Dashai

------

- Added frontend development documentation.
- improved backend development documentation.
- Change front package management to yarn.
- Improved gitignore for frontend.
- Added missing packages that allow the front end to run locally
- Added prettier as default frontend formatter for the project.
- Formatted all front end files according to prettier styles.
- Changed navbar to a MUI responsive implementation.
- Fix fronend router and links.
- Add design colors to MUI theme provider.
- Moved route handling to App.jsx
- Change page container to MUI container component.
- The experiment table and new experiment button is now implemented using MUI's
Datagrid and Button. Its styles and design were also improved.
- Moved the experiments API to its own module in `src/api/experiments`. Is by default typed.
- ExperimentsTable component now imports these methods and executes them in an asynchronous wrapper inside the component.
- Interactivity is added through snackbars/toast and loading animations. These can be seen when the getExperiments request fails and when an experiment is deleted (correctly or incorrectly).
- Typescript is enabled on the front end.
- Interfaces are declared for the main application objects: Experiment, Dataset and Run. These can be found in the `src/types`` folder.

------

**Main change**: Dataset endpoints
New dataset endpoints on `DashAI/back/api_v1` follow the RESTful api paradigm

Changes
*******

- Implemented put and delete on datasets
- Implemented get on /dataset/ and /dataset/id
- Now all datasets raise HTTPException if an error occurs.
- Fixed docstring styling

------

**Main change**: Legacy Endpoints
Enabled again some endpoints that have been purged after the repository changes.
We added multiple api versions and fixed frontend to work with the new api structure.

Changes
*******

- Added api_v0 and api_v1
- Legacy endpoints are implemented in `DashAI/back/api/api_v0/endpoints/old_endpoints.py`
- Fixed some bugs on dataloaders typing
- Fixed a bug on `DashAI/back/dataloaders/classes/dataloader.py` when splitting and given test and train sizes to train_test_split method (float sum error)

- Changed .env file to point new endpoints
- Fixed react navigate from '/...' to '/app/...'

-----

**Main change**: Merged ModelInstance and Run in one table

Changes
*******

- Base was refactored to database/models, so fixed import in init
- Merged the ModelInstance table with the Run table

-----

**Main change**: Serving from FastAPI
Remove a secondary server to serve our front end. Also, it enhances GitHub workflow to build the frontend and then run pytest.

Fixed some .gitignore redundancies

Changes
*******

- Unified pytest and npm test to a single workflow with two jobs
- pytest job uses the react build generated on the npm cli cycle.
- Removed the secondary process that ran the server on python's http.server
- DashAI/back/database/db.py
- Fixed deprecated import in SQLAlchemy 2.0
- Enabled frontend serve and static files on FastAPI
- Renamed app path from / to /app
- Renamed api path from / to /api
- Added database health check on execution.
- Renamed a bunch of paths from images/* to /images/*
- Renamed NavBar's paths from /* to /app/*

-----

**Main change**: Database model
Cleaned database model.

Changes
*******

- We are now using the modern version of SQLAlchemy Mapped syntax
- We have 4 tables now: Dataset, Experiment, Model and Run
- Dataset One-to-Many Experiment
- Experiment One-to-Many Model
- Model One-to-One Run
- Model and Run cascades on deletion.
- Added states for the future state machine.