Customization

The DevFlow backend is highly customizable to suit different use cases and requirements.

---

Supported Languages

The backend supports indexing and parsing for multiple programming languages:

Fully Supported

• Python (.py) - Functions, classes, docstrings
• JavaScript (.js) - Functions, classes, JSDoc
• TypeScript (.ts) - Functions, classes, interfaces, types
• Java (.java) - Methods, classes, interfaces
• Go (.go) - Functions, structs, methods
• C++ (.cpp, .cc, .cxx) - Functions, classes
• C# (.cs) - Methods, classes, interfaces
• Rust (.rs) - Functions, structs, traits
• PHP (.php) - Functions, classes
• C (.c) - Functions, structs

Language Mapping

LANGUAGE_MAP = {
    'py': 'python', 'js': 'javascript', 'ts': 'typescript', 
    'java': 'java', 'go': 'go', 'rb': 'ruby',
    'cpp': 'cpp', 'cc': 'cpp', 'cxx': 'cpp', 'cs': 'csharp', 
    'kt': 'kotlin', 'php': 'php', 'c': 'c',
    'rs': 'rust', 'scala': 'scala', 'swift': 'swift'
}

---

Embedding Models

Default Model

• text-embedding-ada-002: OpenAI's recommended embedding model
• Dimensions: 1536
• Performance: Fast and accurate for code embeddings

Alternative Models

You can configure different embedding models via environment variables:

# OpenAI Models
EMBEDDING_MODEL=text-embedding-ada-002
EMBEDDING_MODEL=text-embedding-3-small
EMBEDDING_MODEL=text-embedding-3-large

# Custom Models (if supported)
EMBEDDING_MODEL=custom-model-endpoint

---

Advanced Configuration

Environment Variables

# API Configuration
API_HOST=127.0.0.1
API_PORT=8000
DEBUG=false

# OpenAI Configuration
OPENAI_API_KEY=your_api_key
OPENAI_MODEL=gpt-4.1-nano
OPENAI_TOKEN_LIMIT=2048
OPENAI_TIMEOUT=30

# Embedding Configuration
EMBEDDING_MODEL=text-embedding-ada-002
EMBEDDING_BATCH_SIZE=100
EMBEDDING_TIMEOUT=60

# Database Configuration
CHROMA_DB_PATH=./data/vector_store
METADATA_DB_PATH=./data/metadata.db

# Cache Configuration
CACHE_TTL=3600
CACHE_MAX_SIZE=1000

# Rate Limiting
RATE_LIMIT_REQUESTS=100
RATE_LIMIT_WINDOW=3600

# Logging
LOG_LEVEL=INFO
LOG_FORMAT=json

# Workspace
WORKSPACE_ROOT=/path/to/workspace

Performance Tuning

# Increase batch size for faster indexing
EMBEDDING_BATCH_SIZE=200

# Adjust cache settings
CACHE_TTL=7200
CACHE_MAX_SIZE=2000

# Modify rate limits
RATE_LIMIT_REQUESTS=500
RATE_LIMIT_WINDOW=3600

---

Extending the Backend

Adding New Languages

1. Update Language Map:

   # In app/api/v1/endpoints.py
   LANGUAGE_MAP['new_ext'] = 'new_language'
   

2. Extend Code Parser:

   # In app/services/code_parser.py
   def parse_new_language(self, code: str):
       # Implement language-specific parsing
       pass
   

3. Add Language Support:

   def extract_functions_new_language(self, parsed_code, source_code):
       # Extract functions for the new language
       pass
   

Custom Embedding Models

1. Create Custom Embedder:

   class CustomEmbedder:
       def embed(self, texts: List[str]) -> List[List[float]]:
           # Implement custom embedding logic
           pass
   

2. Update Configuration:

   # In app/services/embedder.py
   if settings.EMBEDDING_MODEL == "custom":
       embedder = CustomEmbedder()
   

Custom Search Logic

1. Extend Vector Store:

   class CustomVectorStore(VectorStore):
       def custom_search(self, query: str, **kwargs):
           # Implement custom search logic
           pass
   

2. Add Custom Endpoints:

   @router.post("/custom_search")
   async def custom_search(request: CustomSearchRequest):
       # Implement custom search endpoint
       pass
   

---

Deployment Customization

Docker Configuration

# Custom Dockerfile
FROM python:3.9-slim

# Install custom dependencies
RUN pip install custom-package

# Copy custom configuration
COPY custom_config.py /app/

# Set custom environment
ENV CUSTOM_SETTING=value

Production Settings

# Production environment variables
DEBUG=false
LOG_LEVEL=WARNING
API_HOST=0.0.0.0
API_PORT=8000

# Security settings
ENABLE_AUTH=true
API_KEY=your_secure_api_key

# Performance settings
WORKERS=4
MAX_CONNECTIONS=1000

---

Monitoring and Logging

Custom Logging

# Custom logging configuration
import logging

logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
    handlers=[
        logging.FileHandler('devflow.log'),
        logging.StreamHandler()
    ]
)

Metrics Collection

# Add custom metrics
from prometheus_client import Counter, Histogram

search_requests = Counter('search_requests_total', 'Total search requests')
search_duration = Histogram('search_duration_seconds', 'Search request duration')

---

Best Practices

Performance

• Use appropriate batch sizes for embeddings
• Configure caching for frequently accessed data
• Monitor memory usage with large codebases
• Use SSD storage for vector database

Security

• Keep API keys secure and rotate regularly
• Use HTTPS in production
• Implement proper authentication
• Validate all input data

Maintenance

• Regularly update dependencies
• Monitor disk space for vector storage
• Backup metadata database
• Clear old embeddings periodically

The backend's modular architecture makes it easy to customize and extend for specific requirements.