LLM Configuration

LLM configuration loading module.

This module provides functionality to load and validate LLM configuration from YAML files, including environment variable expansion.

class spacylize.llm_config.LLMConfig(*, model, api_key=None, api_base=None, max_tokens=1024)

Bases: BaseModel

Configuration model for LLM settings.

Defines the structure and validation rules for LLM configuration including model selection, authentication, and generation parameters.

Parameters:

• model (str)

• api_key (str | None)

• api_base (str | None)

• max_tokens (int)

model

The model identifier for LiteLLM.

Type:

str

api_key

Optional API key for authentication.

Type:

str | None

api_base

Optional custom API base URL.

Type:

str | None

max_tokens

Maximum number of tokens to generate.

Type:

int

model: str

api_key: str | None

api_base: str | None

max_tokens: int

model_config = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

spacylize.llm_config.expand_env_vars(value)

Recursively expand environment variable references in configuration values.

Replaces ${VAR_NAME} patterns with their environment variable values. Supports nested dictionaries and lists.

Parameters:

value (Any) – Configuration value to process (str, dict, list, or other).

Returns:

The value with environment variables expanded.

Return type:

Any

spacylize.llm_config.load_llm_config(path)

Load and validate LLM configuration from a YAML file.

Parameters:

path (Path) – Path to the YAML configuration file.

Returns:

Validated LLM configuration object.

Return type:

LLMConfig

Raises:

RuntimeError – If the configuration is invalid or fails validation.