dwf-analyzer-mcp
An MCP server for analyzing DWF files using Amazon Bedrock's Nova Pro model, enabling visual analysis, metadata extraction, and image extraction from DWF drawings.
README
DWF Analyzer MCP Server
A Model Context Protocol (MCP) server for analyzing DWF (Design Web Format) files using Amazon Bedrock's Nova Pro model for visual analysis.
Features
- DWF File Parsing: Extract metadata, layers, and object information from DWF files
- Image Extraction: Extract embedded images from DWF files with format detection and conversion
- Visual Analysis: Analyze drawings using Amazon Nova Pro multimodal AI model
- Comprehensive Analysis: Combined metadata extraction, image processing, and visual analysis
Installation
Prerequisites
- Python 3.12 or higher
- AWS account with Bedrock access
- AWS credentials configured
Install from Source
git clone https://github.com/jesamkim/dwf-analyzer-mcp.git
cd dwf-analyzer-mcp
uv install
Using uvx (Recommended)
uvx dwf-analyzer-mcp
Configuration
Environment Variables
Set the following environment variables:
export AWS_ACCESS_KEY_ID="your-access-key"
export AWS_SECRET_ACCESS_KEY="your-secret-key"
export AWS_DEFAULT_REGION="us-west-2" # Optional, defaults to us-west-2
export LOG_LEVEL="INFO" # Optional, defaults to INFO
AWS Permissions
Your AWS credentials need the following permissions:
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"bedrock:InvokeModel",
"bedrock:InvokeModelWithResponseStream"
],
"Resource": [
"arn:aws:bedrock:*::foundation-model/us.anthropic.claude-3-7-sonnet-20250219-v1:0",
"arn:aws:bedrock:*::foundation-model/us.amazon.nova-pro-v1:0"
]
}
]
}
Usage
Running the Server
# Using the installed package (HTTP transport)
dwf-analyzer-mcp
# Using uvx (HTTP transport)
uvx dwf-analyzer-mcp
# Using FastMCP directly (HTTP transport)
fastmcp run dwf_analyzer_mcp.server:mcp --transport streamable-http
# Custom host and port
HOST=0.0.0.0 PORT=8080 dwf-analyzer-mcp
MCP Client Configuration
Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"dwf-analyzer": {
"url": "http://localhost:8080/mcp",
"transport": "http",
"env": {
"AWS_ACCESS_KEY_ID": "your-access-key",
"AWS_SECRET_ACCESS_KEY": "your-secret-key",
"AWS_DEFAULT_REGION": "us-west-2"
}
}
}
}
Cursor IDE
Add an HTTP-type server:
{
"url": "http://localhost:8080/mcp",
"transport": "http"
}
Zed Editor
Add to your MCP settings:
{
"dwf-analyzer": {
"url": "http://localhost:8080/mcp",
"transport": "http"
}
}
Available Tools
Core Analysis Tools
extract_dwf_metadata
Extract metadata and basic information from a DWF file.
Parameters:
file_path(string): Path to the DWF file
Returns: Dictionary containing file metadata, layers, and object summary
extract_dwf_images
Extract embedded images from a DWF file.
Parameters:
file_path(string): Path to the DWF fileoutput_dir(string, optional): Directory to save extracted images
Returns: Dictionary containing extraction results and image information
analyze_dwf_visual
Perform visual analysis of a DWF file using Amazon Nova Pro.
Parameters:
file_path(string): Path to the DWF filefocus_area(string): Analysis focus area (general)include_metadata(boolean): Whether to include file metadata in results
Returns: Dictionary containing visual analysis results
analyze_dwf_comprehensive
Perform comprehensive analysis including metadata, images, and visual analysis.
Parameters:
file_path(string): Path to the DWF file
Returns: Dictionary containing complete analysis results
System Tools
health_check
Perform a health check of the service.
Returns: Dictionary containing service health status
Available Resources
dwf://analysis/config
Get the current analysis configuration schema.
dwf://analysis/supported-formats
Get information about supported DWF formats and capabilities.
Supported Formats
- DWF Versions: V00.22, V00.30, V00.55, V06.00
- Image Formats: PNG, JPEG, BMP, GIF, WEBP
- Max File Size: 100MB
- AWS Regions: us-west-2, us-east-1, eu-west-1, ap-southeast-1
Analysis Focus Areas
- general: Overall drawing analysis
Development
Setup Development Environment
git clone https://github.com/jesamkim/dwf-analyzer-mcp.git
cd dwf-analyzer-mcp
uv install --dev
Running Tests
uv run pytest
Code Formatting
uv run black src tests
uv run ruff check src tests
Type Checking
uv run mypy src
Examples
Basic Usage
# Extract metadata from a DWF file
result = extract_dwf_metadata("/path/to/drawing.dwf")
print(result["metadata"]["header"]["version"])
# Extract images
images = extract_dwf_images("/path/to/drawing.dwf", "/output/directory")
print(f"Extracted {images['extracted_images']} images")
# Perform visual analysis
analysis = analyze_dwf_visual("/path/to/drawing.dwf", focus_area="general")
print(analysis["visual_analysis"])
Comprehensive Analysis
# Perform complete analysis
result = analyze_dwf_comprehensive("/path/to/drawing.dwf")
# Access different parts of the analysis
metadata = result["metadata"]
images = result["image_extraction"]
visual = result["visual_analysis"]
Troubleshooting
Common Issues
-
AWS Credentials Error
- Ensure AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY are set
- Verify your AWS account has Bedrock access
-
Model Access Error
- Check if Nova Pro model is available in your AWS region
- Verify your AWS permissions include bedrock:InvokeModel
-
Image Processing Error
- Ensure Pillow is installed:
pip install pillow - Check if the DWF file contains extractable images
- Ensure Pillow is installed:
-
File Not Found Error
- Verify the DWF file path is correct and accessible
- Check file permissions
Debug Mode
Enable debug logging:
export LOG_LEVEL=DEBUG
dwf-analyzer-mcp
License
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
Changelog
v0.2.0
- BREAKING CHANGE: Migrated from STDIO to Streamable HTTP transport
- Updated server configuration for HTTP deployment
- Added uvicorn and starlette dependencies for HTTP server support
- Updated client configuration examples for HTTP transport
- Improved scalability and production deployment capabilities
v0.1.0
- Initial release
- Basic DWF parsing functionality
- Image extraction capabilities
- Amazon Nova Pro integration
- FastMCP server implementation
- Comprehensive analysis tools
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。