- [x] Added real data based test as well as pure synthetic - [x] Added load and dump function to exchange standardized CSV - [x] Corrected sigma management when only scalar or None are used inplace of array - [x] Added synthetic dataset generation, bound for unit test operations - [x] Started test with real dataset
0.1.3
- [x] Added special solvers: Debye Heat Capacity, Crank Diffusion, Raney Keton Dehydrogenation - [x] Added log scale mode for fit and loss figures - [x] Updated nox commands for building and cache - [x] Updated notebooks and documentation
0.1.2
- [x] Added error surface in addition with contour levels - [x] Added book test suite for resources generation - [x] Added dataset export from solver interface - [x] Updated CI pipeline - [x] Corrected typo in figures - [x] Adapted quality test to be more realistic - [x] Added fake sigma capability to tests bad chi square regression - [x] Completed publication workflow on PyPi
0.1.1
- [x] Added lot of logistic solvers - [x] Added minimizer to check `curve_fit` - [x] Added CI pipeline for GitHub
0.1.0
- [x] First beta version of the package - [x] Created first solver interface for fitting problems (`FitSolverInterface`) - [x] Complete fitting procedure with `SciKit-Learn` compliant interface - [x] Implemented Chi Square Goodness of Fit tests for fitting procedure - [x] Created a bunch of fit model (linear and scientific namespaces) - [x] Created a bunch of tests to assess capabilities - [x] Created summary figures: - [x] Fit Plot to check adjustment - [x] Loss Plot to check parameters convergence and uniqueness (low dimensional and scatter) - [x] Chi Square Goodness of Fit Plot to check fitting compliance - [x] Started Sphinx documentation for whole package - [x] Created quick start guide using Jupyter notebooks