHotXp-4xb2gWFf-Dy_t25qNdY,6660
-lightpdf_aipdf_mcp/converter.py,sha256=XTrMwzXUV1eG_Wlr6l0HrdL9UlEMS4ediVSrBX7YHUM,17090
-lightpdf_aipdf_mcp/create_pdf.py,sha256=oALIhOBo60D3Gu_li7d7FF0COhFfSTM-BJpc63r9iAs,2465
-lightpdf_aipdf_mcp/editor.py,sha256=cYJ6NlS9q_HJwL-Aw7mVwCT5CECMLWYlmR_ePhw_Ja4,30081
-lightpdf_aipdf_mcp/server.py,sha256=huFvfA77pcDIfU2NOD5Fk0T6vwawnoCjlifJIf_HXUc,66983
-lightpdf_aipdf_mcp/translator.py,sha256=NbFDz-mZSD4qCNQVyV0W_0x6xXwbqs_7FiBU13JAxZs,4243
-lightpdf_aipdf_mcp-0.1.136.dist-info/METADATA,sha256=MUl3xxU0cN9YC1FlIzzTwTNTpJV9wK9s5Ics3gMKwbo,8120
-lightpdf_aipdf_mcp-0.1.136.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-lightpdf_aipdf_mcp-0.1.136.dist-info/entry_points.txt,sha256=X7TGUe52N4sYH-tYt0YUGApeJgw-efQlZA6uAZmlmr4,63
-lightpdf_aipdf_mcp-0.1.136.dist-info/RECORD,,