Empty generated text can return 500s

[evaline-ju]

Describe the bug

When generated_text from an LLM is empty and any detectors are requested on classification-with-text-generation, this can trigger 500

Sample Code

curl -v -H "Content-Type: application/json" --request POST \
--data '{"inputs": "!",
    "model_id": "my-model",
    "guardrail_config": {
        "input": {
            "models": {}
        },
        "output": {
            "models": {}
        }
    }
}' \
    http://localhost:8033/api/v1/task/classification-with-text-generation

Expected behavior

200 with no generated text (and no classification results, since the detectors would be acting on empty text

Observed behavior

500

Additional context

Potentially make sure contents is passed as [""] instead of "" or [] which could trigger 422 on detectors because contents is a required field on /api/v1/text/contents [ref. https://foundation-model-stack.github.io/fms-guardrails-orchestrator/?urls.primaryName=Detector+API#/default/text_content_analysis_unary_handler]

[flaviabeo]

Can you assign me on this one? Debugging this 🔍