Generated AI helper

This commit is contained in:
Bill Thiede 2025-09-11 16:13:02 -07:00
parent 9452a2b014
commit 2c1c7abf0a

40
.github/copilot-instructions.md vendored Normal file
View File

@ -0,0 +1,40 @@
# Copilot/AI Agent Instructions for Letterbox
## Project Overview
- **Letterbox** is a Rust monorepo for a mail/newsreader system with a web frontend and a Rocket/GraphQL backend.
- Major crates: `server` (backend, Rocket+async-graphql), `web` (Seed-based WASM frontend), `notmuch` (mail integration), `shared` (common types), `procmail2notmuch` (migration/utility).
- Data flows: Email/news data is indexed and queried via the backend, exposed to the frontend via GraphQL. SQLx/Postgres is used for persistence. Notmuch and custom SQL are both used for mail storage/search.
## Key Workflows
- **Development**: Use `dev.sh` to launch a tmux session with live-reloading for both frontend (`trunk serve`) and backend (`cargo watch ... run`).
- **Build/Release**: Use `just patch|minor|major` for versioned releases (runs SQLx prepare, bumps versions, pushes). `Makefile`'s `release` target does similar steps.
- **Frontend**: In `web/`, use `cargo make serve` and `cargo make watch` for local dev. See `web/README.md` for Seed-specific details.
- **Backend**: In `server/`, run with `cargo run` or via the tmux/dev.sh workflow. SQL migrations are in `server/migrations/`.
## Project Conventions & Patterns
- **GraphQL**: All API boundaries are defined in `server/src/graphql.rs`. Use the `Query`, `Mutation`, and `Subscription` roots. Types are defined with `async-graphql` derive macros.
- **HTML Sanitization**: See `server/src/lib.rs` for custom HTML/CSS sanitization and transformation logic (e.g., `Transformer` trait, `sanitize_html`).
- **Tag/Query Parsing**: The `Query` struct in `server/src/lib.rs` parses user queries into filters for notmuch/newsreader/tantivy.
- **Shared Types**: Use the `shared` crate for types and helpers shared between frontend and backend.
- **Custom SQL**: Raw SQL queries are in `server/sql/`. Use these for complex queries not handled by SQLx macros.
- **Feature Flags**: The `tantivy` feature enables full-text search via Tantivy. Check for `#[cfg(feature = "tantivy")]` in backend code.
## Integration Points
- **Notmuch**: Integrated via the `notmuch` crate for mail indexing/search.
- **Postgres**: Used for newsreader and other persistent data (see `server/migrations/`).
- **GraphQL**: All client-server communication is via GraphQL endpoints defined in the backend.
- **Seed/Trunk**: Frontend is built with Seed (Rust/WASM) and served via Trunk.
## Examples
- To add a new GraphQL query, update `server/src/graphql.rs` and expose it in the `QueryRoot`.
- To add a new frontend page, add a module in `web/src/` and register it in the Seed app's router.
- To run the full dev environment: `./dev.sh` (requires tmux, trunk, cargo-watch, etc.).
## References
- See `web/README.md` for frontend/Seed workflow details.
- See `Justfile` and `Makefile` for release/versioning automation.
- See `server/src/lib.rs` and `server/src/graphql.rs` for backend architecture and conventions.
- See `server/sql/` for custom SQL queries.
---
If any conventions or workflows are unclear, please ask for clarification or check the referenced files for examples.