v0.3.3.1 (2018-02-10) --------------------- * Fixed bug in which MarkovChain.steady_state might use the wrong eigenvector. * Added warnings if transition probabilities are negative or don't sum to 1. v0.3.3.0 (2018-02-08) --------------------- * plot_path now only plots the path if the probability is nonzero * plot_path takes in a starting condition to be consistent with prob_of_path * simulate_path takes an optional parameter to plot the path v0.3.2.1 (2018-02-08) --------------------- * Only show real component of steady state distribution when computing left eigenvector. v0.3.2.0 (2018-01-18) --------------------- * Restore functionality for calling `probability_function` for joint distributions. v0.3.1.1 (2018-01-08) --------------------- * MarkovChain.distribution() supports a state as a starting distribution. v0.3.1.0 (2018-01-08) --------------------- * Code refactor to follow PEP8 * All new Markov Chain module using numpy backend * Function definitions largely the same for common functions * MarkovChains can now be constructed using additional class functions * Functions renamed: * Table.expected_value -> Table.ev * Table.variance -> Table.var * Unit Tests! Bumped to around 66% code coverage. v0.2.9.0 (2017-03-19) --------------------- * Plot_3d v0.2.8.1 (2017-03-18) --------------------- * Plot_continuous now accepts python functions too v0.2.8.0 (2017-03-13) --------------------- * Updated unconstrain to rearrange_2 and nicefy to rearrange_1 v0.2.7.1 (2017-03-11) --------------------- * SymPy integration being finalized - added `unconstrain` and updated `declare` v0.2.7.0 (2017-03-10) --------------------- * sample renamed to sample_from_dist to avoid conflicts with datascience v0.2.6.3 (2017-03-09) --------------------- * Fixed documentation for plots * plots removed from global v0.2.6.2 (2017-03-09) --------------------- * Plot_continuous works with sympy v0.2.6.1 (2017-03-09) --------------------- * Plot_continuous now works with any function passed in as func v0.2.6.0 (2017-03-06) --------------------- * Wrapper for plotting continuous functions v0.2.5.1 (2017-03-06) --------------------- * Beginning to add SymPy integration in *symbolic_math.py* v0.2.5.0 (2017-02-22) --------------------- * Added log_probability_of_path v0.2.4.4 (2017-02-20) --------------------- * Fixing installation issues v0.2.4.3a (2017-02-20) ---------------------- * fixed mfpt v0.2.4.2 (2017-02-16) --------------------- * Fixed typo in steady_state, not sure how it happened v0.2.4.1 (2017-02-16) --------------------- * Documentation fix v0.2.4.0 (2017-02-13) --------------------- * Removed T and S from markov chains * added .column * states now sorted v0.2.3.8 (2017-02-13) --------------------- * Added get target v0.2.3.7 (2017-02-12) --------------------- * Deprecation error fix v0.2.3.6 (2017-02-12) --------------------- * Distribution now shows states with probability 0 v0.2.3.5 (2017-02-11) --------------------- * Added show_ev for conditional distributions v0.2.3.4 (2017-02-11) --------------------- * state --> states v0.2.3.3 -------- * Documentation v0.2.3.2 (2017-02-11) --------------------- * Changed label for empirical distribution to state * mc.distribution accepts states v0.2.3.1 (2017-02-11) --------------------- * Fixed mean_first_passage_times v0.2.3.0 (2017-02-11) --------------------- * Renamed a ton of functions * Implemented starting conditions v0.2.2.0 (2017-02-11) --------------------- * Begin wrapping of pykov v0.2.1.3 (2017-02-08) --------------------- * Plots uses plt.bar instead of Table.hist * Added optional parameter edges= v0.2.1.2 (2017-02-04) --------------------- * Added show_ave as optional parameter v0.2.1.1 (2017-02-04) --------------------- * Added show_ev and show_sd as optional parameters for plot v0.2.1.0 (2017-02-04) --------------------- * Added sample for single variable distributions * Added CDF for single variable distributions v0.2.0.0 (2017-02-03) --------------------- * Pykov v0.1.8.1 (2017-02-01) --------------------- * Renamed emp_dist values to proportions rather than probabilities v0.1.8.0 (2017-01-30) --------------------- * Added emp_dist to allow for empirical distributions v0.1.7.6 (2017-01-19) --------------------- * __version__ instead of version v0.1.7.5 (2017-01-18) --------------------- * Joint Distributions no longer give a warning if probabilities rounded to 6 decimal places = 1 v0.1.7.4 (2017-01-17) --------------------- * Single variable distributions now check that probabilities sum to 1 v0.1.7.3 (2017-01-17) --------------------- * Plot now adds edge border if there are fewer than 75 bins * Plot now has an optional parameter edge that accepts a boolean * Added marginal_dist which returns a single variable distribution v0.1.7.2 (2017-01-17) --------------------- * .values is now an alias for .domain v0.1.7.1 (2017-01-17) --------------------- * Fixed vertical axis for Plot v0.1.7.0 (2017-01-16) --------------------- * Removed marginal_of_X, marginal_of_Y, etc * conditional_dist_given(given) is now conditional_dist(label, given) v0.1.6.4 (2017-01-15) --------------------- * Joint Distribution functions can have arbitrary number of arguments again v0.1.6.3 (2017-01-15) --------------------- * fixed a bug in which toJoint just renamed the x-columns rather than changing the order v0.1.6.2 (2017-01-14) --------------------- * toJoint now preserve original order v0.1.6.1 (2017-01-14) --------------------- * JointDistribution probabilities don't have to sum to 1, v0.1.6 (2017-01-14) ------------------- * Added probability_function for JointDistribution * probability_function now checks number of arguments in pfunc v0.1.5.1 (2017-01-12) --------------------- * Added JointDistribution to the init v0.1.5 (2017-01-12) ------------------- * Plotting width now works with events and masks * JointDistribution can now be used with any variable v0.1.4.3 (2016-12-20) --------------------- * Changed the colors for plots v0.1.4.2 -------- * Slight modifications to plot labels v0.1.4a ------- * Single distribution plotting moved from the ``plot_dist`` method to the ``Plot`` function * Multiple distribution plotting moved from the ``Plot`` function to the ``Plots`` function * Events are now plotted by passing an argument to ``Plot`` v0.1.3 ------ * Added joint distributions * All ``FiniteDistribution`` objects changed to become ``datascience.tables.Table`` objects * Began renaming v0.1.2 ------ Initial Release