cnks 0.1.0__py3-none-any.whl → 0.2.1__py3-none-any.whl

cnks/server.py

@@ -7,8 +7,11 @@ import subprocess
 import sys
 import time
 import logging
+import webbrowser
+import traceback
 from pathlib import Path
 from urllib.parse import quote
+from typing import Dict, List, Any, Optional, Union
 
 from mcp.server.models import InitializationOptions
 import mcp.types as types
@@ -41,6 +44,16 @@ browser_instance = None
 
 server = Server("cnks")
 
+# 导入我们新创建的extractor模块
+try:
+    from . import chrome_extractor as extractor
+except ImportError:
+    try:
+        import chrome_extractor as extractor
+    except ImportError:
+        extractor = None
+        logger.warning("无法导入chrome_extractor模块，批量提取功能将不可用")
+
 def find_chrome_executable():
     """查找Chrome可执行文件路径"""
     system = platform.system()
@@ -97,6 +110,8 @@ def open_chrome(url):
 
 async def search_with_playwright(keywords):
     """使用playwright在知网搜索关键词"""
+    global page_content
+    
     if not PLAYWRIGHT_AVAILABLE:
         return "需要安装playwright模块：uv add playwright"
     
@@ -216,7 +231,7 @@ async def search_with_playwright(keywords):
                                             # 查找所有包含"article/abstract?v="字样的链接
                                             links_count = await find_and_count_abstract_links(page)
                                             
-                                            return f"已完成全部操作：搜索关键词、设置每页显示50条、勾选CSSCI来源类别。找到{links_count}条包含article/abstract?v=的链接。浏览器将保持打开状态。"
+                                            return links_count
                                         else:
                                             logger.debug("[DEBUG] 在来源类别区域未找到CSSCI选项")
                                             
@@ -231,9 +246,11 @@ async def search_with_playwright(keywords):
                                                 # 查找所有包含"article/abstract?v="字样的链接
                                                 links_count = await find_and_count_abstract_links(page)
                                                 
-                                                return f"已完成全部操作：搜索关键词、设置每页显示50条、勾选CSSCI来源类别。找到{links_count}条包含article/abstract?v=的链接。浏览器将保持打开状态。"
+                                                return links_count
                                             else:
-                                                return "已完成搜索和设置每页显示50条，但未找到CSSCI选项。浏览器将保持打开状态。"
+                                                # 查找所有包含"article/abstract?v="字样的链接
+                                                links_count = await find_and_count_abstract_links(page)
+                                                return links_count
                                     else:
                                         logger.debug("[DEBUG] 未找到来源类别区域")
                                         
@@ -248,61 +265,127 @@ async def search_with_playwright(keywords):
                                             # 查找所有包含"article/abstract?v="字样的链接
                                             links_count = await find_and_count_abstract_links(page)
                                             
-                                            return f"已完成全部操作：搜索关键词、设置每页显示50条、勾选CSSCI来源类别。找到{links_count}条包含article/abstract?v=的链接。浏览器将保持打开状态。"
+                                            return links_count
                                         else:
-                                            return "已完成搜索和设置每页显示50条，但未找到来源类别区域或CSSCI选项。浏览器将保持打开状态。"
+                                            # 查找所有包含"article/abstract?v="字样的链接
+                                            links_count = await find_and_count_abstract_links(page)
+                                            return links_count
                                 except Exception as e:
                                     logger.debug(f"[DEBUG] 勾选CSSCI选项时出错: {str(e)}")
-                                    return f"已完成搜索和设置每页显示50条，但勾选CSSCI时出错: {str(e)}。浏览器将保持打开状态。"
+                                    # 查找所有包含"article/abstract?v="字样的链接
+                                    links_count = await find_and_count_abstract_links(page)
+                                    return links_count
                                 
-                                return "已完成全部操作：搜索关键词、点击排序下拉框、选择每页显示50条。浏览器将保持打开状态。"
+                                # 查找所有包含"article/abstract?v="字样的链接
+                                links_count = await find_and_count_abstract_links(page)
+                                return links_count
                             else:
                                 logger.debug("[DEBUG] 未找到'50'选项")
-                                return "已搜索并点击下拉框，但未找到'50'选项。浏览器将保持打开状态。"
+                                page_content = {
+                                    "count": 0,
+                                    "links": [],
+                                    "error": "已搜索并点击下拉框，但未找到'50'选项"
+                                }
+                                return 0
                         else:
                             logger.debug("[DEBUG] 未找到排序下拉框")
-                            return "已搜索，但未找到排序下拉框。浏览器将保持打开状态。"
+                            page_content = {
+                                "count": 0,
+                                "links": [],
+                                "error": "已搜索，但未找到排序下拉框"
+                            }
+                            return 0
                     except Exception as e:
                         logger.debug(f"[DEBUG] 点击下拉框或选项时出错: {str(e)}")
-                        return f"已搜索，但在点击下拉框或选项时出错: {str(e)}。浏览器将保持打开状态。"
-                    
-                    # 不关闭浏览器，让它保持打开状态
-                    # 注意：不调用 browser.close() 和 playwright.stop()
+                        page_content = {
+                            "count": 0,
+                            "links": [],
+                            "error": f"已搜索，但在点击下拉框或选项时出错: {str(e)}"
+                        }
+                        return 0
                 else:
                     # 不关闭浏览器
-                    return f"已填写搜索关键词: {keywords}，但未找到搜索按钮。请手动点击搜索。"
+                    page_content = {
+                        "count": 0,
+                        "links": [],
+                        "error": f"已填写搜索关键词: {keywords}，但未找到搜索按钮"
+                    }
+                    return 0
             else:
                 # 不关闭浏览器
-                return f"未找到搜索框。已打开知网页面，请手动搜索: {keywords}"
+                page_content = {
+                    "count": 0,
+                    "links": [],
+                    "error": f"未找到搜索框，无法搜索: {keywords}"
+                }
+                return 0
         except Exception as e:
             logger.debug(f"[DEBUG] 填写搜索框或点击搜索按钮时出错: {str(e)}")
             # 不关闭浏览器
-            return f"自动搜索过程中出错，请手动在页面中搜索: {keywords}"
+            page_content = {
+                "count": 0,
+                "links": [],
+                "error": f"自动搜索过程中出错: {str(e)}"
+            }
+            return 0
     except Exception as e:
         error_msg = str(e)
         logger.debug(f"[DEBUG] Playwright错误: {error_msg}")
         
         # 如果是找不到Chrome的错误，提供更明确的指导
         if "Executable doesn't exist" in error_msg and "ms-playwright" in error_msg:
-            return f"需要安装Playwright的浏览器: playwright install\n如果您想使用系统Chrome，请重新启动服务器。\n\n{error_msg}"
-        
-        # 如果Playwright启动失败，使用传统方式打开Chrome
-        return f"使用Playwright启动Chrome失败: {error_msg}。尝试使用传统方式打开浏览器。"
+            error_message = f"需要安装Playwright的浏览器: playwright install\n如果您想使用系统Chrome，请重新启动服务器。\n\n{error_msg}"
+        else:
+            error_message = f"使用Playwright启动Chrome失败: {error_msg}"
+            
+        page_content = {
+            "count": 0,
+            "links": [],
+            "error": error_message
+        }
+        return 0
 
 def search_with_direct_chrome(keywords):
     """直接使用Chrome搜索，不使用playwright"""
-    logger.debug("[DEBUG] 正在使用search_with_direct_chrome函数")
-    
-    # 构建知网搜索URL - 知网不支持URL参数搜索，所以只能打开页面
-    url = "https://kns.cnki.net/kns8s/search"
+    global page_content
     
-    # 打开Chrome
-    result = open_chrome(url)
+    logger.debug("[DEBUG] 正在使用search_with_direct_chrome函数")
     
-    if result is True:
-        return f"已打开知网页面。请在搜索框中输入并搜索: {keywords}"
-    else:
-        return f"打开Chrome浏览器失败: {result}"
+    try:
+        url = f"https://kns.cnki.net/kns8s/search?q={quote(keywords)}"
+        logger.debug(f"[DEBUG] 打开URL: {url}")
+        
+        result = open_chrome(url)
+        
+        if isinstance(result, str) and "打开Chrome" in result:
+            logger.debug(f"[DEBUG] 直接打开Chrome结果: {result}")
+            
+            page_content = {
+                "count": 0,
+                "links": [],
+                "error": f"直接打开Chrome搜索: {result}"
+            }
+            
+        else:
+            logger.debug("[DEBUG] 直接打开Chrome成功")
+            
+            page_content = {
+                "count": 0,
+                "links": [],
+                "message": "已打开Chrome并搜索关键词，但无法自动获取链接。请安装playwright以获取完整功能。"
+            }
+        
+        return page_content
+    except Exception as e:
+        logger.debug(f"[DEBUG] search_with_direct_chrome出错: {str(e)}")
+        
+        page_content = {
+            "count": 0,
+            "links": [],
+            "error": f"使用Chrome搜索时出错: {str(e)}"
+        }
+        
+        return page_content
 
 def get_page_content():
     """获取当前页面内容（简化模拟）"""
@@ -514,49 +597,25 @@ async def handle_get_prompt(
 @server.list_tools()
 async def handle_list_tools() -> list[types.Tool]:
     """列出可用工具"""
-    return [
-        types.Tool(
-            name="open-cnki",
-            description="打开中国知网搜索页面",
-            inputSchema={
-                "type": "object",
-                "properties": {},
-                "required": [],
-            },
-        ),
-        types.Tool(
-            name="search-keywords",
-            description="在知网搜索关键词",
-            inputSchema={
-                "type": "object",
-                "properties": {
-                    "keywords": {"type": "string", "description": "搜索关键词"},
-                },
-                "required": ["keywords"],
-            },
-        ),
-        types.Tool(
-            name="add-note",
-            description="添加笔记",
-            inputSchema={
-                "type": "object",
-                "properties": {
-                    "name": {"type": "string", "description": "笔记名称"},
-                    "content": {"type": "string", "description": "笔记内容"},
+    tools = []
+    
+    # 只添加搜索并提取的组合工具
+    if extractor is not None and PLAYWRIGHT_AVAILABLE:
+        tools.append(
+            types.Tool(
+                name="mcp_cnks_search_and_extract",
+                description="搜索知网关键词并提取所有论文的详细内容",
+                inputSchema={
+                    "type": "object",
+                    "properties": {
+                        "keywords": {"type": "string", "description": "搜索关键词"},
+                    },
+                    "required": ["keywords"],
                 },
-                "required": ["name", "content"],
-            },
-        ),
-        types.Tool(
-            name="get-abstract-links",
-            description="获取最近一次搜索找到的论文摘要链接",
-            inputSchema={
-                "type": "object",
-                "properties": {},
-                "required": [],
-            },
+            )
         )
-    ]
+    
+    return tools
 
 @server.call_tool()
 async def handle_call_tool(
@@ -565,25 +624,7 @@ async def handle_call_tool(
     """处理工具执行请求"""
     global current_url, page_content
     
-    if name == "open-cnki":
-        current_url = "https://kns.cnki.net/kns8s/search"
-        result = open_chrome(current_url)
-        if result is True:
-            return [
-                types.TextContent(
-                    type="text",
-                    text="已打开中国知网搜索页面。"
-                )
-            ]
-        else:
-            return [
-                types.TextContent(
-                    type="text",
-                    text=f"打开中国知网时出错: {result}"
-                )
-            ]
-    
-    elif name == "search-keywords":
+    if name == "mcp_cnks_search_and_extract" and extractor is not None and PLAYWRIGHT_AVAILABLE:
         if not arguments:
             raise ValueError("缺少参数")
         
@@ -591,72 +632,80 @@ async def handle_call_tool(
         if not keywords:
             raise ValueError("缺少关键词")
         
-        # 优先使用playwright进行搜索
-        if PLAYWRIGHT_AVAILABLE:
-            result = await search_with_playwright(keywords)
+        try:
+            # 第一步：执行搜索
+            logger.info(f"开始执行搜索并提取：关键词 '{keywords}'")
+            links_count = await search_with_playwright(keywords)
             current_url = "https://kns.cnki.net/kns8s/search"
             
-            return [
-                types.TextContent(
-                    type="text",
-                    text=result
-                )
-            ]
-        else:
-            # 如果没有playwright，回退到传统方式
-            result = search_with_direct_chrome(keywords)
-            current_url = "https://kns.cnki.net/kns8s/search"
+            # 检查搜索结果
+            if not isinstance(page_content, dict) or "links" not in page_content or not page_content["links"]:
+                return [
+                    types.TextContent(
+                        type="text",
+                        text=json.dumps({
+                            "error": "搜索未返回有效链接",
+                            "count": 0,
+                            "results": []
+                        }, ensure_ascii=False)
+                    )
+                ]
+            
+            # 提取链接
+            urls = [link["url"] for link in page_content["links"] if "url" in link]
+            if not urls:
+                return [
+                    types.TextContent(
+                        type="text",
+                        text=json.dumps({
+                            "error": "未找到有效链接",
+                            "count": 0,
+                            "results": []
+                        }, ensure_ascii=False)
+                    )
+                ]
+            
+            # 第二步：执行提取
+            logger.info(f"搜索成功，找到 {len(urls)} 个链接，开始提取内容")
+            results = await extractor.batch_extract_contents(urls)
+            
+            # 包装结果
+            result_json = {
+                "keywords": keywords,
+                "count": len(results),
+                "results": results,
+                "success_count": sum(1 for r in results if "error" not in r or not r["error"]),
+                "error_count": sum(1 for r in results if "error" in r and r["error"])
+            }
             
             return [
                 types.TextContent(
                     type="text",
-                    text=f"{result}。如需自动搜索功能，请安装: uv add playwright"
+                    text=json.dumps(result_json, ensure_ascii=False)
                 )
             ]
-    
-    elif name == "add-note":
-        if not arguments:
-            raise ValueError("缺少参数")
-        
-        note_name = arguments.get("name")
-        content = arguments.get("content")
-        
-        if not note_name or not content:
-            raise ValueError("缺少名称或内容")
-        
-        # 更新服务器状态
-        notes[note_name] = content
-        
-        # 通知客户端资源已更改
-        await server.request_context.session.send_resource_list_changed()
-        
-        return [
-            types.TextContent(
-                type="text",
-                text=f"已添加笔记 '{note_name}': {content}"
-            )
-        ]
-    
-    elif name == "get-abstract-links":
-        if not page_content or "找到" not in page_content:
+        except Exception as e:
+            logger.error(f"搜索并提取时出错: {str(e)}")
+            logger.error(traceback.format_exc())
             return [
                 types.TextContent(
                     type="text",
-                    text="尚未执行搜索或未找到链接。请先使用search-keywords工具搜索。"
+                    text=json.dumps({
+                        "error": f"搜索并提取内容时出错: {str(e)}",
+                        "keywords": keywords,
+                        "count": 0,
+                        "results": []
+                    }, ensure_ascii=False)
                 )
             ]
-        
-        return [
-            types.TextContent(
-                type="text",
-                text=page_content
-            )
-        ]
     
-    raise ValueError(f"未知工具: {name}")
+    else:
+        raise ValueError(f"未知工具: {name}")
 
 async def find_and_count_abstract_links(page):
     """查找并统计包含article/abstract?v=的链接"""
+    global page_content
+    
     try:
         logger.debug("[DEBUG] 开始查找所有包含article/abstract?v=的链接")
         
@@ -690,11 +739,11 @@ async def find_and_count_abstract_links(page):
         else:
             logger.debug(f"[DEBUG] 链接数量为{links_count}条，多于预期的50条")
         
-        # 存储结果 - 只包含编号和链接，不包含标题和连字符
-        global page_content
-        page_content = f"找到{links_count}条包含article/abstract?v=的链接\n\n" + "\n".join([
-            f"{link['index']}. {link['href']}" for link in links_info
-        ])
+        # 存储结果 - 使用字典结构而不是纯文本
+        page_content = {
+            "count": links_count,
+            "links": [{"index": link['index'], "url": link['href']} for link in links_info]
+        }
         
         return links_count
     except Exception as e:
@@ -710,7 +759,7 @@ async def main():
             write_stream,
             InitializationOptions(
                 server_name="cnks",
-                server_version="0.1.0",
+                server_version="0.2.1",
                 capabilities=server.get_capabilities(
                     notification_options=NotificationOptions(),
                     experimental_capabilities={},
@@ -725,33 +774,54 @@ def create_fastmcp_server():
         from mcp.server.fastmcp import FastMCP
         fast_mcp = FastMCP("知网搜索")
         
-        @fast_mcp.tool()
-        def open_cnki_search():
-            """打开中国知网搜索页面"""
-            return open_chrome("https://kns.cnki.net/kns8s/search")
-        
-        @fast_mcp.tool()
-        async def search_keywords(keywords: str) -> str:
-            """在知网搜索关键词"""
-            logger.debug("[DEBUG] 正在使用FastMCP的search_keywords函数")
-            if PLAYWRIGHT_AVAILABLE:
-                result = await search_with_playwright(keywords)
-                return result
-            else:
-                result = search_with_direct_chrome(keywords)
-                return f"{result}。如需自动搜索功能，请安装: uv add playwright"
-        
-        @fast_mcp.tool()
-        def get_abstract_links() -> str:
-            """获取最近一次搜索找到的论文摘要链接"""
-            if not page_content or "找到" not in page_content:
-                return "尚未执行搜索或未找到链接。请先使用search_keywords工具搜索。"
-            return page_content
-        
-        @fast_mcp.resource("webpage://current")
-        def get_current_webpage() -> str:
-            """获取当前网页内容"""
-            return get_page_content()
+        # 只添加搜索并提取的工具
+        if extractor is not None and PLAYWRIGHT_AVAILABLE:
+            @fast_mcp.tool()
+            async def mcp_cnks_search_and_extract(keywords: str) -> dict:
+                """搜索关键词并提取所有论文的详细内容"""
+                logger.debug("[DEBUG] 正在使用FastMCP的mcp_cnks_search_and_extract函数")
+                try:
+                    # 第一步：执行搜索
+                    result_count = await search_with_playwright(keywords)
+                    
+                    # 检查搜索结果
+                    if not isinstance(page_content, dict) or "links" not in page_content or not page_content["links"]:
+                        return {
+                            "error": "搜索未返回有效链接",
+                            "keywords": keywords,
+                            "count": 0,
+                            "results": []
+                        }
+                    
+                    # 提取链接
+                    urls = [link["url"] for link in page_content["links"] if "url" in link]
+                    if not urls:
+                        return {
+                            "error": "未找到有效链接",
+                            "keywords": keywords,
+                            "count": 0,
+                            "results": []
+                        }
+                    
+                    # 第二步：执行提取
+                    results = await extractor.batch_extract_contents(urls)
+                    
+                    # 包装结果
+                    return {
+                        "keywords": keywords,
+                        "count": len(results),
+                        "results": results,
+                        "success_count": sum(1 for r in results if "error" not in r or not r["error"]),
+                        "error_count": sum(1 for r in results if "error" in r and r["error"])
+                    }
+                except Exception as e:
+                    logger.error(f"搜索并提取时出错: {str(e)}")
+                    return {
+                        "error": f"搜索并提取内容时出错: {str(e)}",
+                        "keywords": keywords,
+                        "count": 0,
+                        "results": []
+                    }
         
         return fast_mcp
     except ImportError: