document_summary

Check which files already have cached AI summaries without triggering generation.

Code
Parameters
----------
params : dict
    {
        "file_ids": ["uuid-1", "uuid-2"]  (required)
    }

Returns
-------
dict
    {
        "summaries": [
            {
                "file_id": "uuid-1",
                "summary": "This document is a...",
                "success": true
            }
        ]
    }

Generate AI-powered summaries for documents attached to a policy.

Code
Retrieves all attachments (or a filtered subset) for the given policy,
extracts their text content, and sends each to an LLM for summarization.

Parameters
----------
params : dict
    {
        "policy_id": "uuid-1"  (required),
        "file_ids": ["uuid-2", "uuid-3"]  (optional, defaults to all documents)
    }

Returns
-------
dict
    {
        "policy_id": "uuid-1",
        "policy_number": "1234",
        "total_documents": 2,
        "summaries": [
            {
                "file_id": "uuid-2",
                "filename": "declaration.pdf",
                "mime_type": "application/pdf",
                "size": 102400,
                "summary": "This document is a...",
                "error": null,
                "success": true
            },
            ...
        ]
    }