Pypesto

------------------

* Finite Differences:
* Adaptive finite differences (671)
* Add helper function for checking gradients of objectives (690)
* Small bug fixes (711, 714)

* Storage:
* Store representation of the objective (669)
* Minor fixes in HDF5 history (679)
* HDF5 reader for ensemble predictions (681)
* Update storage demo jupyter notebook (699)
* Option to trim trace to be monotonically decreasing (705)

* General:
* Improved tests and bug fixes of validation intervals (676, 685)
* Add input file validation via PEtab linter for PEtab import (678)
* Remove default values from docstring (680)
* Minor fixes/improvements of ensembles (687, 688)
* Fix sorting of optimization values including `NaN` values (691)
* Specify axis limits for plotting (693)
* Minor fixes in visualization (696)
* Add installation option `all_optimizers` (695)
* Improve installation documentation (689)
* Update `pysb` and `BNG` version on GitHub Actions (697)
* Bug fix in steady state guesses (715)

------------------

* Objective:
* Basic finite differences (666)
* Fix factor 2 in res/fval values (619)

* Optimization:
* Sort optimization results when appending (668)
* Read optimizer result from HDF5 (previously only CSV) (663)

* Storage:
* Load ensemble from HDF5 (640)

* CI:
* Add flake8 checks as pre-commit hook (662)
* Add efficient biological conversion reaction test model (619)

* General:
* No automatic import of the predict module (657)
* Assert unique problem parameter names (665)
* Load ensemble from optimization result with and without history usage
(640)
* Calculate validation profile significance (658)
* Set pypesto screen logger to "INFO" by default (667)

* Minor fixes:
* Fix axis variable overwriting in `visualize.sampling_parameter_traces`
(665)

------------------

* Objectives:
* New Aesara objectve (623, 629, 635)

* Sampling:
* New Emcee sampler (606)
* Fix compatibility to new Theano version (650)

* Storage:
* Improve hdf5 storage documentation (612)
* Hdf5 history for MultiProcessEngine (650)
* Minor fixes (637, 638, 645, 649)

* Visualization:
* Fix bounds of parameter plots (601)
* Fix waterfall plots with multiple results (611)

* CI:
* Move CI tests on GitHub Actions to python 3.9 (598)
* Add issue template (604)
* Update BionetGen Link (630)
* Introduce project.toml (634)

* General:
* Introduce progress bar for optimization, profiles and ensembles (641)
* Extend gradient checking functionality (644)

* Minor fixes:
* Fix installation of ipopt (599)
* Fix Zenodo link (601)
* Fix duplicates in documentation (603)
* Fix least squares optimizers (617 631 632)
* Fix trust region options (616)
* Fix slicing for new AMICI release (621)
* Refactor and document latin hypercube sampling (647)
* Fix missing SBML name in PEtab import (648)

------------------

* Ensembles/Sampling:
* General ensemble analysis, visualization, storage (557, 565, 568)
* Calculate and plot MCMC parameter and prediction CIs via ensemble
definition, parallelize ensemble predictions (490)

* Optimization:
* New optimizer: SciPy Differential Evolution (543)
* Set fides default to hybrid (578)

* AMICI:
* Make `guess_steadystate` less restrictive (561) and have a more
intuitive default behavior (562, 582)
* Customize time points (490)

* Storage:
* Save HDF5 history with SingleCoreEngine (564)
* Add read/write function for whole results (589)

* Engines:
* MPI based distributed parallelization (542)

* Visualization:
* Speed up waterfall plots by resizing scales only once (577)
* Change waterfall default offset to 1 - minimum (593)

* CI:
* Move GHA CI tests to pull request level for better cooperability (574)
* Streamline test environments using tox and pre-commit hooks (579)
* Test profile and sampling storage (585)
* Update for Ubuntu 20.04, add rerun on failure (587)

* Minor fixes (release notes 558, nlop tests 559, close files 495,
visualization 554, deployment 560, flakiness 570,
aggregated deepcopy 572, respect user-provided offsets 576,
update to SWIG 4 591, check overwrite in profile writing 566)

------------------

* New optimizers:
* FIDES (506, 503 500)
* NLopt (493)

* Extended PEtab support:
* PySB import (437)
* Support of PEtab's initializationPriors (535)
* Support of prior parameterScale{Normal,Laplace} (520)
* Example notebook for synthetic data generation (482)

* General new and improved functionality:
* Predictions (544)
* Move tests to GitHub Actions (524)
* Parallelize profile calculation (532)
* Save `x_guesses` in `pypesto.problem` (494)
* Improved finite difference gradients (464)
* Support of unconstrained optimization (519)
* Additional NaN check for fval, grad and hessian (521)
* Add sanity checks for optimizer bounds (516)

* Improvements in storage:
* Fix hdf5 export of optimizer history (536)
* Fix reading `x_names` from hdf5 history (528)
* Storage does not save empty arrays (489)
* hdf5 storage sampling (546)
* hdf5 storage parameter profiles (546)

* Improvements in the visualization routines:
* Plot parameter values as histogram (485)
* Fix y axis limits in waterfall plots (503)
* Fix color scheme in visualization (498)
* Improved visualization of optimization results (486)

* Several small bug fixes (547, 541, 538, 533, 512, 508)

------------------

* New optimizer: CMA-ES (457)
* New plot: Optimizer convergence summary (446)

* Fixes in visualization:
* Type checks for reference points (460)
* y_limits in waterfall plots with multiple results (475)
* Support of new amici release (469)

* Multiple fixes in optimization code:
* Remove unused argument for dlib optimizer (466)
* Add check for installation of ipopt (470)
* Add maxiter as default option of dlib (474)

* Numpy based subindexing in amici_util (462)
* Check amici/PEtab installation (477)