iflytek/astron-agent · error · ValueError

Unsupported model source

Error message

Unsupported model source: {model_source}

What it means

ChatAIFactory.get_chat_ai dispatches on the model_source string against ModelProviderEnum values (OPENAI, ANTHROPIC, GOOGLE, ...). If model_source does not match any known provider, it raises a plain ValueError 'Unsupported model source: {model_source}'. This is a configuration/enum-validation guard so unknown providers fail fast instead of producing a None client.

Solutions

1. Check the model_source value configured for the agent and correct it to an exact ModelProviderEnum value (openai/anthropic/google as defined in the enum).
2. Trim and lower-case/normalize the stored model_source before calling the factory.
3. If a new provider is intended, add an elif branch in get_chat_ai mapping it to its ChatAI implementation.
4. Add a startup-time validation that rejects unknown model_source values with a clear message.

Example fix

// before
model_source = config['model_source']  # e.g. 'OpenAI'
llm = ChatAIFactory.get_chat_ai(model_source, **kwargs)
// after
model_source = str(config['model_source']).strip().lower()
assert model_source in [e.value for e in ModelProviderEnum], f'unknown provider {model_source}'
llm = ChatAIFactory.get_chat_ai(model_source, **kwargs)

VALID = {e.value for e in ModelProviderEnum}
def validate_model_source(src: str) -> str:
    s = (src or '').strip().lower()
    if s not in VALID:
        raise ValueError(f'model_source must be one of {sorted(VALID)}, got {src!r}')
    return s

try:
    llm = ChatAIFactory.get_chat_ai(model_source, **kwargs)
except ValueError as e:
    log.error('bad model_source: %s', e)
    llm = ChatAIFactory.get_chat_ai(ModelProviderEnum.OPENAI.value, **kwargs)  # fallback default

When it happens

Trigger: get_chat_ai receives a model_source string that is not one of the ModelProviderEnum values — e.g. a typo like 'anthropicc', an uppercase/untrimmed value like 'Anthropic ', or a newly added provider not yet handled by the factory's if/elif chain.

Common situations: Misconfigured agent model settings in the database/console, hand-written config with wrong casing or whitespace, or a contributor adding a new ModelProviderEnum member without adding a branch in chat_ai_factory.

Related errors

• size limit is invalid
• INVITE_UNSUPPORTED_TYPE
• DATABASE_TYPE_ILLEGAL
• 8309
• Invalid target language

AI-assisted analysis of iflytek/astron-agent@5e758547a8 (2026-09-12). Data as JSON: /api/errors/4ada7164e5c4fd2e. Report an issue: GitHub.

Appendix: source

Thrown at core/workflow/infra/providers/llm/chat_ai_factory.py:53

        Create and return a chat AI instance based on the specified model source.

        :param model_source: The model provider identifier (e.g., 'xinghuo', 'openai')
        :param kwargs: Additional keyword arguments to pass to the chat AI constructor
        :return: An instance of the appropriate chat AI class
        :raises ValueError: If the specified model source is not supported
        """

        # Retrieve the chat AI class from the registry
        if model_source == ModelProviderEnum.XINGHUO.value:
            return SparkChatAi(**kwargs)
        elif model_source == ModelProviderEnum.OPENAI.value:
            return OpenAIChatAI(**kwargs)
        elif model_source == ModelProviderEnum.ANTHROPIC.value:
            return AnthropicChatAI(**kwargs)  # Use new implementation
        elif model_source == ModelProviderEnum.GOOGLE.value:
            return GoogleChatAI(**kwargs)  # Use new implementation
        else:
            raise ValueError(f"Unsupported model source: {model_source}")

View on GitHub (pinned to 5e758547a8)