InitConfig.py 2.8 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788
  1. import logging
  2. import os
  3. from chromadb.utils import embedding_functions
  4. from embedchain.config.BaseConfig import BaseConfig
  5. class InitConfig(BaseConfig):
  6. """
  7. Config to initialize an embedchain `App` instance.
  8. """
  9. def __init__(self, log_level=None, ef=None, db=None, host=None, port=None):
  10. """
  11. :param log_level: Optional. (String) Debug level
  12. ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL'].
  13. :param ef: Optional. Embedding function to use.
  14. :param db: Optional. (Vector) database to use for embeddings.
  15. """
  16. self._setup_logging(log_level)
  17. # Embedding Function
  18. if ef is None:
  19. from chromadb.utils import embedding_functions
  20. self.ef = embedding_functions.OpenAIEmbeddingFunction(
  21. api_key=os.getenv("OPENAI_API_KEY"),
  22. organization_id=os.getenv("OPENAI_ORGANIZATION"),
  23. model_name="text-embedding-ada-002",
  24. )
  25. else:
  26. self.ef = ef
  27. if db is None:
  28. from embedchain.vectordb.chroma_db import ChromaDB
  29. self.db = ChromaDB(ef=self.ef, host=host, port=port)
  30. else:
  31. self.db = db
  32. self.ef = ef
  33. self.db = db
  34. return
  35. def _set_embedding_function(self, ef):
  36. self.ef = ef
  37. return
  38. def _set_embedding_function_to_default(self):
  39. """
  40. Sets embedding function to default (`text-embedding-ada-002`).
  41. :raises ValueError: If the template is not valid as template should contain $context and $query
  42. """
  43. if os.getenv("OPENAI_API_KEY") is None or os.getenv("OPENAI_ORGANIZATION") is None:
  44. raise ValueError("OPENAI_API_KEY or OPENAI_ORGANIZATION environment variables not provided")
  45. self.ef = embedding_functions.OpenAIEmbeddingFunction(
  46. api_key=os.getenv("OPENAI_API_KEY"),
  47. organization_id=os.getenv("OPENAI_ORGANIZATION"),
  48. model_name="text-embedding-ada-002"
  49. )
  50. return
  51. def _set_db(self, db):
  52. if db:
  53. self.db = db
  54. return
  55. def _set_db_to_default(self):
  56. """
  57. Sets database to default (`ChromaDb`).
  58. """
  59. from embedchain.vectordb.chroma_db import ChromaDB
  60. self.db = ChromaDB(ef=self.ef)
  61. def _setup_logging(self, debug_level):
  62. level = logging.WARNING # Default level
  63. if debug_level is not None:
  64. level = getattr(logging, debug_level.upper(), None)
  65. if not isinstance(level, int):
  66. raise ValueError(f"Invalid log level: {debug_level}")
  67. logging.basicConfig(
  68. format="%(asctime)s [%(name)s] [%(levelname)s] %(message)s", level=level
  69. )
  70. self.logger = logging.getLogger(__name__)
  71. return