IntegrateMDHistoWorkspace

[Owen Arnold]

This is a requirement for #11572. See that ticket for some of the use cases.

This is somewhat similar to ​http://docs.mantidproject.org/nightly/algorithms/Integration-v1.html except for that we IncludePartialBins as ON by default.

Note that input workspaces should be IMDHistoWorkspaces. There are two reasons for this.

1) It makes processing a lot simpler. 2) For the use-cases in the aforementioned ticket, which demonstrates this, if the input to CutMD is an MDEventWorkspace, BinMD can do the whole job in one step, so this functionality already exists and is properly handled.

[Owen Arnold]

refs #11573. Enhance MDBoxImplicitFunction

We need to be able to determine the fractional overlap between a rectilinear box represented by a set of vertexes and the box implicit function. This is a general feature.

I've added quite a few functional tests for this, but have not done anything about performance yet.

Changeset: 97fa168083106c2053ab48f0771dceb18a55838a

[Owen Arnold]

refs #11573. One failing test.

New algorithm added Some tests implemented MDBoxImplicitFunction updated and tested MDIterator updated to give bounding box extents

Last change should be key to fixing current failing test.

Changeset: cd8c18f10ca81ac379dc767085fda5b9d2d5de6d

[Owen Arnold]

refs #11573. Fix failing test.

Fix the failing test by using the extents rather than trying to calculate extents from vertexes. This should be faster too.

Added a performance test. On my machine this is running in approx 5.5 seconds.

Changeset: 9cfd2bd82737db2f1ed3e22ba08e82ef3ed3c719

[Owen Arnold]

refs #11573. Reduce from O(N^{2) to more like a O(N) problem}

Reduce from O(N^{2) problem to something way more like a O(N) problem. Results in large speed improvement. Have not yet used OMP. That will be next.}

Changeset: cdc7744992764ff2d59f289daee86a05c5d8fa33

[Owen Arnold]

refs #11573. OMP loop over outer iterator

Add progress reporting

Add omp loop over the output structure of iterators. This is actually a nice parallel problem.

Changeset: eb322badad4a3e9cbbe1b9743165169f42a2c494

[Owen Arnold]

refs #11573. Docs and doctest

Changeset: 2cee28130b8d7df40d3c85423b0907fe293722d4

[Owen Arnold]

This is being verified as ​pull request #618.

[Owen Arnold]

refs #11573. Fix compilation issue

Changeset: 0101f10ab7699966901e208c72ee302282f71e10

[Owen Arnold]

refs #11573. Fix doxygen warning.

Changeset: 1b3e1b160dd389e1d8ed08aca7177cec49b4bc00

[Owen Arnold]

refs #11573. Fix doxygen warning.

Changeset: 51aced26e65b168c4e5836aefdef0826eb8eae7d

[Owen Arnold]

refs #11573. Fix doxygen warning.

Changeset: 66c3958121907765147b91f12c39f7b5df73bbec

[Owen Arnold]

refs #11573. Fix merge conflict. merge master.

Merge branch 'master' of github.com:mantidproject/mantid into 11573_integration

Conflicts:

Code/Mantid/Vates/VatesSimpleGui/ViewWidgets/inc/MantidVatesSimpleGuiViewWidgets/BackgroundRgbProvider.h

Changeset: 3d32d2c79cdf3d2185d3b35666533376ba7d8476

[Owen Arnold]

refs #11573. Property removed. Update unittests.

Changeset: 4a8f38661b8f53d4b3f659a071df2fa7867ca065

[Owen Arnold]

refs #11573. Uninitialised variable.

Changeset: 7825817f3e86ff69d67f21f5e259b1dbc89f030f

[Federico Montesino Pouzols]

I'm compiling this (I'm afraid it's rebuillding most of Mantid and it'll take ~1 h) and my gcc gave me these two warnings: ` MDHistoWorkspaceIterator.cpp:223:45: warning: conversion to ‘Mantid::coord_t {aka float}’ from ‘size_t {aka long unsigned int}’ may alter its value [-Wconversion]

coord_t dRound = size_t(dExact + 0.5); Round to nearest bin edge.

MDHistoWorkspaceIterator.cpp:225:20: warning: conversion to ‘gnu_cxx::alloc_traits<std::allocator<long unsigned int> >::value_type {aka long unsigned int}’ from ‘Mantid::coord_t {aka float}’ may alter its value [-Wfloat-conversion]

indexes[d] = dRound;

` The '-pedantic' builds on the CI machines didn't complain apparently, so I'd say is not worth delaying this PR, but just noting it here, as there seems to be more work following on this?

[Owen Arnold]

refs #11573. Fix type warning

Changeset: 5082ee197e3ed0df9cdf0d91c3bfe044af48d34d

[Owen Arnold]

refs #11573. Fix type warnings

Changeset: 8830e5096bf6f76d72d34e9fd5e5ee0eb844fd4f

[Federico Montesino Pouzols]

All looks good and thoroughly tested and documented. I did a few manual tests as in the PR description and it behaved well.

I'm just waiting for the CI builds to finish. There's a bit of a queue for some CI platforms today, so if this is urgent just ping me and I'll just merge it, The last few changes are harmless and this had previously got the green tick.

[Owen Arnold]

refs #11573. Implies int generic list fixed.

Changeset: 500a691a6e857455450b9c46bec9a8c3f6416e09

[Federico Montesino Pouzols]

refs #11573, fix extra ';' warning picked by gcc in Ubuntu

Changeset: 1ff27b367e9e74a24888fd8c805bccd2464282cf

[Federico Montesino Pouzols]

There was a failed build on ubuntu because of a ';' warning that I took the liberty to fix as Owen is away today. And then there's a failed OSX build because of a couple of warnings coming from another, unrelated branch (SaveNXSPETest.h). Ready to merge.

[Federico Montesino Pouzols]

Merge pull request #618 from mantidproject/11573_integration

Add integration algorithm for MDHistoWorkspaces

Full changeset: ​bf13dd638090989706d901e4de3e0157a5ec2332

[Stuart Campbell]

This ticket has been transferred to github issue ​12411