lightpdf-aipdf-mcp 0.1.135__py3-none-any.whl → 0.1.137__py3-none-any.whl

lightpdf_aipdf_mcp/ocr.py

@@ -0,0 +1,96 @@
+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)
+        api_endpoint = os.getenv("API_ENDPOINT", "techsz.aoscdn.com/api")
+        self.api_base_url = f"https://{api_endpoint}/tasks/document/ocr"
+
+    async def ocr_document(self, file_path: str, format: str = "pdf", password: Optional[str] = None, original_name: Optional[str] = None, language: 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:
+        await self.logger.log("info", "正在提交OCR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):
+            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OCR任务") 

lightpdf_aipdf_mcp/server.py

@@ -6,6 +6,7 @@ import sys
 import argparse
 import json
 from typing import List, Dict, Any, Callable, TypeVar, Optional, Union
+from urllib.request import url2pathname
 
 # 第三方库导入
 from dotenv import load_dotenv
@@ -281,8 +282,25 @@ async def process_tool_call(
     editor = Editor(logger, file_handler)
     extra_params = operation_config.get("extra_params", {})
 
+    # 新增：OCR操作分支
+    if operation_config.get("is_ocr_operation"):
+        from .ocr import OcrClient
+        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"),
+                password=password,
+                original_name=original_name,
+                language=extra_params.get("language", "English,Digits,ChinesePRC")
+            ),
+            "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(
@@ -1124,6 +1142,49 @@ async def handle_list_tools() -> list[types.Tool]:
                 "required": ["files", "target"]
             }
         ),
+        types.Tool(
+            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": {
+                    "files": {
+                        "type": "array",
+                        "items": {
+                            "type": "object",
+                            "properties": {
+                                "path": {
+                                    "type": "string",
+                                    "description": "URL of the file to be recognized, supports http/https/oss."
+                                },
+                                "password": {
+                                    "type": "string",
+                                    "description": "Document password, if any."
+                                },
+                                "name": {
+                                    "type": "string",
+                                    "description": "Original filename."
+                                }
+                            },
+                            "required": ["path"]
+                        },
+                        "description": "List of files to be recognized, each item contains path and optional password, name."
+                    },
+                    "format": {
+                        "type": "string",
+                        "description": "Output format, supports pdf/docx/pptx/xlsx/txt, default is pdf.",
+                        "enum": ["pdf", "docx", "pptx", "xlsx", "txt"],
+                        "default": "pdf"
+                    },
+                    "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="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.",
@@ -1271,6 +1332,10 @@ 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"]
+        },
     }
     
     DEFAULTS = {
@@ -1303,6 +1368,12 @@ async def handle_call_tool(name: str, arguments: dict | None) -> list[types.Text
         if isinstance(file_objects, dict):
             file_objects = [file_objects]
             
+        # file_objects中的path需要处理file://协议
+        for file_obj in file_objects:
+            path = file_obj.get("path")
+            if path and path.startswith("file://"):
+                file_obj["path"] = url2pathname(path.removeprefix('file:'))
+        
         config = TOOL_CONFIG[name]
         operation_config = dict(config)  # 复制配置
         

{lightpdf_aipdf_mcp-0.1.135.dist-info → lightpdf_aipdf_mcp-0.1.137.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lightpdf-aipdf-mcp
-Version: 0.1.135
+Version: 0.1.137
 Summary: MCP Server for LightPDF AI-PDF
 Author: LightPDF Team
 License: Proprietary

{lightpdf_aipdf_mcp-0.1.135.dist-info → lightpdf_aipdf_mcp-0.1.137.dist-info}/RECORD

@@ -3,9 +3,10 @@ lightpdf_aipdf_mcp/common.py,sha256=PhTf7Zg6mEgn1rTmJDHotXp-4xb2gWFf-Dy_t25qNdY,
 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=15J_dPZN1lurRUkHG4BTMKUDNrdPoiArOnSqh1Ru8Hk,66670
+lightpdf_aipdf_mcp/ocr.py,sha256=myiKlT6mIb-ns4dAiHuMCEqvW_Cwgnp0UoBF-mC0oN8,3849
+lightpdf_aipdf_mcp/server.py,sha256=vMwBs2pj8w_yNhUBzWqEJeZYfHeEkRobwqLAeGY9K5E,71437
 lightpdf_aipdf_mcp/translator.py,sha256=NbFDz-mZSD4qCNQVyV0W_0x6xXwbqs_7FiBU13JAxZs,4243
-lightpdf_aipdf_mcp-0.1.135.dist-info/METADATA,sha256=PTUZP-BIKA-iCmNiZuju_tr73guyLGjBZprDbcKP47M,8120
-lightpdf_aipdf_mcp-0.1.135.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-lightpdf_aipdf_mcp-0.1.135.dist-info/entry_points.txt,sha256=X7TGUe52N4sYH-tYt0YUGApeJgw-efQlZA6uAZmlmr4,63
-lightpdf_aipdf_mcp-0.1.135.dist-info/RECORD,,
+lightpdf_aipdf_mcp-0.1.137.dist-info/METADATA,sha256=JViKTkcjHF4FY0HGeGOfldroXtaR-2WZG4IEzHG3Juc,8120
+lightpdf_aipdf_mcp-0.1.137.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+lightpdf_aipdf_mcp-0.1.137.dist-info/entry_points.txt,sha256=X7TGUe52N4sYH-tYt0YUGApeJgw-efQlZA6uAZmlmr4,63
+lightpdf_aipdf_mcp-0.1.137.dist-info/RECORD,,