🐍 Python Implementation Prompt

📋 Instructions

This page contains the comprehensive prompt for implementing the MongoDB Airbnb Workshop REST API in Python. Use this prompt with Cline to generate a complete Python application.

🎯 How to Use This Prompt

  1. Copy the entire prompt from the section below
  2. Paste it into your AI assistant of choice
  3. Make sure the swagger.json file is used as context when prompted
  4. Review and test the generated code
  5. Iterate and refine as needed

📂 Expected Deliverables

After running this prompt, you should receive:

  • ✅ Complete Python REST API application
  • ✅ Organized file structure with separated concerns
  • ✅ MongoDB indexes configuration file
  • ✅ Environment setup instructions
  • ✅ Integration with Atlas Search and Vector Search
  • ✅ AI Chatbot implementation with RAG

🤖 Python Implementation Prompt

Copy the prompt below and use it with your AI assistant:

Create a Python REST API application based on the swagger.json specification file.

**Core Requirements:**
- Read and analyze the swagger.json file to understand all API endpoints
- Build a Python web application using Flask or FastAPI on port 5000
- **Implement ONLY the APIs defined in the swagger.json file - no additional endpoints**
- Organize code into multiple files based on the API categories you identify from the swagger
- Separate concerns: CRUD operations, search functionality, vector search, and chatbot

**Key Implementation Areas:**

1. **CRUD Operations**: Standard database operations for listings and reviews
2. **Search & Filtering**: Basic filtering using MongoDB aggregation pipelines and standard queries
3. **Statistics**: Statistical analysis using MongoDB aggregation framework
4. **Atlas Search**: Implement using MongoDB Atlas Search operators (autocomplete, facet, text search)
5. **Vector Search**: Implement using MongoDB Atlas Vector Search with automated embeddings
   - **Embedding Field**: Configure automated embeddings on the `description` field for semantic search capabilities
6. **AI Chatbot**: RAG implementation using Atlas Vector Search + AWS Bedrock LLM integration via LangChain
7. **Results**: Workshop participant management

**Implementation Scope:**
- **Swagger-Only Implementation**: Create endpoints ONLY as specified in the swagger.json file
- **No Extra Features**: Do not add endpoints, routes, or functionality beyond what's documented in the OpenAPI specification

**MongoDB Integration:**
- Use the available MCP (Model Context Protocol) to thoroughly analyze and understand the collection schemas before implementing any queries
- **Schema Discovery First**: Query the MCP to examine the actual data structure, field types, and document format in the collection
- Implement proper database queries based on the actual collection structure and field types discovered through MCP
- Ensure all field references and data types match the real database schema, not assumptions

**Database Collection Usage:**
- **Primary Collection**: All listing-related endpoints (`/api/listingsAndReviews/*`) use the `listingsAndReviews` collection
- **Atlas Search**: Search endpoints (`/autocomplete`, `/facet`, `/search`) operate on the `listingsAndReviews` collection
- **Vector Search**: The `/vectorsearch` endpoint uses the `listingsAndReviews` collection with automated embeddings
- **Chat System**: Chat endpoints (`/api/chat/*`) use the `listingsAndReviews` collection for RAG operations
- **Results Data**: Only the Results endpoints (`/api/results/*`) use data from the `airbnb_arena` database

**Atlas Search & Vector Search Requirements:**
- **Vector Search Index**: Create a separate vector search index specifically for the `/vectorsearch` endpoint with automated embeddings on the `description` field
- **Lexical Search Index**: Create a single Atlas Search index that supports all lexical search operations (`/search`, `/facet`, `/autocomplete`) using appropriate operators and analyzers

**Atlas Search Field Mappings:**
- **name** (for autocomplete): Enable autocomplete search on listing names
- **amenities** (for filtering): Support faceted filtering by available amenities  
- **property_type** (for filtering): Allow filtering by property type categories
- **beds** (for numeric filtering): Enable range filtering on number of beds

**Vector Search Configuration:**
- **numCandidates**: Set to 100 for optimal performance and accuracy
- **limit**: Return top 10 most relevant results
- **Embedding Field**: Configure automated embeddings on the `description` field
- **RAG Architecture**: Use vector search for retrieval in chatbot responses

**Required Deliverables:**
1. **Complete Python application** with proper file organization
2. **mongodb_indexes file** containing all required MongoDB indexes:
   - One comprehensive Atlas Search index for all lexical search operations (search, facet, autocomplete) with proper field mappings
   - One dedicated Vector Search index for semantic search with automated embeddings on the `description` field
   - Standard database indexes for performance optimization
   - Include clear comments and creation instructions
   - MongoDB Atlas UI or Atlas CLI is not available
3. **requirements.txt file** without version pinning - let pip resolve the latest compatible versions

**Environment Variables:**
Load configuration from the existing `.env` file which contains:
- MONGODB_URI, AWS_REGION, LLM_MODEL, and other necessary configuration
- Use python-dotenv to load these environment variables into your application

**Additional Features:**
- Proper error handling and validation
- Environment configuration support
- CORS for frontend integration
- Compatible with existing test cases in rest-lab/ folder
- LangChain integration for AWS Bedrock connectivity
- LangChain MongoDB chat memory

**Technical Implementation Notes:**
- Separate filtering/statistics from advanced Atlas Search operations
- Use appropriate MongoDB operators for each search type
- Design RAG pipeline for contextual chatbot responses
- Ensure the single Atlas Search index supports text operators, autocomplete operators, and facet operators with proper field mappings
- Create a dedicated vector search index separate from the lexical search index
- No authorization/credentials is required

Analyze the swagger.json file first, then design and implement the complete application with appropriate file structure and all required indexes.

🔍 Automated Embeddings Implementation Note

Automated embeddings is a newer MongoDB Atlas feature. If your AI assistant doesn’t implement automated embeddings correctly or lacks current documentation knowledge, use this enhanced prompt:

@web Reference the MongoDB Atlas Vector Search automated embedding documentation at https://www.mongodb.com/docs/atlas/atlas-vector-search/automated-embedding/ for implementation details. Use curl to retrieve the information, not chromium.

🔧 Post-Generation Checklist

After receiving your generated code, verify:

  • File Structure: Code is organized by functionality (CRUD, search, vector, chat, etc.)
  • Environment Setup: .env file template and requirements.txt included
  • MongoDB Indexes: file with all necessary indexes
  • Atlas Search: Proper implementation of lexical search features
  • Vector Search: Automated embeddings integration for semantic search
  • AI Chatbot: RAG implementation with LangChain and Bedrock
  • Error Handling: Proper HTTP status codes and error responses
  • Testing: Compatible with existing test files in rest-lab folder
  • Documentation: Clear setup and usage instructions

💡 Tips for Success

  1. Start with swagger.json: Always provide the OpenAPI specification as context
  2. Test incrementally: Test each endpoint category as it’s implemented
  3. Use the MCP: Leverage the Model Context Protocol for schema understanding
  4. Follow the patterns: The generated code should match the workshop’s architectural patterns
  5. Iterate: Don’t hesitate to ask for refinements or specific improvements

🚀 Ready to Code?

Copy the prompt above and start building your MongoDB Airbnb Workshop API in Python!