ai-agents-structured-output-re-when-using-llamaindex-s-structured-llm-the-respons-9a33c235

THEN Implement a retry mechanism that catches Pydantic validation errors and re-generates the structured output up to N times. This ensures the output conforms to the expected model before proceeding.

from pydantic import ValidationError

for attempt in range(3):
    try:
        response = sllm.chat(messages)
        validated = output_cls.model_validate_json(response.content)
        break
    except (AttributeError, ValidationError):
        continue
else:
    raise Exception("Failed after retries")