• fixed issues with numpy > = 2.0 (see #140, numpy.NaN deprecated and removed)
• fixed issues with refactor of scipy.stats (see #139)

• added testing for python 3.12
• fixed issue with matplotlib >= 3.8.0 (see #129)
• added missing documents to dream manual (Convergence tests, Parallel coordinates plot)
• added numba.njit-accelerated fallback bounds.apply methods to dream (still uses compiled C DLL if available)
• provide MAX_CORR attribute on the CorrelationView; clear the figure if the number of variables exceeds MAX_CORR

• added support for python 3.11, scipy 1.10, numpy 1.24, wx 4.1.1
• fixed covariance calculation for n-D datasets
• fixed batch mode I/O redirection cleanup
• fixed issue with DREAM bounds checker when running in parallel
• default to single precision derivatives with lm (fixes issue in SasView where OpenCL models failed with Levenberg-Marquardt)
• improved support for repeat fitting within scripts and notebooks (start_mapper should now work after stop_mapper)

• use MPFit in place of scipy.leastsq for bounds-constrained Levenberg-Marquardt

Breaking change: * simultaneous fit now scales individual nllfs by squared weight rather than weight

• "apply parameters" action added to GUI menu (does the same as --pars flag in CLI)
• operators refactored (no more eval)
• BoundedNormal keywords renamed (sigma, mu) -> (std, mean)
• support for numba usage in models
• fixed Parameters view jumping to top after toggling fit (linux, Mac)
• fixed Summary view sliders disappearing in linux
• fixed uncertainty plots regenerating at each parameter update
• improved documentation of uncertainty analysis

Breaking change: * python 2.7 support discontinued

• add stopping conditions to DREAM, using --alpha=p-value to reject convergence
• require --overwrite or --resume when reusing a store directory
• enable outlier trimming in DREAM with --outliers=iqr
• add fitted slope and loglikelihood distribution to the loglikelihood plot
• display seed value used for fit so it can be rerun with --seed
• save MCMC files using gzip
• remove R stat from saved state
• restore --pars option, which was broken in 0.7.17
• terminate the MPI session when the fit is complete instead of waiting for the allocation to expire
• allow a series of fits in the same MPI session
• support newest matplotlib

• restore python 2.7 support

• restore DREAM fitter efficiency (it should now require fewer burn-in steps)
• errplot.reload_errors allows full path to model file
• clip values within bounds at start of fit so constraints aren't infinite
• allow --entropy=gmm|mvn|wnn|llf to specify entropy estimation algorithm
• allow duplicate parameter names in model on reload
• expand tilde in path names
• GUI: restore parallel processing
• GUI: suppress uncertainty updates during fit to avoid memory leak
• disable broken fitters: particle swarm, random lines, snobfit
• minor doc changes

• improved handling of parameters for to_dict() json pickling

• parallel fitting suppressed in GUI for now---need to reuse thread pool
• support limits=(min, max) for pm and pmp parameter ranges
• cleaner handling of single/multiple fit specification
• fix --entropy command line option
• better support for pathlib with virtual file system

• support for --checkpoint=n, which updates the .mc files every n hours
• fix bug for stuck fits on --resume: probabilities contain NaN
• better error message for missing store directory
• Python 3.8 support (time.clock no longer exists)

• fix pickle problem for parameterized functions
• support multi-valued functions in Curve, shown with a coupled ODE example
• update support for newer numpy and matplotlib

• --parallel defaults to using one process per CPU.
• --pop=-k sets population size to k rather than k times num parameters
• --resume=- resumes from --store=/path/to/store
• use expanded canvas for parameter histograms to make plots more readable
• use regular spaced tics for parameter histograms rather than 1- and 2-sigma
• improve consistency between values of cov, stderr and chisq
• fix handling of degenerate ranges on parameter output
• add entropy calculator using gaussian mixture models (default is still Kramer)
• vfs module allows loading of model and data from zip file (not yet enabled)
• warn when model has no fitted parameters
• update mpfit to support python 3
• support various versions of scipy and numpy

• add support for parameter serialization

• restructure parameter table in gui

• full support for python 3 in wx GUI
• allow added or missing parameters in reloaded .par file
• add dream state to return from fit() call

• fix source distribution (bin directory was missing)

• merge in amdahl branch for improved performance
• update plot so that the displayed "chisq" is consistent with nllf
• slight modification to the DREAM DE crossover ratio so that no crossover weight ever goes to zero.
• par.dev(std) now uses the initial value of the parameter as the center of the distribution for a gaussian prior on par, as stated in the documentation. In older releases it was incorrectly defaulting to mean=0 if the mean was not specified.
• save parameters and uncertainties as JSON as well as text
• convert discrete variables to integer prior to computing DREAM statistics
• allow relative imports from model files
• support latest numpy/matplotlib stack
• initial support for wxPhoenix/python 4 GUI (fit ranges can't yet be set)

• add --view option to command line which gets propagated to the model plotter
• add support for probability p(x) for vector x using VectorPDF(f,x0)
• rename DirectPDF to DirectProblem, and allow it to run in GUI
• data reader supports multi-part files, with parts separated by blank lines
• add gaussian mixture and laplace examples
• bug fix: plots were failing if model name contains a '.'
• miscellaneous code cleanup

• gui: undo code cleaning operation which broke the user interface

• population initializers allow indefinite bounds
• use single precision criterion for levenberg-marquardt and bfgs
• implement simple, faster, less accurate Hessian & Jacobian
• compute uncertainty estimate from Jacobian if problem is sum of squares
• gui: fit selection window acts like a dialog

• accept model.par output from a different model
• show residuals with curve fit output
• only show correlations for selected variables
• show tics on correlations if small number
• improve handling of uncertainty estimate from curvature
• tweak dream algorithm -- maybe improve the acceptance ratio?
• allow model to set visible variables in output
• improve handling of arbitrary probability density functions
• simplify loading of pymc models
• update to numdifftools 0.9.14
• bug fix: improved handling of ill-conditioned fits
• bug fix: avoid copying mcmc chain during run
• bug fix: more robust handling of --time limit
• bug fix: support newer versions of matplotlib and numpy
• miscellaneous tweaks and fixes

• add entropy calculator (still unreliable for high dimensional problems)
• adjust scaling of likelihood (the green line) to match histogram area
• use --samples to specify the number of samples from the distribution
• mark this and future releases with a DOI at zenodo.org

• tweak uncertainty calculations so they don't fail on bad models

• documentation updates

• use relative rather than absolute noise in dream, which lets us fit target values in the order of 1e-6 or less.
• fix covariance population initializer

• use --time to stop after a given number of hours
• Levenberg-Marquardt: fix "must be 1-d or 2-d" bug
• improve curvefit interface

• pull numdifftools dependency into the repository

• improve the load_model interface

• Pure python release