code/config.yml

embedding_options:
  embedd_files: False # bool
  data_path: 'storage/data' # str
  url_file_path: 'storage/data/urls.txt' # str
  expand_urls: True # bool
  db_option : 'FAISS' # str [FAISS, Chroma, RAGatouille]
  db_path : 'vectorstores' # str
  model : 'sentence-transformers/all-MiniLM-L6-v2' # str [sentence-transformers/all-MiniLM-L6-v2, text-embedding-ada-002']
  search_top_k : 3 # int
  score_threshold : 0.0 # float
  lambda_mult: 0.5 # float - Determines Diversity of the retrieved results
llm_params: 
  use_history: True # bool
  memory_window: 3 # int
  llm_loader: 'local_llm' # str [local_llm, openai]
  openai_params:
    model: 'gpt-4' # str [gpt-3.5-turbo-1106, gpt-4]
  local_llm_params:
    model: "storage/models/tinyllama-1.1b-chat-v0.3.Q5_K_M.gguf"
    model_type: "llama"
    temperature: 0.2
splitter_options:
  use_splitter: True # bool
  split_by_token : True # bool
  remove_leftover_delimiters: True # bool
  remove_chunks: False # bool
  chunk_size : 300 # int
  chunk_overlap : 30 # int
  chunk_separators : ["\n\n", "\n", " ", ""] # list of strings
  front_chunks_to_remove : null # int or None
  last_chunks_to_remove : null # int or None
  delimiters_to_remove : ['\t', '\n', '   ', '  '] # list of strings