jordan/stemedb
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
8f6506b70a
stemedb/uat/consumer-health/README.md
jordan 8f6506b70a feat: Aphoria scan modes + stemedb-ontology crate + consumer health UAT 
Major additions:
- Staged scanning modes (working tree, staged, committed) with git integration
- Drift detection for baseline vs current state comparisons
- Hosted API handlers for policy CRUD operations via StemeDB API
- stemedb-ontology crate with domain definitions and medical extractors
- Consumer health vertical UAT scenarios (GLP-1, gastroparesis, etc.)
- Aphoria development skill documentation

Code organization:
- Split large files into focused modules to stay under 500-line limit
- Extracted config tests, episteme helpers/drift/aliases, API helpers

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-02-04 21:57:33 -07:00

2.8 KiB
Raw Blame History

Consumer Health UAT Scenarios

User Acceptance Testing for the Ontology Layer + Medical Vertical.

Prerequisites

  1. StemeDB running: cargo run --bin stemedb-api
  2. Ontology crate built: cargo build -p stemedb-ontology
  3. Pharma CLI available: cargo build --bin steme-pharma

Scenario Overview

GLP-1 Living Systematic Review Scenarios

Scenario File Tests
Muscle Loss Contradiction glp1-muscle-loss-contradiction.md Skeptic Lens conflict detection
FDA Label Paradigm Shift glp1-fda-label-paradigm-shift.md Epoch supersession O(1)
Pre-print vs Peer Review glp1-preprint-vs-peer-review.md Multi-sig weighting
Semantic Decay glp1-semantic-decay.md 73-day half-life
Visual Anchoring glp1-visual-anchoring.md pHash validation

Consumer Health Intelligence Scenarios

Scenario File Tests
Gastroparesis Multi-Source gastroparesis-multi-source.md Source-class hierarchy
Anecdotal Signal Precedence anecdotal-signal-precedence.md Cluster escalation
Guidance Change Propagation guidance-change-propagation.md "What changed since?"
Layered Consensus layered-consensus.md Per-tier positions
Time Travel Query time-travel-query.md as_of snapshot
Disagreement Dashboard disagreement-dashboard.md Resolved/Active/Emerging

Running Scenarios

Each scenario file contains:

  1. Scenario description - What we're testing
  2. Test matrix - Expected vs actual results
  3. Commands - Exact curl/CLI commands to run
  4. Sign-off checklist - Manual verification points

Example Workflow

# 1. Start StemeDB
cargo run --bin stemedb-api &

# 2. Run a scenario
# Follow commands in glp1-muscle-loss-contradiction.md

# 3. Record results
# Update the test matrix with actual values

# 4. Archive results
cp glp1-muscle-loss-contradiction.md results/2024-XX-XX-muscle-loss.md

Weekly Execution Schedule

Week Scenarios Why
1 (none) Building domain definition
2 (none) Building extractor
3 glp1-muscle-loss-contradiction First conflict demo
4 gastroparesis-multi-source, layered-consensus Source hierarchy
5 glp1-fda-label-paradigm-shift Epochs
6 Full suite Integration validation

Success Criteria

From GLP-1 Living Review:

  • Query muscle_sparing_effect with Skeptic lens returns conflict_score > 0.5
  • Epoch supersession invalidates assertions O(1), not O(N)
  • Multi-sig: Lancet reviewer signature has higher weight

From Consumer Health Intelligence:

  • Tier 0 (FDA) wins over 100x Tier 5 (Reddit) volume
  • lens=layered-consensus returns per-tier positions
  • Source-aware decay: NEJM 8mo old ~0.87 effective; Reddit 26mo old expired