Adding new Jacobians
FitBenchmarking allows the use of custom methods for evaluating the Jacobian of the model. This section describes how to add further methods within FitBenchmarking
In order to add a new Jacobian evaluation method, <jac_method>
,
you will need to:
Create
fitbenchmarking/jacobian/<jac_method>_jacobian.py
, which contains a new subclass ofJacobian
. Then implement the method:- abstract Jacobian.eval()
Evaluates Jacobian of the model, \(\nabla_p f(x,p)\), at the point given by the parameters.
- Parameters
params (list) – The parameter values at which to evaluate the Jacobian
- Returns
Computed Jacobian
- Return type
numpy array
The numerical method is set sequentially within
loop_over_jacobians()
by using themethod
attribute of the class.Enable the new method as an option in Fitting Options, following the instructions in Adding new Options. Specifically:
Amend the
VALID_FITTING
dictionary so that the element associated with thejac_method
key contains the new<jac_method>
.Extend the
VALID_JACOBIAN
dictionary to have a new key<jac_method>
, with the associated element being a list of valid options for this Jacobian.Extend the
DEFAULT_JACOBIAN
dictionary to have a new key<jac_method>
, with the associated element being a subset of the valid options added inVALID_JACOBIAN
in the previous step.Amend the file
fitbenchmarking/utils/tests/test_options_jacobian.py
to include tests for the new options.
Document the available Jacobians by:
adding a list of available
method
options to the docs for Jacobian Options, and including licencing information, if appropriate.updating any example files in the
examples
directory
Create tests for the Jacobian evaluation in
fitbenchmarking/jacobian/tests/test_jacobians.py
.
The FittingProblem
class
When adding new Jacobian, you will find it helpful to make use of the
following members of the FittingProblem
class:
- class fitbenchmarking.parsing.fitting_problem.FittingProblem(options)
Definition of a fitting problem, which will be populated by a parser from a problem definition file.
Onces populated, this should include the data, the function and any other additional requirements from the data.
- data_e
numpy array The errors or weights
- data_x
numpy array The x-data
- data_y
numpy array The y-data
- eval_model(params, **kwargs)
Function evaluation method
- Parameters
params (list) – parameter value(s)
- Returns
data values evaluated from the function of the problem
- Return type
numpy array