Skip to main content

gpt_4o

What it is​

  • A non-functional template for configuring a gpt-4o ChatOpenAI model intended for the Financial Controller domain.
  • Provides a create_model() factory that fetches OPENAI_API_KEY via naas_abi.secret and returns a configured ChatOpenAI instance (or None if missing).

Public API​

  • create_model() -> ChatOpenAI | None
    • Logs a warning that this is a template.
    • Retrieves OPENAI_API_KEY.
    • Returns a ChatOpenAI configured with:
      • model="gpt-4o"
      • temperature=0.1
      • max_tokens=4000
    • Returns None if the API key is not found.
  • model
    • Module-level placeholder set to None (comment indicates it “would be” create_model()).

Configuration/Dependencies​

  • Dependencies:
    • langchain_openai.ChatOpenAI
    • naas_abi.secret (expects secret key retrieval)
    • naas_abi_core.logger
  • Required secret:
    • OPENAI_API_KEY must be available via secret.get("OPENAI_API_KEY").

Usage​

from naas_abi_marketplace.domains.financial-controller.models import gpt_4o

model = gpt_4o.create_model()
if model is None:
    raise RuntimeError("Missing OPENAI_API_KEY")

# Model is a LangChain ChatOpenAI instance
print(model.model_name)

Caveats​

  • The module is explicitly marked “NOT FUNCTIONAL YET - template only”.
  • The module-level model is always None unless you call create_model() yourself.