Jenn

Fix

- missing dependencies (`jsonschema`, `jsonpointer`)
- missing data (*.json was not being included in build, so added MANIFEST.in)
- typing oversight for python 3.8 (in `cost.py` and `sythetic.py`)

Fix

- Fixed random seed not working (previously not being passed to parameter initialization)
- Fixed `minibatch` issue throwing error below when `shuffle=False` and more than one batch

Traceback (most recent call last):
File "C:\[...]\jenn\model.py", line 141, in fit
self.history = train_model(
File "C:\[...]\jenn\core\training.py", line 121, in train_model
batches = data.mini_batches(batch_size, shuffle, random_state)
File "C:\[...]\jenn\core\data.py", line 229, in mini_batches
batches = mini_batches(X, batch_size, shuffle, random_state)
File "C:\[...]\jenn\core\data.py", line 51, in mini_batches
if mini_batch:
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()


Refactor

- Added jsonschema to validate reloaded parameters and check array shapes
- Added levels as input to `plot.contours`

Features

- Added optional ability to prioritize individual training points (useful to ensure more accuracy in known regions of interest)
- Added optional ability to warmstart; i.e. continue training from current parameters (without initialization)
- Exposed more hyperparameters pertaining to optimizer (e.g. tolerance stopping criteria)
- Added option to use finite difference for generating synthetic data partials (used to study effect noisy partials)

Documentation

- Added airfoil notebook as example of large dataset
- Added surrogate-based optimization notebook to demonstrate benefit of JENN
- Updated theory.pdf

Fix

- Updated annotations in `jenn.utils.plot` which were incompatible with Python 3.8 (causing runtime errors)
- Manually updated `__version__` number inside `__init__` (previous oversight)

Documentation

- Update demo examples to use `from jenn.utils import plot` instead of `jenn.utils.plot` (which failed a test on Python 3.11.7)

Feature

- Added support for `python >= 3.8`

Documentation

- Update [paper link](https://github.com/shb84/JENN/blob/master/docs/theory.pdf) to point to version on `master` instead of `refactor` branch

Fix

- Fixed `minibatch` which was previously not reusing parameters from one batch to another
- Fixed `random_state` which was previously was not being passed everywhere it should
- Resolved rrror messages when optional `matplotlib` library not installed

Feature

- Added support for `python >= 3.9`

_This release introduces breaking changes but makes the algorithm about 5x faster._

Feature

- Added static, sensitivity profiles as new plotting utility
- Added `synthetic` module to create example data from canonical test functions
- Added `evaluate` method to `NeuralNet` model (which does `predict` and `predict_partials` in one step more efficiently)

Documentation

- Added documentation using `sphinx` and published on GitHub Pages
- Added more example notebooks

Refactor

- Moved core API into its own subpackage
- Moved core API data management functionality into their own classes: `Parameters`, `DataSet`, `Cache`
- Moved plotting module and metrics into utilities subpackage
- Renamed core API modules, classes, and functions
- Renamed user API modules, classes, and functions
- Changed user API by adding NeuralNet model
- Changed user API plotting utility functions names and kwards
- Changed datastructure exposed to user (compared to `v0.1.0`, `X, Y, J` are now transposed)
- Simplified almost all functions for easier maintainability/readability

Style

- Using `ruff`, `docformatter`, `black` and `mypy` for linting

Performance

- Arrays are now updated in place (code about 5x faster)

Tests

- Added unit tests using `pytest`