Documentation index¶

Use this page to answer a question without reading the entire documentation set.

What is the product?¶

CONTEXT.md: canonical language and domain boundaries.
PRODUCT_VISION_STRATEGY.md: product direction and upstream rationale.
Approved service specification: observable requirements and acceptance examples.

Engineering rationale: a current deep dive into concepts, tradeoffs, consistency, resource costs, category/geographic strategies, Swimlanes, ANN, and known limitations.
Implementation workflow: how to trace, implement, verify, and document a change, with concrete examples and evidence requirements.
Repository guide: book-length walkthrough of the product, code paths, algorithms, storage, operations, tests, and recorded versus inferred design rationale.
Architecture: component ownership, dependency direction, state, and external boundaries.
Technical implementation: as-built code, data flow, algorithms, persistence, and lifecycle reference.
Synthetic data generation walkthrough: an educational explanation of profiles, cohorts, random streams, shopping contexts, digests, and verified loading, with Mermaid diagrams and reproducibility exercises.
Service design: approved implementation design.
Serving admission-control design: reusable optional protection boundary and its qualification limits.

Repository command reference: one-page commands for setup, tests, simulation, synthetic sources, benchmarks, service runtime, and documentation deployment.
AGENTS.md: agent entry point, invariants, change loop, and handoff contract.
Testing: command ladder, test layers, CI, generated artifacts, and regression discipline.
Troubleshooting: setup, lock, test, packaging, Docker, and CI failures.
Agent readiness: evidence, gaps, and measurement status.
Contributing: review and contribution expectations.

Recommendations operations: signals, alerts, and response procedures.
Commerce simulation harness: isolated local/test simulation.
Generated relational source: deterministic SQLite or PostgreSQL generation, lineage, local Docker setup, and verification limits, including the step-by-step append and training workflow.
Shopper personalization: authorization, state, privacy, and rollout contract.

Start with the decision map and tradeoffs, then use its source and test links to check the implementation. Recorded decisions and inferred rationale are explicitly distinguished.
Architecture decision records: decision lifecycle and template.
Ownership: accountable documents and tooling.
Research documents under docs/research/ support decisions but do not override approved specifications, source, migrations, or tests.