https://github.com/tomashubelbauer/apple-notes-database
https://github.com/tomashubelbauer/apple-notes-database
apple-notes apple-script
Last synced: 4 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/tomashubelbauer/apple-notes-database
- Owner: TomasHubelbauer
- Created: 2023-08-28T20:14:55.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2023-09-04T16:30:23.000Z (almost 2 years ago)
- Last Synced: 2024-12-31T22:13:57.545Z (6 months ago)
- Topics: apple-notes, apple-script
- Language: JavaScript
- Homepage:
- Size: 42 KB
- Stars: 2
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Apple Notes Database
[](https://github.com/TomasHubelbauer/apple-notes-database/actions/workflows/playwright.yml)
This project is an experiment in using Apple Notes as a storage layer for a web
application.
It is not meant to be practical, it is a creative exercise.
## Running
Run `PORT=3000 node .` to run the server.
The UI is at `/` and allows listing notes and reading and writing notes by ID.
There is also a to-do list sample running at `/todo.html`.
## Design
Entity/type approximates a database table and is implemented using Apple Notes
folders.
There is an encompassing folder for the application named after `appName.js`.
## API
- `GET type` - list the IDs of entities of the given type
The ID can be any string of characters except for a line break and a comma.
- `GET type/id` - get the content of the entity of the given type by its ID
The return value is a JSON object literal with string key-value pairs.
- `POST type/id` - set the content of the entity of the given type by its ID
The expected body is of the same format as the corresponding GET method.
There is no return value.
- `DELETE type/id` - delete the note of the given type.
There is no return value.
## Testing
The app uses Playwright for E2E testing.
Run the tests using `npm test`.
The test runner will start the app server itself.
## To-Do
### See if the Playwright browser download could be cached for say a week
We want to refresh it every once in a while, but every test run seems too much.
Let's see if I can use the GitHub Actions cache to do this.
### Switch the macOS image to macOS 12 while it is latest
I got help making the AppleScript work for macOS 13:
https://github.com/actions/runner-images/issues/8214
This image is in beta though.
I should find a way to do this in macOS 12 and only switch to macOS 13 once it
it latest.
### Consider splitting `POST` and `PUT` into different endpoints
Maybe I should respect the semantics for `POST` being for creating new items and
`PUT` being for updating existing items and erroring if the item doesn't exist.
In this scenario, `POST` would be a single-component path of just `POST /type`
and would automatically assign a unique ID or it could optionally have a
two-component path variant `POST /type/id` with suggested ID (based on say a
friendly name) which would be checked for uniqueness.
`PUT` would always require the ID: `PUT /type/id`.
### Add support for ID-less multi-deletion on `DELETE /type`
This will bring parity to ID-less multi-retrival at `GET /type?full`.
It will be useful for building file-system-like experiences on top of the notes.
It should be accompanied with a method for creating a folder (different from
indirectly via `GET /type` or `GET /type?full` through `ensureTypeFolder`) so
that the folder CRUD actions can be carried out end to end over the API.
### Add a new app with multi-select support to demo the multi-deletion feature
I need to test this feature more as well, it seems kind of unreliable and I had
trouble putting together the AppleScript for it.
In some versions which I attempted to make slicker I would get seemingly racy
errors, sometimes it would delete the notes successfully, sometimes it would
error with unhelpful vague error messages.
The current version when passed a set of IDs sometimes succeeds but doesn't
actually seem to delete the notes?
I might need to craft better AppleScript or add checks after the command with
retry logic maybe?