Please sign in to comment.
Updating sensitivity analysis class to be general for any model param…
…eters (#426) * Updating Sensitivity analysis class to be used with any parameters, not just initial conditions. Also added ability to pass a list of parameter names rather than requiring all from the model to be used. * Adding check to if passed in list of parameters is actually in the model. * Adding examples in documentation for sens_type='params' and with a list of parameters. Adding tests for parameter list. * Adding test for using deprecated class name (InitialsSensitivity). This will maintain compatibility. * Fixing doc tests. Printing differences between python versions gave some errors. They seem to be fixed. * Adding last test case where sens_type=None and sample_list=None * Removed a lot of the initializations of the solver. Not needed. Added test to catch a non pysb simulator. * Fixing logic in checking if sens_type sample_list is None and sens_type is not in approved list. * Addressing changes for pull request. * Changing doctest to verify size of array, rather than printing. Windows represents np.array.shape as type L, where on linux it uses it uses int. This is due to differences in C types. Pointless fix in the long run, but now it should pass doctests on python2 for windows. It is discussed here. numpy/numpy#5809 * Adding attribute and parameter descriptions. I made some attributes private that were supposed to be for internal calculations. * Documentation fixes for sensitivity analysis class
- Loading branch information...
Showing with 295 additions and 134 deletions.
|@@ -0,0 +1,5 @@|
|Sensitivity anaylsis (:py:mod:`pysb.tools.sensitivity_analysis`)|
|.. automodule:: pysb.tools.sensitivity_analysis|
Oops, something went wrong.