This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
With this PR, Model gets a new compute method which realizes the computation of the model's solution and of derived quantities such as outputs or error estimates.
solve, output and estimate_error are implemented by calling compute with appropriate arguments.
By relying on a central method, we can avoid having to pass computed values from the outside, e.g. the solution to estimate_error. (Note that it might depend on the Model which values have to be passed.)
compute returns a dict of all computed values. solve, etc. remain as convenience methods.
compute can be easily extended to return further quantities of interest. E.g. an optional stages argument might indicate to return intermediate Newton stages as an additional item of the return value.
Model.compute has a default implementation which relays individual computational tasks to dedicated _compute_solution, _compute_output, etc. methods which allow to more easily customize compute when subclassing. Before these methods are called _compute is called in which, e.g., simultaneous solution and output computation can be realized. For values that already have been computed, the respective _compute_... methods will then be skipped.
The signature of compute should stay stable for the foreseeable future. The base class implementation may change in the future.
I decided to keep automatic caching of the solution in the default implementation of compute. In particular, for models using _compute_solution, a call to m.output(mu) will not trigger an additional solution of the model, when m.solve(mu) has been called before for the same mu and caching is enabled.
@pymor/pymor-devs, @HenKlei, @TiKeil, while not much has changed internally, this PR will significantly affect the user expercience. Please take a close look.
Since the parameter.parse call now happens in compute, the part in the tutorial (lines 161 to 169 in the original tutorial) where parameters.parse is missing should be adjusted.
Since the parameter.parse call now happens in compute, the part in the tutorial (lines 161 to 169 in the original tutorial) where parameters.parse is missing should be adjusted.
Since the parameter.parse call now happens in compute, the part in the tutorial (lines 161 to 169 in the original tutorial) where parameters.parse is missing should be adjusted.
Done.
I meant this part:
That did not work too well! In pyMOR, all parametric objects expect the
`mu` argument to be an instance of the :class:`~pymor.parameters.base.Mu`
class. :meth:`~pymor.models.interface.Model.solve` is an exception: for
convenience, it accepts as a `mu` argument anything that can be converted
to a :class:`~pymor.parameters.base.Mu` instance using the
:meth:`~pymor.parameters.base.Parameters.parse` method of the
:class:`~pymor.parameters.base.Parameters` class. In fact, if you look
back at the implementation of :meth:`~pymor.models.interface.Model.solve`,
you see the explicit call to :meth:`~pymor.parameters.base.Parameters.parse`.
Since the parameter.parse call now happens in compute, the part in the tutorial (lines 161 to 169 in the original tutorial) where parameters.parse is missing should be adjusted.
Done.
I meant this part:
That did not work too well! In pyMOR, all parametric objects expect the
`mu` argument to be an instance of the :class:`~pymor.parameters.base.Mu`
class. :meth:`~pymor.models.interface.Model.solve` is an exception: for
convenience, it accepts as a `mu` argument anything that can be converted
to a :class:`~pymor.parameters.base.Mu` instance using the
:meth:`~pymor.parameters.base.Parameters.parse` method of the
:class:`~pymor.parameters.base.Parameters` class. In fact, if you look
back at the implementation of :meth:`~pymor.models.interface.Model.solve`,
you see the explicit call to :meth:`~pymor.parameters.base.Parameters.parse`.
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.
With this PR,
Modelgets a newcomputemethod which realizes the computation of the model's solution and of derived quantities such as outputs or error estimates.solve,outputandestimate_errorare implemented by callingcomputewith appropriate arguments.By relying on a central method, we can avoid having to pass computed values from the outside, e.g. the solution to
estimate_error. (Note that it might depend on theModelwhich values have to be passed.)computereturns a dict of all computed values.solve, etc. remain as convenience methods.computecan be easily extended to return further quantities of interest. E.g. an optionalstagesargument might indicate to return intermediate Newton stages as an additional item of the return value.Model.computehas a default implementation which relays individual computational tasks to dedicated_compute_solution,_compute_output, etc. methods which allow to more easily customizecomputewhen subclassing. Before these methods are called_computeis called in which, e.g., simultaneous solution and output computation can be realized. For values that already have been computed, the respective_compute_...methods will then be skipped.The signature of
computeshould stay stable for the foreseeable future. The base class implementation may change in the future.I decided to keep automatic caching of the solution in the default implementation of
compute. In particular, for models using_compute_solution, a call tom.output(mu)will not trigger an additional solution of the model, whenm.solve(mu)has been called before for the samemuand caching is enabled.The
solve_d_muandoutput_d_mumethods from Towards linear optimization (dual problem, sensitivity problem, output gradient) #1110 will also make use ofcompute.See discussion in #1095.
Documentation still needs to be updated.
@pymor/pymor-devs, @HenKlei, @TiKeil, while not much has changed internally, this PR will significantly affect the user expercience. Please take a close look.