`options.base_options`

source

This module contains the BaseOptions class that defines configuration settings for all experiments in lettuce. It handles initialization, validation, and management of various configuration options, particularly focusing on database connections, LLM model settings, embedding configurations, and inference parameters.

The class uses Pydantic BaseSettings to automatically load configuration from environment variables and .env files, providing a robust and type-safe configuration management system.

Classes

`BaseOptions`

class BaseOptions(BaseSettings)

A configuration settings class for the lettuce pipeline that manages all configuration options including database connections, LLM model settings, embedding configurations, and inference parameters. It uses Pydantic BaseSettings to handle environment variable loading and configuration validation with sensible defaults for all configuration options.

Configuration Fields

Database Configuration

Field	Environment Variable	Type	Default	Description
`db_host`	`DB_HOST`	`str`	`"localhost"`	Database host address
`db_user`	`DB_USER`	`str`	`"postgres"`	Database username
`db_password`	`DB_PASSWORD`	`str`	`"password"`	Database password
`db_name`	`DB_NAME`	`str`	`"omop"`	Database name
`db_port`	`DB_PORT`	`int`	`5432`	Database port
`db_schema`	`DB_SCHEMA`	`str`	`"cdm"`	Database schema name
`db_vectable`	`DB_VECTABLE`	`str`	`"embeddings"`	Name of the vector embeddings table
`db_vecsize`	`DB_VECSIZE`	`int`	`384`	Dimension size of embedding vectors

Inference Configuration

Field	Environment Variable	Type	Default	Description
`inference_type`	`INFERENCE_TYPE`	`InferenceType`	`InferenceType.OLLAMA`	Type of inference backend to use
`ollama_url`	`OLLAMA_URL`	`str`	`"http://localhost:11434"`	URL for Ollama server

LLM Model Configuration

Field	Environment Variable	Type	Default	Description
`llm_model`	`LLM_MODEL`	`LLMModel`	`LLMModel.LLAMA_3_1_8B`	LLM model to use for inference
`temperature`	`TEMPERATURE`	`float`	`0.0`	Sampling temperature for LLM generation
`local_llm`	`LOCAL_LLM`	`str \| None`	`None`	Path to local LLM weights file
`debug_prompt`	`DEBUG_PROMPT`	`bool`	`False`	Enable prompt debugging output

Embedding Configuration

Field	Environment Variable	Type	Default	Description
`embedding_model`	`EMBEDDING_MODEL`	`EmbeddingModelName`	`EmbeddingModelName.BGESMALL`	Embedding model to use
`embedding_top_k`	`EMBEDDING_TOP_K`	`int`	`5`	Number of top embeddings to retrieve

Authentication

Field	Environment Variable	Type	Default	Description
`auth_api_key`	`AUTH_API_KEY`	`str \| None`	`None`	API key for authentication

Methods

`connection_url`

def connection_url() -> str:

Generate a PostgreSQL connection URL from the database configuration. Constructs a connection string in the format required by SQLAlchemy and other database libraries using the configured database parameters.

Returns

str

PostgreSQL connection URL in the format: “postgresql://user:password@host:port/database”

Examples

settings = BaseOptions()
settings.connection_url()
'postgresql://postgres:password@localhost:5432/omop'

`print`

def print() -> None:

Print all configuration settings in a formatted display. Outputs all current configuration values in a readable format, useful for debugging and verification of loaded settings.

Examples

settings = BaseOptions()
settings.print()
------------ Options -------------
db_host: localhost
db_user: postgres
db_password: password
...
-------------- End ---------------

Environment Variables

The BaseOptions class automatically loads configuration from environment variables and .env files. Environment variables should be prefixed with the field name in uppercase:

Using .env Files

If you have configuration different from the defaults, create a .env file in your project root:

DB_HOST=my-database-server
DB_USER=myuser
DB_PASSWORD=mypassword
LLM_MODEL=LLAMA_3_1_8B
TEMPERATURE=0.7

The configuration will be automatically loaded:

settings = BaseOptions()

Options Pipeline Options