Content
# 🌊 HRFCO Intelligent Water Search System
AI-friendly Korean water data search system with natural language processing.
## 🚀 Features
- **Natural Language Search**: "한강 수위", "서울 강우량" → Automatic station discovery
- **Smart Matching**: 16 regions + river name mapping + similarity scoring
- **Response Optimization**: All responses < 1KB (prevents LLM token overflow)
- **OpenAI Compatible**: Ready for ChatGPT Function Calling integration
## 📡 API Endpoints
- `/.netlify/functions/search-station` - Search stations by location name
- `/.netlify/functions/get-water-info` - One-stop water information query
- `/.netlify/functions/recommend-stations` - Recommend nearby stations
- `/.netlify/functions/openai-functions` - OpenAI Function definitions
## 🔧 Environment Variables
```
HRFCO_API_KEY=your-api-key-here
```
## 🧪 Test Examples
```bash
# Search stations
curl -X POST https://your-site.netlify.app/.netlify/functions/search-station \
-d '{"location_name": "한강", "limit": 3}'
# Get water info
curl -X POST https://your-site.netlify.app/.netlify/functions/get-water-info \
-d '{"query": "서울 수위", "limit": 5}'
```
## 📊 Performance
- **Data Source**: 1,366 water level + 742 rainfall observatories
- **Response Size**: 346-522 bytes (optimized for LLM)
- **Search Accuracy**: 90%+ for Korean location names
- **Response Time**: < 3 seconds
Built with TypeScript + Netlify Functions
Connection Info
You Might Also Like
firecrawl
Firecrawl MCP Server enables web scraping, crawling, and content extraction.
markitdown
MarkItDown-MCP is a lightweight server for converting URIs to Markdown.
markitdown
Python tool for converting files and office documents to Markdown.
Filesystem
Node.js MCP Server for filesystem operations with dynamic access control.
Sequential Thinking
A structured MCP server for dynamic problem-solving and reflective thinking.
Fetch
Retrieve and process content from web pages by converting HTML into markdown format.