config.yaml

# Professional RAG System Configuration
app:
  name: "Professional RAG Document Assistant"
  version: "1.0.0"
  debug: false
  max_upload_size: 50 # MB
  max_concurrent_uploads: 3

# Model configurations
models:
  embedding:
    name: "sentence-transformers/all-MiniLM-L6-v2"
    max_seq_length: 256
    batch_size: 32
    device: "auto" # auto, cpu, cuda
  
  reranker:
    name: "cross-encoder/ms-marco-MiniLM-L-6-v2"
    max_seq_length: 512
    batch_size: 16
    enabled: true

# Document processing
processing:
  chunk_size: 512
  chunk_overlap: 50
  min_chunk_size: 100
  max_chunks_per_doc: 1000
  supported_formats: ["pdf", "docx", "txt"]

# Search configuration
search:
  default_k: 10
  max_k: 20
  vector_weight: 0.7
  bm25_weight: 0.3
  rerank_top_k: 50
  final_top_k: 10

# Caching
cache:
  embedding_cache_size: 10000
  query_cache_size: 1000
  cache_ttl: 3600 # seconds
  enable_disk_cache: true
  cache_dir: "./cache"

# UI settings
ui:
  theme: "soft"
  title: "Professional RAG Assistant"
  description: "Upload documents and ask questions with AI-powered retrieval"
  max_file_size: "50MB"
  allowed_extensions: [".pdf", ".docx", ".txt"]
  show_progress: true
  show_analytics: true

# Logging
logging:
  level: "INFO"
  format: "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
  file: "logs/rag_system.log"
  max_size: "10MB"
  backup_count: 5