This project aims to be a replacement for a collection of tools I use regularly for managing my ebook library, among which are calibre and calibre-web. While both of those tools are amazing, the fact that I need at least those two to get everything I want is enough reason for me to try and improve my situation with something new.
As this project is still in its early phases, I'm focusing on creating a minimum viable product that can replace calibre and calibre-web, as well as improve on some key missing features. This includes
- Adding/Editing/Deleting ebooks in my library
- Managing metadata associated with each book
- Having books that may belong to multiple series/universes
- Having proper views of authors and series
- Using automatic fetching of metadata with sensible standards and ways of replacing data
- Basic data collection regarding books I've read over time (when started/finished, etc.)
- Ideally provide an app for as many platforms as possible (native, mobile, web)
Currently, I'm using the following app components and versions to build everything:
- Tauri CLI: 4.6.2
- pnpm: 10.18.0
Tauri only allows creating native desktop apps and mobile apps. There is currently no way to use Tauri for web apps directly, so the future of that part is unclear. Ideally, I would like to create something that runs on the web as well because this would be a great way of getting access to my library even from devices that don't have Promethea installed, but at the same time this would likely require almost an entirely separate code base with its own HTTP server implementation. This would likely happen with a backend written with axum and an interface that lets me adapt the frontend to both Tauri and the axum implementation.
For now, Tauri is being developed mainly for the desktop, but once it's in a more stable and polished form, I'll look into bringing it to mobile as well.
To enforce cleaner commit messages, I've decided to try commit linting with husky and commitlint. Whenever a commit message is created via git commit -m "some-message", these two tools enforce the structure "type(scope): message" where type can be things such as "feat", "fix", "revert" etc. The "scope" is intended to show where the type of change has been made roughly. Finally, the message is just any type of message.
Since both husky and commitlint were added to the project via pnpm, the packages should be automatically installed on any system that uses this code base. As such, no additional setup effort should be required.
Managed to run the debugger using nvim-dap with overseer:
- Use overseer to run the task defined in
.vscode/tasks.json::OverseerRun, then selectdev server (vite) - Launch the debugger by opening
main.rsand pressing F5
Theoretically, it should be possible to automatically launch the Vite dev server when debugging starts, but I haven't managed it yet.
Also works in VS Code by starting vite server with pnpm vite and then clicking the "Debug" button in main function
In order to use sqlx's query!() macro, you temporarily need a database at compile time and set its path as the DATABASE_URL environment variable. Then, update the offline cache via CLI:
# ./backend
DATABASE_URL=sqlite:///C:/path/to/sqlite.db cargo sqlx prepare --workspace
This command creates an offline store for compile-time checks of sqlx::query!() without needing a constantly available database. The command has to be rerun occasionally, for example when the schema changes or when a new query is added.