- 2D Scatter Statistics
- Regressions (linear, poly, custom)
- Contribution to sample mean and contribution to sample variance
- Get custom distributions working
- Tornado plots (?)
- Variable contribution plots
- Example: Evidence-Based Scheduling
- 2D/3D Line statistics
- Correlated inputs
- Switch from prints to logging
- Ability to plot derived data in addition to vars
- Star Sampling
- Demo running on remote server (AWS, etc)
- Variogram Sensitivity Analysis
- repr's for all main objects
- Flag on invals to tell if they were loaded from file
- Histogram/spaghetti plot coloring a-la aleatory
- Cannot plot a varstat in percentile space
- Python 3.13 support
- Python 3.9 support
- Fixed docs build not including docstrings
- Numpy 2.0 real compatibility & wheels on PyPi
- Testing for partial failures of cases while running with different
debugflags (GH-10)
- Support newer versions of
dask.distributed(GH-12)
tqdm_dask_distributedsupporting function, replaced bydask.distributed.progress
- Numpy 2.0 code compatibility attempt
- Allow specifying
distsanddistkwargsfor invars loaded from file
- Fix cases not being populated with invars and invals when loading from file (GH-9)
- Fix nummaps not being per-variable when loading outvars from file
- Fix histogram not plotting for imported invars (GH-8)
- Fix broken docs build
- Automated github publishing to PyPI
- Update github action runners
- Fix underflow issue in dvars (GH-7)
- Python 3.12 support
- Documentation cleanup
- Halton sequence seeded draws changed with scipy 1.11, see scipy issue #18079
- Python 3.8 support
- More plotting test coverage.
- Allow passing in varname strings to
sim.plot.
- License changed from GPLv3 to more permissive MIT (ok because single author project).
- Fix
sim.plotscalarvars not being used (GH-6)
- Python 3.11 support
numbais now an optional dependency
- Automated testing for plots and multiplots.
- Parallel processing now chains preprocessing, running, and postprocessing into a single dask task graph via Sim.executeAllFcns(), giving large speed boost
- Sim functions that take in cases default to None (all cases)
- Fixed bug plotting histograms
- Fixed another bug generating and plotting varstat ranges from mixed-length data
- Fixed bug generating and plotting percentile ranges from mixed-length data
mc_multi_plot.multi_plot_grid_rectplotting, made default formc.plot__repr__for Cases and Vals
mc_multi_plot.multi_plot_2d_scatter_gridrenamed tomulti_plot_grid_tri- Fixed bug when plotting against simulation steps of different lengths
- Make Sims singlethreaded by default
- Export outvar nums to a csv or json file with
sim.exportOutVars - Import invars from a csv or json with
sim.importInVars - Optionally plot a contour underneath a 2D scatter plot with a third variable
- Flags to keep siminput and simrawoutput for each case (default True)
- Added Sim.vars to reference all invars and outvars
- Variance, skewness, kurtosis, and moment added as varstats
sim.exportInVarNumsrenamed tosim.exportInVarssim.importOutValsrenamed tosim.importOutVars- Refined and removed upper triangle of multiplot grid
- Choose whether to sort sensisitivities while plotting.
- Added
Sim.plot()method - Export invar nums to a csv or json file with
sim.exportInVarNums - Import outvals from a csv or json with
sim.importOutVals, and convert to outvars - 2.5D plotting (scalar vs vectors)
- Prevent overwriting existing InVars or OutVals with an already used name.
- Do not save sim and case data by default
- Fix splitting pairs of variables when plotting
- DVARS fleshed out, moved out of alpha
- Can now plot sensitivity indices and ratios
- Added
singlethreadedkwarg toSiminitialization. - Added
daskkwargskwarg toSiminitialization. - Added
percentilevarstat.
- For all datafiles, switch from
dilltocloudpicklefor pickling. - Parallel processing backend moved from
pathostodask. Sim.PreProcessCase,Sim.RunCase, andSim.PostProcessCasebroken out tocase_runners.pyfunctions.
- Removed
coreskwarg fromSiminitialization. - Removed python 3.7 support, to align with dask.
- Plots of all the example statistical distributions
- Bootstrapping confidence interval for a VarStat
- Plot bootstrapped confidence intervals
- Copy of template in jupyter notebook format
- Change color palette to match matplotlib defaults
- Discrete InVar plots show stem pmfs
- Vars get their own
plot()method as a shorthand - Baseball example
- Rocket example
- Experimental D-VARS sensitivity analysis
Sim.scalarOutVars()dict andSim.noutvars- You can specify a custom varstat
- You can plot plot in percentile space rather than nums
- Removed mc_ and MC prefixes from all functions and classes
- Python 3.7.0 support (pandas testing made optional)
- All functions are now imported into top-level package, so you don't need to dig through modules
- Python 3.7 support (>=3.7.1 to match pandas)
Val.nums were made numpy arraysVals have ashapeattribute rather thansize, to match with numpyhelper_functions.get_tuple()changed toget_list()- Simplify code for extracting valmaps
- More typing
- Python 3.8 support
- Pre-commit hooks for linting
- Dusted off the code with lots of linting
- Made all modules lowercase
- pandas is now an optional dependency
- Documentation is now up on readthedocs!
- Docstrings for all classes and functions, roughly follows the numpy docstrings convention
- Analysis process diagram
- More unit tests
- Documentation and images were moved around.
- Project logo
- Added multi_plot_2d_scatter_grid()
- Python 3.10 support
import monaconow imports all modules- The 'nom' case is changed to the 'median' case
- No more inline tests
- Documentation updates
- Better type consistency
- Testing on more platforms
- Small bugfixes
- Initial release!