# Implementation Plan: Full UX Verification Tests
**Branch**: `009-full-ux-tests` | **Date**: 2026-01-02 | **Spec**: [specs/009-full-ux-tests/spec.md](specs/009-full-ux-tests/spec.md)
**Input**: Feature specification from `/specs/009-full-ux-tests/spec.md`
## Summary
The primary requirement is to implement a comprehensive automated verification suite using Playwright and Chrome DevTools protocols to validate the production-readiness of the Medical GraphRAG Assistant on EC2. The technical approach involves creating a standalone verification package that targets the live Streamlit UI and verifies end-to-end flows including FHIR search, knowledge graph interactions, medical image retrieval, and agent memory persistence.
## Technical Context
**Language/Version**: Python 3.11
**Primary Dependencies**: Playwright, pytest, pytest-playwright, requests, intersystems-iris
**Storage**: InterSystems IRIS (Target environment validation)
**Testing**: pytest (Playwright integration)
**Target Platform**: EC2 (Linux) / Chrome DevTools MCP
**Project Type**: Single (Verification Suite package)
**Performance Goals**: Full suite execution < 15 minutes
**Constraints**: Must run against remote URL provided via `TARGET_URL` env var
**Scale/Scope**: 14+ MCP tools, multi-modal search, memory editor, interactive viz
## Constitution Check
*GATE: Must pass before Phase 0 research. Re-check after Phase 1 design.*
1. **Library-First**: The verification suite will be implemented as a standalone test package. (Pass)
2. **CLI Interface**: The suite will be invokable via a simple `pytest` command with environment variables. (Pass)
3. **Test-First**: The test cases themselves serve as the definitive specification for "implemented features". (Pass)
4. **Integration Testing**: This feature *is* the integration testing framework for the project. (Pass)
5. **Observability**: The suite will produce structured HTML/JSON reports for easy debugging. (Pass)
## Project Structure
### Documentation (this feature)
```text
specs/009-full-ux-tests/
├── plan.md # This file
├── research.md # Phase 0 output
├── data-model.md # Phase 1 output
├── quickstart.md # Phase 1 output
├── contracts/ # Phase 1 output
└── tasks.md # Phase 2 output
```
### Source Code (repository root)
```text
tests/
└── ux/
├── playwright/ # Playwright test scripts
│ ├── conftest.py # Shared fixtures (browser, auth)
│ ├── test_search.py # Multi-modal search tests
│ ├── test_memory.py # Agent memory tests
│ └── test_viz.py # Visualization tests
├── utils/ # Test helpers (IRIS queries, etc.)
└── config/ # Environment-specific test configs
```
**Structure Decision**: Single project structure within the existing `tests/ux/` directory to maintain consistency with the repository's testing pattern while providing a clean separation for Playwright-based tests.
## Complexity Tracking
| Violation | Why Needed | Simpler Alternative Rejected Because |
|-----------|------------|-------------------------------------|
| Playwright Dependency | Browser automation required for Streamlit | Manual verification is non-repeatable |
| Chrome DevTools MCP | Required for deep inspection | Standard HTTP checks miss client-side rendering |
