feat(intelligence): implement policy matching for demo procedure

[rsalus]

Note for implementers: The design outlined here is a guideline, not a strict specification. Feel free to pivot from the suggested approach if you discover a better solution during implementation. Update the issue with your rationale when deviating significantly.

Summary

Match procedures to payer policies and load criteria for evidence extraction.

Context

• Existing code: apps/intelligence/src/policies/example_policy.py defines the expected policy structure
• Analyze endpoint: has hardcoded SUPPORTED_PROCEDURE_CODES but no policy loading
• Gap: No mechanism to match (procedure_code, payer_id) to a specific policy

Dependencies

Issue	Relationship
#5	Blocked by - need procedure/payer selection
#6, #7	Unblocks - policy criteria feed into LLM prompts

Policy Structure

{
    "policy_id": "string",
    "procedure_codes": ["CPT codes"],
    "criteria": [
        {
            "id": "criterion_id",
            "description": "Human-readable description",
            "evidence_patterns": ["regex patterns"],
            "required": bool
        }
    ],
    "form_field_mappings": {
        "pdf_field_name": "data_field_identifier"
    }
}

Tasks

Phase 1: Policy Module Infrastructure

• Create PolicyRegistry class
• Implement get_policy(procedure_code, payer_id) -> Policy | None
• Add Policy Pydantic model

Phase 2: Integrate with Analyze Endpoint

• Add optional payer_id to AnalyzeRequest
• Replace hardcoded procedure codes with registry lookup
• Pass loaded policy to evidence extraction

Phase 3: Demo Policy (after #5)

• Create concrete policy for selected procedure/payer
• Add unit tests

Files

File	Action
apps/intelligence/src/policies/registry.py	Create
apps/intelligence/src/policies/models.py	Create
apps/intelligence/src/api/analyze.py	Modify

Design References

• §4.3 Policy Definition
• §4.4 LLM Reasoning Chain