Refactor PiecewiseBase add coarsened tabulated linear function #14221

dschwen · 2019-10-23T18:25:49Z

I had some refactoring to do on piecewise function. I will need help fixing app fallout (Mastodon in particular). This PR adds a new Piecewise Linear function that performs a coarsening of the supplied data. (this makes sense mainly for data supplied as CSV files)

Refs #14220

moosebuild · 2019-10-23T19:12:32Z

Job Documentation on a65861b wanted to post the following:

View the site here

This comment will be updated on new commits.

permcody · 2019-10-24T14:51:34Z

Looks ok - but I bet you can fix the breakage before this goes in.

dschwen · 2019-10-24T15:18:12Z

I cannot fix pronghorn (no idea where those exodiffs come from...) or mastodon (ping @sveerara), but I fixed bison.

dschwen · 2019-10-29T13:47:15Z

Plot from the test file. Reduced from 10000 to 10 points (using a tolerance of 0.1).

dschwen · 2019-10-30T13:15:21Z

@milljm the Minimum gcc civet environment is missing the numpy python package (which is present everywhere else)

functions/coarsened_piecewise_linear.prepare_data: Working Directory: /opt/civet/build_0/moose/test/tests/functions/coarsened_piecewise_linear
functions/coarsened_piecewise_linear.prepare_data: Running command: ./generate_data.py
functions/coarsened_piecewise_linear.prepare_data: Traceback (most recent call last):
functions/coarsened_piecewise_linear.prepare_data:   File "./generate_data.py", line 3, in <module>
functions/coarsened_piecewise_linear.prepare_data:     import numpy as np
functions/coarsened_piecewise_linear.prepare_data: ImportError: No module named numpy
functions/coarsened_piecewise_linear.prepare_data: 
functions/coarsened_piecewise_linear.prepare_data: ################################################################################
functions/coarsened_piecewise_linear.prepare_data: Tester failed, reason: CODE 1
functions/coarsened_piecewise_linear.prepare_data: 
    FAIL functions/coarsened_piecewise_linear.prepare_data FAILED (CODE 1)

milljm · 2019-10-30T13:20:09Z

test/tests/functions/coarsened_piecewise_linear/tests

+  [./prepare_data]
+    type = RunCommand
+    command = './generate_data.py'
+    requirement = "Generate the fine tabulated function data for the coarsened_piecewise_linear test"


Add the following, so this test is properly skipped, if numpy is not available:

required_python_packages = 'numpy'

Does it make sense to just modify the script to not use numpy if that's an issue? Do we have systems where numpy isn't available?

I don't think there are systems where MOOSE is available, but it is impossible to get numpy. The test boxes certainly have it and my priority is to make those run as efficiently as possible.

dschwen · 2019-10-30T17:42:34Z

@sveerara, Mastodon is the only app that fails now with

/opt/civet/build_0/mastodon/src/transfers/HazardCurveTransfer.C:4:10: fatal error: Piecewise.h: No such file or directory
 #include "Piecewise.h"
          ^~~~~~~~~~~~~

I think the fix is simply to include PiecewiseBase.h now.

sveerara · 2019-10-30T17:46:46Z

@dschwen, I will add this fix now to MASTODON. Thanks!

sveerara · 2019-10-30T18:06:03Z

I have added a merged request to Mastodon (258) to be merged after this one gets merged.

dschwen · 2019-11-01T21:10:07Z

@bwspenc would you mind taking a look at this to wrap it up?

permcody · 2019-11-04T21:09:29Z

@bwspenc - ping

bwspenc

Overall, looks good. Just minor stuff. It looks like some conflicts need to be resolved too. Sorry about not getting to this sooner.

framework/doc/content/source/functions/CoarsenedPiecewiseLinear.md

framework/doc/content/source/vectorpostprocessors/PiecewiseFunctionTabulate.md

framework/include/functions/CoarsenedPiecewiseLinear.h

framework/include/utils/PointReduction.h

framework/include/functions/PiecewiseLinearBase.h

framework/include/vectorpostprocessors/PiecewiseFunctionTabulate.h

framework/src/utils/PointReduction.C

test/tests/functions/coarsened_piecewise_linear/tests

bwspenc · 2019-11-06T20:34:33Z

test/tests/functions/coarsened_piecewise_linear/tests

+  [./prepare_data]
+    type = RunCommand
+    command = './generate_data.py'
+    requirement = "Generate the fine tabulated function data for the coarsened_piecewise_linear test"


Does it make sense to just modify the script to not use numpy if that's an issue? Do we have systems where numpy isn't available?

dschwen · 2019-12-06T21:56:14Z

Mastodon fix is still needed, but Swetha mentioned that is has been ready for a month.

permcody · 2019-12-09T16:17:30Z

@dschwen - I can merge the Mastodon patch (1 liner). It has to go in after this one though.

dschwen · 2019-12-09T19:24:22Z

Thanks @permcody , the failure above is a seemingly unrelated python unittest TIMEOUT

dschwen · 2019-12-10T21:05:55Z

@bwspenc could you please re-review this?

Please re-review. I think we can reasonably request numpy to be installed in a MOOSE environment.

dschwen · 2019-12-24T17:12:33Z

♫ 🎶♪ All I want for Christmas is @bwspenc's rev-youuuuuuuu... ♫ 🎶♪

dschwen · 2019-12-24T17:29:08Z

bwspenc · 2019-12-24T21:10:19Z

Merry Christmas!

tophmatthews · 2020-09-22T16:57:43Z

framework/src/functions/PiecewiseBase.C

+                                     "All function data, supplied in abscissa, ordinate pairs");
+  params.addParam<std::vector<Real>>("x", "The abscissa values");
+  params.addParam<std::vector<Real>>("y", "The ordinate values");
+  params.addParam<FileName>("data_file", "File holding CSV data for use with Piecewise");


Aie! How did these sneak in...I created PiecewiseBase so I could make a custom linear function with inputs not like x and y, but rather ramp_rates, and hold_times, while still allowing me to pass it to IterationAdaptiveDT, which requires a PiecewiseBase. I think I need to make a PiecewiseBaseBase now :(

It looks like PiecewiseBase is only used in PiecewiseConstant and PiecewiseLinearBase, at least in MOOSE. Maybe I can just move these input parameters into those two files...

Yeah, you took away Piecewise which gave me the flexibility I needed to not rely on input parameters to build my interpolation @dschwen!

@tophmatthews, is it maybe this PR? Was the problem you raised ever solved?

Yeah, I thought I fixed this...

Also, that isn't me, @dschwen ...sorry @topher

That's what he deserves for snatching that username away from you :-D

Right! Doesn't seem to be active either. :)

dschwen force-pushed the coarse_function_14220 branch 2 times, most recently from 9165efd to b449763 Compare October 24, 2019 03:09

moosebuild added the PR: Failed but allowed label Oct 24, 2019

dschwen force-pushed the coarse_function_14220 branch from b449763 to 13fc591 Compare October 24, 2019 20:11

moosebuild removed the PR: Failed but allowed label Oct 24, 2019

dschwen force-pushed the coarse_function_14220 branch 2 times, most recently from 70f3d96 to ab7e1b9 Compare October 29, 2019 13:18

dschwen force-pushed the coarse_function_14220 branch 3 times, most recently from 5c43315 to e2b8401 Compare October 30, 2019 12:46

milljm reviewed Oct 30, 2019

View reviewed changes

dschwen force-pushed the coarse_function_14220 branch from e2b8401 to 06b4ff3 Compare October 30, 2019 16:03

dschwen added the PR: Ready for review/merge label Oct 30, 2019

dschwen changed the title ~~Refactor PiecewiseBase~~ Refactor PiecewiseBase add coarsened tabulated linear function Oct 30, 2019

moosebuild added the PR: Failed but allowed label Oct 30, 2019

dschwen assigned permcody Oct 30, 2019

dschwen requested a review from bwspenc October 31, 2019 14:51

permcody assigned bwspenc and unassigned permcody Oct 31, 2019

bwspenc previously requested changes Nov 6, 2019

View reviewed changes

bwspenc added PR: Address comments and removed PR: Ready for review/merge labels Nov 6, 2019

dschwen added 4 commits December 6, 2019 13:07

Refactor PiecewiseBase (idaholab#14220)

e1a76f4

Factor out piecewise linear base class (idaholab#14220)

931d6e3

Add point reduction algorithm (idaholab#14220)

d3a4da7

Add test correct spelling (idaholab#14220)

edac05e

dschwen force-pushed the coarse_function_14220 branch from 06b4ff3 to edac05e Compare December 6, 2019 20:23

moosebuild removed PR: Address comments PR: Failed but allowed labels Dec 6, 2019

Address comments (idaholab#14220)

a65861b

dschwen requested a review from bwspenc December 6, 2019 21:56

moosebuild added the PR: Failed but allowed label Dec 10, 2019

bwspenc approved these changes Dec 24, 2019

View reviewed changes

bwspenc merged commit 9a91b04 into idaholab:next Dec 24, 2019

tophmatthews reviewed Sep 22, 2020

View reviewed changes

dschwen mentioned this pull request Jan 31, 2022

Allow for multiple timestep limiting functions #17653

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Refactor PiecewiseBase add coarsened tabulated linear function #14221

Refactor PiecewiseBase add coarsened tabulated linear function #14221

dschwen commented Oct 23, 2019 •

edited

moosebuild commented Oct 23, 2019 •

edited

permcody commented Oct 24, 2019

dschwen commented Oct 24, 2019

dschwen commented Oct 29, 2019

dschwen commented Oct 30, 2019

milljm Oct 30, 2019

bwspenc Nov 6, 2019

dschwen Dec 13, 2019 •

edited

dschwen commented Oct 30, 2019

sveerara commented Oct 30, 2019

sveerara commented Oct 30, 2019

dschwen commented Nov 1, 2019

permcody commented Nov 4, 2019

bwspenc left a comment

bwspenc Nov 6, 2019

dschwen commented Dec 6, 2019

permcody commented Dec 9, 2019

dschwen commented Dec 9, 2019

dschwen commented Dec 10, 2019

dschwen commented Dec 24, 2019

dschwen commented Dec 24, 2019

bwspenc commented Dec 24, 2019

tophmatthews Sep 22, 2020

tophmatthews Sep 22, 2020

dschwen Jan 31, 2022 •

edited

tophmatthews Jan 31, 2022

tophmatthews Jan 31, 2022

dschwen Jan 31, 2022

tophmatthews Jan 31, 2022

Refactor PiecewiseBase add coarsened tabulated linear function #14221

Refactor PiecewiseBase add coarsened tabulated linear function #14221

Conversation

dschwen commented Oct 23, 2019 • edited

moosebuild commented Oct 23, 2019 • edited

permcody commented Oct 24, 2019

dschwen commented Oct 24, 2019

dschwen commented Oct 29, 2019

dschwen commented Oct 30, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dschwen Dec 13, 2019 • edited

Choose a reason for hiding this comment

dschwen commented Oct 30, 2019

sveerara commented Oct 30, 2019

sveerara commented Oct 30, 2019

dschwen commented Nov 1, 2019

permcody commented Nov 4, 2019

bwspenc left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dschwen commented Dec 6, 2019

permcody commented Dec 9, 2019

dschwen commented Dec 9, 2019

dschwen commented Dec 10, 2019

dschwen commented Dec 24, 2019

dschwen commented Dec 24, 2019

bwspenc commented Dec 24, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dschwen Jan 31, 2022 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dschwen commented Oct 23, 2019 •

edited

moosebuild commented Oct 23, 2019 •

edited

dschwen Dec 13, 2019 •

edited

dschwen Jan 31, 2022 •

edited