Add clinical multimodal DB, Flask API, docs and tests

[pifu-dabai]

Motivation

• Provide a lightweight, notebook-friendly multimodal clinical database and API to support automated extraction, semantic retrieval, IQA scoring, longitudinal change detection and basic RAG-style answers for research and data management workflows.

Description

• Add ClinicalMultimodalDB implementing an SQLite-backed schema, FTS5 semantic search, regex-based auto-fill, IQA scoring, lesion/lab delta detection, knowledge-graph extraction and research query helpers in database/clinical_multimodal_db.py and export types in database/__init__.py.
• Add a Flask application factory create_app in server/clinical_api.py exposing endpoints for patient/visit/medication/lesion/lab/asset ingest, auto-fill extraction, IQA scoring, semantic search/RAG, graph extraction, evolution and research queries.
• Include a project-level plan document docs/data_management_research_engine_plan.md describing architecture, feature roadmap and acceptance criteria for intelligent data management and research engine capabilities.
• Add end-to-end unit tests in tests/test_clinical_api.py covering ingest, auto-fill, asset download, search, RAG, IQA, change detection, snapshot, graph functions, audits and research query flows.

Testing

• Ran the added test suite with python -m unittest discover which executed tests/test_clinical_api.py and the end-to-end flow and validation guardrail tests completed successfully.

---

Codex Task