README.md•7.62 kB
# Visum Thinker MCP Server
A Model Context Protocol (MCP) server that provides structured sequential thinking capabilities for problem-solving and analysis. This server enables AI assistants to break down complex problems into manageable steps, revise thoughts, and explore alternative reasoning paths.
## Features
- **Step-by-step reasoning**: Break down complex problems into sequential thoughts
- **Dynamic revision**: Revise and refine thoughts as understanding deepens
- **Branching logic**: Branch into alternative reasoning paths
- **Adaptive planning**: Adjust the total number of thoughts dynamically
- **State management**: Maintain thinking context across multiple tool calls
- **Progress tracking**: Monitor completion status and thought progression
- **PDF Analysis**: Load and analyze PDF documents for problem-solving context
- **Content Search**: Find relevant sections in PDFs based on queries and search terms
- **Persistent Storage**: Auto-save state to disk, survive server restarts
- **Knowledge Transfer**: Export/import thinking sessions between servers
- **🚗 Visum Integration**: Complete PTV Visum COM API integration for transportation planning
- **Smart Path Learning**: Automatically discovers and remembers custom Visum installation paths
- **Zero Configuration**: Once set up, works seamlessly across server restarts
- **Demo Mode**: Full functionality testing without requiring Visum installation
- **Transportation Analysis**: Load models, run calculations, analyze networks and matrices
## Installation
### Quick Installation
```bash
# Option 1: Install from NPM
npm install -g visum-thinker-mcp-server
# Option 2: Use with npx (no installation)
npx visum-thinker-mcp-server
# Option 3: Clone from GitHub
git clone https://github.com/yourusername/visum-thinker-mcp-server.git
cd visum-thinker-mcp-server
npm install && npm run build
```
See [INSTALLATION.md](./INSTALLATION.md) for detailed setup instructions.
## Prerequisites
- Node.js 16 or higher
- npm or yarn
## Usage
### With Claude Desktop
Add to your Claude Desktop configuration (`claude_desktop_config.json`):
```json
{
"mcpServers": {
"visum-thinker": {
"command": "node",
"args": ["/absolute/path/to/sequential_thinking/build/index.js"]
}
}
}
```
### With VS Code
The project includes a `.vscode/mcp.json` configuration file for VS Code MCP integration.
### Direct Usage
```bash
npm run dev
```
## Tools
### `sequential_thinking`
Main tool for step-by-step reasoning process.
**Parameters:**
- `thought` (string): The current thinking step
- `nextThoughtNeeded` (boolean): Whether another thought step is needed
- `thoughtNumber` (integer): Current thought number
- `totalThoughts` (integer): Estimated total thoughts needed
- `isRevision` (boolean, optional): Whether this revises previous thinking
- `revisesThought` (integer, optional): Which thought is being reconsidered
- `branchFromThought` (integer, optional): Branching point thought number
- `branchId` (string, optional): Branch identifier
- `needsMoreThoughts` (boolean, optional): If more thoughts are needed
### `load_pdf`
Load a PDF file to provide context for analysis.
**Parameters:**
- `filePath` (string): Absolute path to the PDF file
### `analyze_pdf_section`
Analyze specific sections of the loaded PDF.
**Parameters:**
- `query` (string): What to look for or analyze in the PDF
- `startPage` (integer, optional): Starting page number (1-based)
- `endPage` (integer, optional): Ending page number (1-based)
- `searchTerms` (array of strings, optional): Specific terms to search for
### `reset_thinking`
Clears the current thinking state to start fresh.
### `get_thinking_summary`
Returns a summary of the current thinking session including PDF context if loaded.
### `export_knowledge`
Export the current thinking state and PDF knowledge to a file.
**Parameters:**
- `exportPath` (string): Absolute path where to save the exported knowledge file
### `import_knowledge`
Import thinking state and PDF knowledge from an exported file.
**Parameters:**
- `importPath` (string): Absolute path to the exported knowledge file to import
### 🚗 Visum Transportation Planning Tools
The server includes comprehensive PTV Visum integration with intelligent path learning:
- **`check_visum`**: Check Visum availability and learn custom installation paths
- **`load_visum_model`**: Load transportation models (.ver files)
- **`run_visum_calculation`**: Execute transportation calculations and analyses
- **`get_network_statistics`**: Analyze network topology and characteristics
- **`analyze_visum_matrices`**: Examine demand and flow matrices
- **`export_visum_results`**: Export analysis results to various formats
**Key Features:**
- **🧠 Smart Path Learning**: Automatically remembers custom Visum installation paths
- **🔄 Zero Setup**: Works seamlessly after initial path discovery
- **🎯 Demo Mode**: Full testing capability without Visum installation
- **📊 Complete Analysis**: All major transportation planning workflows supported
See [VISUM-PATH-LEARNING.md](./VISUM-PATH-LEARNING.md) for detailed information about the intelligent path learning system.
## 🤖 GitHub Copilot Integration
The Sequential Thinking MCP Server includes comprehensive GitHub Copilot integration for enhanced AI-assisted development:
### 🚀 Quick Start with Copilot
1. **Server Status**: Ensure MCP server is running (`npm run dev`)
2. **Open Copilot Chat**: `Ctrl+Shift+I` in VS Code
3. **Test Integration**: Ask `@copilot List available MCP tools`
4. **Start Solving**: `@copilot Use sequential thinking to solve [your problem]`
### 🎯 Copilot Capabilities
- **🧠 Sequential Thinking**: AI-guided step-by-step problem solving
- **📄 PDF Analysis**: Intelligent document processing and analysis
- **🚗 Transportation Planning**: Expert Visum integration and workflow automation
- **🔧 Smart Configuration**: Automatic Visum path learning and persistence
- **💡 Context-Aware Suggestions**: Code completion with domain knowledge
### 💬 Example Copilot Interactions
```
@copilot Can you use sequential thinking to analyze this transportation network problem?
@copilot Check if Visum is available and help me load a network model
@copilot Use the PDF analysis tools to extract data from this traffic report
@copilot Create a complete workflow for transportation demand analysis
```
See [COPILOT-INTEGRATION.md](./COPILOT-INTEGRATION.md) for comprehensive setup and usage guide.
## Development
### Project Structure
```
sequential_thinking/
├── src/
│ └── index.ts # Main server implementation
├── build/ # Compiled JavaScript output
├── .vscode/
│ └── mcp.json # VS Code MCP configuration
├── .github/
│ └── copilot-instructions.md
├── package.json
├── tsconfig.json
└── README.md
```
### Scripts
- `npm run build`: Compile TypeScript to JavaScript
- `npm run dev`: Build and run the server
- `npm test`: Run tests (placeholder)
### Debugging
The server logs to stderr for compatibility with STDIO transport. Use VS Code's debugging features or add console.error statements for debugging.
## Architecture
The server maintains a global thinking state that tracks:
- All thoughts in the current session
- Current progress and estimated completion
- Revision and branching relationships
- Session completion status
Each tool call updates this state and provides formatted responses that help users follow the thinking process.
## License
MIT