htmlgen-mcp 0.2.2__py3-none-any.whl → 0.3.1__py3-none-any.whl

{agents → htmlgen_mcp/agents}/web_tools/navigation.py

@@ -333,6 +333,35 @@ def create_responsive_navbar(file_path: str, brand_name: str = "公司名称", n
                         # 没找到 body，兜底：前置插入
                         content = navbar_html + "\n" + content
 
+            # 确保 Bootstrap 依赖存在（导航栏需要）
+            bootstrap_css = '<link href="https://cdn.jsdelivr.net/npm/bootstrap@5.3.0/dist/css/bootstrap.min.css" rel="stylesheet">'
+            bootstrap_js = '<script src="https://cdn.jsdelivr.net/npm/bootstrap@5.3.0/dist/js/bootstrap.bundle.min.js"></script>'
+            fontawesome_css = '<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.4.0/css/all.min.css">'
+            
+            # 添加 Bootstrap CSS（如果不存在）
+            if 'bootstrap' not in content.lower():
+                if '</head>' in content:
+                    # 在 </head> 前插入，但要在自定义 CSS 之前
+                    if '<link rel="stylesheet" href="assets/css/style.css">' in content:
+                        content = content.replace(
+                            '<link rel="stylesheet" href="assets/css/style.css">',
+                            f'{bootstrap_css}\n    {fontawesome_css}\n    <link rel="stylesheet" href="assets/css/style.css">'
+                        )
+                    else:
+                        content = content.replace('</head>', f'    {bootstrap_css}\n    {fontawesome_css}\n</head>')
+            
+            # 添加 Bootstrap JS（如果不存在）
+            if 'bootstrap.bundle' not in content.lower():
+                if '</body>' in content:
+                    # 在 </body> 前插入
+                    if '<script src="assets/js/main.js"></script>' in content:
+                        content = content.replace(
+                            '<script src="assets/js/main.js"></script>',
+                            f'{bootstrap_js}\n    <script src="assets/js/main.js"></script>'
+                        )
+                    else:
+                        content = content.replace('</body>', f'    {bootstrap_js}\n</body>')
+            
             # 确保导航锚点对应的区块存在
             id_pattern = re.compile(r'id\s*=\s*["\']([^"\']+)["\']', re.I)
             existing_ids = {m.group(1).strip().lower() for m in id_pattern.finditer(content)}

htmlgen_mcp/config.py

@@ -0,0 +1,326 @@
+"""项目配置管理 - 跨平台支持"""
+import os
+import sys
+import platform
+import tempfile
+from pathlib import Path
+from datetime import datetime
+
+
+class ProjectConfig:
+    """项目配置管理器 - 支持 Windows/macOS/Linux"""
+    
+    @staticmethod
+    def get_system_info() -> dict:
+        """获取系统信息"""
+        return {
+            'system': platform.system(),  # 'Windows', 'Darwin' (macOS), 'Linux'
+            'platform': sys.platform,      # 'win32', 'darwin', 'linux'
+            'home': Path.home(),
+            'temp': Path(tempfile.gettempdir())
+        }
+    
+    @staticmethod
+    def get_default_output_dir() -> Path:
+        """获取默认的项目输出目录（跨平台）
+        
+        优先级：
+        1. 环境变量 WEB_AGENT_OUTPUT_DIR
+        2. 系统特定的文档目录
+        3. 用户主目录下的隐藏目录
+        4. 系统临时目录
+        """
+        
+        # 1. 检查环境变量（所有平台通用）
+        env_dir = os.environ.get('WEB_AGENT_OUTPUT_DIR')
+        if env_dir:
+            output_dir = Path(env_dir)
+            try:
+                output_dir.mkdir(parents=True, exist_ok=True)
+                return output_dir
+            except Exception as e:
+                print(f"⚠️ 无法创建环境变量指定的目录: {e}")
+        
+        system_info = ProjectConfig.get_system_info()
+        home = system_info['home']
+        system = system_info['system']
+        
+        # 2. 系统特定的文档目录
+        if system == 'Windows':
+            # Windows: 使用 Documents 文件夹
+            docs_candidates = [
+                home / 'Documents' / 'WebProjects',
+                home / 'My Documents' / 'WebProjects',  # 旧版 Windows
+                Path(os.environ.get('USERPROFILE', home)) / 'Documents' / 'WebProjects'
+            ]
+        elif system == 'Darwin':  # macOS
+            # macOS: Documents 文件夹
+            docs_candidates = [
+                home / 'Documents' / 'WebProjects',
+                home / 'Projects' / 'WebProjects'  # 有些用户喜欢用 Projects 文件夹
+            ]
+        else:  # Linux 及其他 Unix-like 系统
+            # Linux: 遵循 XDG 标准
+            xdg_documents = os.environ.get('XDG_DOCUMENTS_DIR')
+            docs_candidates = []
+            if xdg_documents:
+                docs_candidates.append(Path(xdg_documents) / 'WebProjects')
+            docs_candidates.extend([
+                home / 'Documents' / 'WebProjects',
+                home / 'projects' / 'web',  # Linux 用户常用小写
+                home / 'Projects' / 'Web'
+            ])
+        
+        # 尝试创建文档目录
+        for doc_dir in docs_candidates:
+            try:
+                if doc_dir.parent.exists():
+                    doc_dir.mkdir(parents=True, exist_ok=True)
+                    return doc_dir
+            except Exception:
+                continue
+        
+        # 3. 用户主目录下的隐藏目录（所有平台）
+        if system == 'Windows':
+            # Windows 使用 AppData
+            app_data = os.environ.get('APPDATA')
+            if app_data:
+                hidden_dirs = [
+                    Path(app_data) / 'WebAgent' / 'projects',
+                    home / 'AppData' / 'Roaming' / 'WebAgent' / 'projects'
+                ]
+            else:
+                hidden_dirs = [home / '.web-agent' / 'projects']
+        else:
+            # macOS 和 Linux 使用点开头的隐藏目录
+            hidden_dirs = [
+                home / '.web-agent' / 'projects',
+                home / '.local' / 'share' / 'web-agent' / 'projects'  # XDG 标准
+            ]
+        
+        for hidden_dir in hidden_dirs:
+            try:
+                hidden_dir.mkdir(parents=True, exist_ok=True)
+                return hidden_dir
+            except Exception:
+                continue
+        
+        # 4. 系统临时目录（最后的选择）
+        temp_base = system_info['temp']
+        temp_dir = temp_base / 'web-agent-projects'
+        try:
+            temp_dir.mkdir(parents=True, exist_ok=True)
+            print(f"⚠️ 使用临时目录: {temp_dir}")
+            print(f"💡 建议设置环境变量 WEB_AGENT_OUTPUT_DIR 到更合适的位置")
+            return temp_dir
+        except Exception as e:
+            # 如果连临时目录都无法创建，使用当前工作目录
+            print(f"⚠️ 无法创建临时目录: {e}")
+            fallback = Path.cwd() / 'web-projects'
+            fallback.mkdir(parents=True, exist_ok=True)
+            return fallback
+    
+    @staticmethod
+    def create_project_directory(
+        project_name: str,
+        base_dir: Path = None,
+        use_timestamp: bool = True
+    ) -> Path:
+        """创建项目目录
+        
+        Args:
+            project_name: 项目名称
+            base_dir: 基础目录，如果不提供则使用默认目录
+            use_timestamp: 是否在目录名中添加时间戳（避免冲突）
+            
+        Returns:
+            创建的项目目录路径
+        """
+        if base_dir is None:
+            base_dir = ProjectConfig.get_default_output_dir()
+        
+        # 清理项目名称，移除特殊字符
+        safe_name = "".join(c for c in project_name if c.isalnum() or c in (' ', '-', '_'))
+        safe_name = safe_name.strip().replace(' ', '-').lower()
+        
+        if use_timestamp:
+            # 添加时间戳避免冲突
+            timestamp = datetime.now().strftime("%Y%m%d-%H%M%S")
+            dir_name = f"{safe_name}-{timestamp}"
+        else:
+            dir_name = safe_name
+        
+        project_dir = base_dir / dir_name
+        
+        # 如果目录已存在且不使用时间戳，添加序号
+        if project_dir.exists() and not use_timestamp:
+            counter = 1
+            while (base_dir / f"{dir_name}-{counter}").exists():
+                counter += 1
+            project_dir = base_dir / f"{dir_name}-{counter}"
+        
+        project_dir.mkdir(parents=True, exist_ok=True)
+        
+        # 创建标准子目录结构
+        (project_dir / "assets" / "css").mkdir(parents=True, exist_ok=True)
+        (project_dir / "assets" / "js").mkdir(parents=True, exist_ok=True)
+        (project_dir / "assets" / "images").mkdir(parents=True, exist_ok=True)
+        
+        # 创建项目信息文件
+        info_file = project_dir / ".project-info.json"
+        import json
+        project_info = {
+            "name": project_name,
+            "created_at": datetime.now().isoformat(),
+            "generator": "htmlgen-mcp",
+            "version": "0.3.0"
+        }
+        info_file.write_text(json.dumps(project_info, ensure_ascii=False, indent=2))
+        
+        return project_dir
+    
+    @staticmethod
+    def get_user_projects_list(base_dir: Path = None) -> list:
+        """获取用户已创建的项目列表
+        
+        Returns:
+            项目信息列表
+        """
+        if base_dir is None:
+            base_dir = ProjectConfig.get_default_output_dir()
+        
+        projects = []
+        if not base_dir.exists():
+            return projects
+        
+        for item in base_dir.iterdir():
+            if item.is_dir():
+                info_file = item / ".project-info.json"
+                if info_file.exists():
+                    try:
+                        import json
+                        info = json.loads(info_file.read_text())
+                        info['path'] = str(item)
+                        projects.append(info)
+                    except:
+                        # 如果没有info文件，仍然添加基本信息
+                        projects.append({
+                            'name': item.name,
+                            'path': str(item),
+                            'created_at': datetime.fromtimestamp(item.stat().st_mtime).isoformat()
+                        })
+        
+        # 按创建时间倒序排序
+        projects.sort(key=lambda x: x.get('created_at', ''), reverse=True)
+        return projects
+    
+    @staticmethod
+    def clean_old_projects(
+        base_dir: Path = None,
+        days_to_keep: int = 7,
+        max_projects: int = 20
+    ) -> int:
+        """清理旧项目
+        
+        Args:
+            base_dir: 基础目录
+            days_to_keep: 保留最近几天的项目
+            max_projects: 最多保留多少个项目
+            
+        Returns:
+            删除的项目数量
+        """
+        if base_dir is None:
+            base_dir = ProjectConfig.get_default_output_dir()
+        
+        projects = ProjectConfig.get_user_projects_list(base_dir)
+        deleted = 0
+        
+        # 如果项目数超过限制，删除最旧的
+        if len(projects) > max_projects:
+            for project in projects[max_projects:]:
+                try:
+                    import shutil
+                    shutil.rmtree(project['path'])
+                    deleted += 1
+                except:
+                    pass
+        
+        # 删除超过指定天数的项目
+        from datetime import timedelta
+        cutoff_date = datetime.now() - timedelta(days=days_to_keep)
+        
+        for project in projects:
+            try:
+                created_at = datetime.fromisoformat(project.get('created_at', ''))
+                if created_at < cutoff_date:
+                    import shutil
+                    shutil.rmtree(project['path'])
+                    deleted += 1
+            except:
+                pass
+        
+        return deleted
+
+
+# 便捷函数
+def get_project_directory(project_name: str = None) -> str:
+    """获取项目目录（供 MCP 工具使用）
+    
+    Args:
+        project_name: 项目名称，如果不提供则生成默认名称
+        
+    Returns:
+        项目目录路径字符串
+    """
+    if not project_name:
+        project_name = f"web-project-{datetime.now().strftime('%Y%m%d')}"
+    
+    config = ProjectConfig()
+    project_dir = config.create_project_directory(project_name, use_timestamp=True)
+    
+    print(f"📁 项目将生成在: {project_dir}")
+    print(f"💡 提示: 可通过设置环境变量 WEB_AGENT_OUTPUT_DIR 来自定义输出目录")
+    
+    return str(project_dir)
+
+
+def list_recent_projects(limit: int = 10) -> list:
+    """列出最近的项目
+    
+    Args:
+        limit: 返回的项目数量
+        
+    Returns:
+        项目列表
+    """
+    config = ProjectConfig()
+    projects = config.get_user_projects_list()
+    return projects[:limit]
+
+
+def clean_temp_projects() -> int:
+    """清理临时项目
+    
+    Returns:
+        删除的项目数量
+    """
+    config = ProjectConfig()
+    
+    # 如果使用的是 /tmp 目录，更积极地清理
+    output_dir = config.get_default_output_dir()
+    if str(output_dir).startswith('/tmp'):
+        # 临时目录只保留1天，最多10个项目
+        return config.clean_old_projects(days_to_keep=1, max_projects=10)
+    else:
+        # 其他目录保留7天，最多20个项目
+        return config.clean_old_projects(days_to_keep=7, max_projects=20)
+
+
+# 导出的功能
+__all__ = [
+    'ProjectConfig',
+    'get_project_directory',
+    'list_recent_projects', 
+    'clean_temp_projects'
+]

htmlgen_mcp/sse_optimizations.py

@@ -0,0 +1,195 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""SSE 模式优化补丁 - 解决超时问题"""
+
+import asyncio
+import json
+import os
+from typing import Dict, Any, List, Optional
+from pathlib import Path
+
+async def read_progress_file_async(
+    file_path: str, 
+    limit: int = 20,
+    chunk_size: int = 8192
+) -> tuple[List[Dict[str, Any]], int]:
+    """异步读取进度日志文件，避免阻塞"""
+    events: List[Dict[str, Any]] = []
+    total_lines = 0
+    
+    try:
+        # 使用异步方式读取文件
+        loop = asyncio.get_event_loop()
+        
+        # 异步读取文件内容
+        def read_file():
+            with open(file_path, 'r', encoding='utf-8') as f:
+                lines = f.readlines()
+                return lines
+        
+        lines = await loop.run_in_executor(None, read_file)
+        total_lines = len(lines)
+        
+        # 处理最后的 limit 行
+        for line in lines[-limit:]:
+            line = line.strip()
+            if not line:
+                continue
+            try:
+                events.append(json.loads(line))
+            except Exception:
+                continue
+                
+        return events, total_lines
+        
+    except Exception as e:
+        raise Exception(f"读取进度文件失败: {str(e)}")
+
+
+async def check_file_exists_async(file_path: str) -> bool:
+    """异步检查文件是否存在"""
+    loop = asyncio.get_event_loop()
+    return await loop.run_in_executor(None, os.path.exists, file_path)
+
+
+def create_streaming_response(data: Dict[str, Any], chunk_size: int = 100) -> List[Dict[str, Any]]:
+    """
+    将大响应分块，适用于 SSE 流式传输
+    避免一次性发送大量数据导致超时
+    """
+    if 'events' in data and len(data['events']) > chunk_size:
+        # 分块发送事件
+        chunks = []
+        events = data['events']
+        
+        for i in range(0, len(events), chunk_size):
+            chunk_data = data.copy()
+            chunk_data['events'] = events[i:i + chunk_size]
+            chunk_data['chunk_index'] = i // chunk_size
+            chunk_data['total_chunks'] = (len(events) + chunk_size - 1) // chunk_size
+            chunks.append(chunk_data)
+        
+        return chunks
+    
+    return [data]
+
+
+async def get_progress_optimized(
+    plan_id: Optional[str] = None,
+    job_id: Optional[str] = None,
+    log_path: Optional[str] = None,
+    limit: int = 20,
+    _job_registry: Dict = None,
+    _progress_log_by_job: Dict = None,
+    _progress_log_by_id: Dict = None,
+    project_root: str = None
+) -> Dict[str, Any]:
+    """
+    SSE 优化版的 get_progress 函数
+    - 使用异步 I/O 避免阻塞
+    - 添加超时控制
+    - 支持流式响应
+    """
+    try:
+        # 设置操作超时（SSE 环境下更短的超时）
+        timeout = 10  # 10秒超时
+        
+        async def _get_progress():
+            if limit <= 0:
+                limit = 20
+            
+            job_info = None
+            resolved_path = None
+            
+            # 查找日志文件路径
+            if job_id and _job_registry:
+                job_info = _job_registry.get(job_id)
+                if job_info and not plan_id:
+                    plan_id = job_info.get("plan_id")
+                if _progress_log_by_job and job_id in _progress_log_by_job:
+                    resolved_path = _progress_log_by_job[job_id]
+                elif job_info and job_info.get("progress_log"):
+                    resolved_path = job_info.get("progress_log")
+            
+            if not resolved_path and plan_id and _progress_log_by_id and plan_id in _progress_log_by_id:
+                resolved_path = _progress_log_by_id[plan_id]
+            
+            if log_path:
+                resolved_path = log_path
+            
+            # 解析路径
+            if resolved_path:
+                if not os.path.isabs(resolved_path):
+                    candidate = os.path.join(project_root or os.getcwd(), resolved_path)
+                    if await check_file_exists_async(candidate):
+                        resolved_path = candidate
+                    else:
+                        alt = os.path.sep + resolved_path.lstrip(os.path.sep)
+                        if await check_file_exists_async(alt):
+                            resolved_path = alt
+            
+            # 检查文件是否存在
+            if not resolved_path or not await check_file_exists_async(resolved_path):
+                return {
+                    "status": "error",
+                    "message": "未找到进度日志，请确认 job_id/plan_id 或提供 log_path"
+                }
+            
+            # 异步读取文件
+            events, total_lines = await read_progress_file_async(resolved_path, limit)
+            
+            response = {
+                "status": "success",
+                "plan_id": plan_id,
+                "job_id": job_id,
+                "log_path": resolved_path,
+                "events": events,
+                "total_records": total_lines,
+                "returned": len(events),
+            }
+            
+            # 添加任务信息
+            if job_info:
+                snapshot_keys = [
+                    "job_id", "status", "plan_id", "progress_log",
+                    "started_at", "updated_at", "completed_at",
+                    "project_directory", "model", "upload_status",
+                    "website_url", "upload_completed_at"
+                ]
+                job_snapshot = {
+                    k: job_info.get(k) for k in snapshot_keys 
+                    if job_info.get(k) is not None
+                }
+                
+                if job_info.get("status") == "completed":
+                    job_snapshot["result_summary"] = {
+                        "report": job_info.get("result", {}).get("report"),
+                        "created_files": job_info.get("result", {}).get("created_files"),
+                    }
+                    if job_info.get("upload_result"):
+                        job_snapshot["upload_result"] = job_info.get("upload_result")
+                
+                if job_info.get("status") == "failed":
+                    job_snapshot["error"] = job_info.get("error")
+                
+                if job_info.get("upload_error"):
+                    job_snapshot["upload_error"] = job_info.get("upload_error")
+                
+                response["job"] = job_snapshot
+            
+            return response
+        
+        # 使用超时控制
+        result = await asyncio.wait_for(_get_progress(), timeout=timeout)
+        return result
+        
+    except asyncio.TimeoutError:
+        return {
+            "status": "error",
+            "message": f"操作超时（{timeout}秒）- SSE 模式下请减少 limit 参数值"
+        }
+    except Exception as exc:
+        return {
+            "status": "error",
+            "message": str(exc)
+        }

{MCP → htmlgen_mcp}/web_agent_server.py

@@ -20,7 +20,8 @@ from typing import Any, Dict, Optional
 
 # 确保项目根目录在模块搜索路径中
 CURRENT_DIR = os.path.dirname(os.path.abspath(__file__))
-PROJECT_ROOT = os.path.dirname(CURRENT_DIR)
+# 当前文件在 src/htmlgen_mcp/ 下，所以需要向上两级到项目根目录
+PROJECT_ROOT = os.path.dirname(os.path.dirname(os.path.dirname(CURRENT_DIR)))
 if PROJECT_ROOT not in sys.path:
     sys.path.insert(0, PROJECT_ROOT)
 
@@ -29,7 +30,7 @@ from fastmcp import FastMCP  # type: ignore
 import uuid
 from pathlib import Path
 
-from agents.smart_web_agent import SmartWebAgent
+from htmlgen_mcp.agents.smart_web_agent import SmartWebAgent
 
 DEFAULT_PROJECT_ROOT = os.path.abspath(
     os.environ.get("WEB_AGENT_PROJECT_ROOT", os.path.join(PROJECT_ROOT, "projects"))
@@ -911,6 +912,9 @@ async def get_progress(
     """
 
     try:
+        # SSE 模式优化：使用异步 I/O
+        loop = asyncio.get_event_loop()
+        
         if limit <= 0:
             limit = 20
 
@@ -935,32 +939,48 @@ async def get_progress(
         if resolved_path:
             if not os.path.isabs(resolved_path):
                 candidate = os.path.join(PROJECT_ROOT, resolved_path)
-                if os.path.exists(candidate):
+                # 异步检查文件存在
+                exists = await loop.run_in_executor(None, os.path.exists, candidate)
+                if exists:
                     resolved_path = candidate
                 else:
                     alt = os.path.sep + resolved_path.lstrip(os.path.sep)
-                    if os.path.exists(alt):
+                    exists_alt = await loop.run_in_executor(None, os.path.exists, alt)
+                    if exists_alt:
                         resolved_path = alt
 
-        if not resolved_path or not os.path.exists(resolved_path):
+        # 异步检查最终路径
+        path_exists = await loop.run_in_executor(
+            None, lambda: resolved_path and os.path.exists(resolved_path)
+        )
+        
+        if not path_exists:
             return {
                 "status": "error",
                 "message": "未找到进度日志，请确认 job_id/plan_id 或提供 log_path（注意绝对路径需以/开头，扩展名为 .jsonl）",
             }
 
-        events: list[Dict[str, Any]] = []
-        total_lines = 0
-        with open(resolved_path, "r", encoding="utf-8") as f:
-            lines = f.readlines()
-            total_lines = len(lines)
-            for line in lines[-limit:]:
-                line = line.strip()
-                if not line:
-                    continue
-                try:
-                    events.append(json.loads(line))
-                except Exception:
-                    continue
+        # 异步读取文件
+        def read_file():
+            events = []
+            total = 0
+            try:
+                with open(resolved_path, "r", encoding="utf-8") as f:
+                    lines = f.readlines()
+                    total = len(lines)
+                    for line in lines[-limit:]:
+                        line = line.strip()
+                        if not line:
+                            continue
+                        try:
+                            events.append(json.loads(line))
+                        except Exception:
+                            continue
+            except Exception:
+                pass
+            return events, total
+        
+        events, total_lines = await loop.run_in_executor(None, read_file)
 
         response: Dict[str, Any] = {
             "status": "success",
@@ -1183,7 +1203,7 @@ async def deploy_folder_or_zip(
     """
     try:
         # 导入EdgeOne部署工具
-        from agents.web_tools.edgeone_deploy import deploy_folder_or_zip_to_edgeone
+        from htmlgen_mcp.agents.web_tools.edgeone_deploy import deploy_folder_or_zip_to_edgeone
 
         # 验证环境变量
         api_token = os.getenv("EDGEONE_PAGES_API_TOKEN")

{htmlgen_mcp-0.2.2.dist-info → htmlgen_mcp-0.3.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: htmlgen-mcp
-Version: 0.2.2
+Version: 0.3.1
 Summary: AI-powered HTML website generator with auto-upload functionality via Model Context Protocol
 Author-email: HTML Generator Team <contact@htmlgen-mcp.com>
 License: MIT