Update generated classes

[MarAlder]

This PR implements recent updates to CPACSGen. Essentially, this allows for cleaner querying of CPACS elements defined as simpleType embedded in complexType (e.g., wallSegment/phi).

Description

This PR is required because the commit ID (a1cadb8) of the linked git-submodule for CPACSGen and the existing cpacs_gen_input do not reproduce the generated C++ classes. This is because XSD restrictions minInclusive and maxInclusive were not yet implemented.

The idea behind this PR is therefore to update the git-submodule-link to the latest developments of CPACSGen, generate the corresponding classes and adapt TiGL accordingly. This should make it easier for developers to add classes for the implementation of new CPACS features.

1. Updated the CPACSGen submodule from a1cadb8 to 6f4fc1b
2. Generated the C++ classes based on the existing cpacs_gen_input definitions
3. Fixed custom classes to be consistent with the generated classes

How Has This Been Tested?

The impemented tests were carried out. One test fails (WingSegmentGuideCurves.tiglWingGetSegmentUpperSurfaceAreaTrimmed), but this is not related to the current changes.

Screenshots, that help to understand the changes(if applicable):

N/A

Checklist:

• A test for the new functionality was added.
• All tests run without failure. (see remark above)
• The new code complies with the TiGL style guide.
• New classes have been added to the Python interface.
• API changes were documented properly in tigl.h.

[joergbrech]

Thanks @MarAlder! This is great.

One test fails (WingSegmentGuideCurves.tiglWingGetSegmentUpperSurfaceAreaTrimmed), but this is not related to the current changes.

But this test was successful before, so it must be related to changes here, right? Let's wait and see if it fails in CI.

[codecov]

Codecov Report

Merging #980 (dad59d4) into master (45a9428) will not change coverage.
Report is 3 commits behind head on master.
The diff coverage is 80.00%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #980   +/-   ##
=======================================
  Coverage   68.90%   68.90%           
=======================================
  Files         298      298           
  Lines       26479    26479           
=======================================
  Hits        18246    18246           
  Misses       8233     8233           

Files	Coverage Δ
src/CCPACSStringVector.cpp	86.79% <100.00%> (ø)
src/cpacs_other/CCPACSExternalObject.cpp	55.07% <100.00%> (ø)
.../structural_elements/CCPACSFuselageWallSegment.cpp	75.83% <50.00%> (ø)

... and 1 file with indirect coverage changes

[joergbrech]

All tests succeed on Linux, so I am guessing the failing test has something to do with your local setup?

The python bindings fail to compile, because there are some explicit references to header files in src/generated, specifically CPACSWallSegment_phi.h and CPACSRectangleProfile_cornerRadius.h.

[MarAlder]

Yep, the test also fails on my local copy of master. Seems to be something with my local settings. I'll check that.

True, the python bindings need to be adjusted as well.

[rainman110]

@MarAlder I could be simply a different OCCT version, leading to slight changes in the area computation.
Another possibility is, that you don't have a patched OCCT for C2 continous coons surfaces. This will also lead to different shapes and areas.

[joergbrech]

I just had a look. Fixing the python bindings boils down to removing four lines from bindings/python_internal/configuration.i:

diff --git a/bindings/python_internal/configuration.i b/bindings/python_internal/configuration.i
index a436d12a..d7049d34 100644
--- a/bindings/python_internal/configuration.i
+++ b/bindings/python_internal/configuration.i
@@ -90,7 +90,6 @@
 #include "generated/CPACSWallPositionUIDs.h"
 #include "generated/CPACSWallPosition.h"
 #include "generated/CPACSWallPositions.h"
-#include "generated/CPACSWallSegment_phi.h"
 #include "generated/CPACSWallSegment.h"
 #include "generated/CPACSWallSegments.h"
 #include "generated/CPACSUIDSequence.h"
@@ -130,7 +129,6 @@
 %boost_optional(tigl::CCPACSWingRibsDefinitions)
 %boost_optional(tigl::CCPACSWingSpars)
 %boost_optional(tigl::generated::CPACSGuideCurve_continuity)
-%boost_optional(tigl::CCPACSRectangleProfile_cornerRadius)
 %boost_optional(tigl::CCPACSRectangleProfile)
 %boost_optional(tigl::CCPACSStandardProfile)
 %boost_optional(tigl::CCPACSWingProfileCST)
@@ -187,7 +185,6 @@
 %include "generated/CPACSWallPositionUIDs.h"
 %include "generated/CPACSWallPosition.h"
 %include "generated/CPACSWallPositions.h"
-%include "generated/CPACSWallSegment_phi.h"
 %include "generated/CPACSWallSegment.h"
 %include "generated/CPACSWallSegments.h"
 %boost_optional(tigl::generated::CPACSWalls)
@@ -405,7 +402,6 @@ class CCPACSWingRibsPositioning;
 %include "generated/CPACSCst2D.h"
 %include "ITiglWingProfileAlgo.h"
 %include "generated/CPACSPosExcl0DoubleBase.h"
-%include "generated/CPACSRectangleProfile_cornerRadius.h"
 %include "generated/CPACSRectangleProfile.h"
 %include "generated/CPACSStandardProfile.h" //TODO: Need to replace with implementation CCPACSStandardProfile.h, once it exists.
 %include "CCPACSWingProfileCST.h"

CPACSWallSegment::phi and CPACSRectangleProfile::cornerRadius are now double values and don't have an extra wrapping type.

@MarAlder, do you want to apply these changes to this PR or do you want me to take over from here?

[MarAlder]

I just had a look. Fixing the python bindings boils down to removing four lines from bindings/python_internal/configuration.i:
[...]
@MarAlder, do you want to apply these changes to this PR or do you want me to take over from here?

Ok, I suspected that too, but have not yet been able to verify it. I'll have to check again locally that I can compile the internal bindings correctly, but that's a different issue that I might get back to you with. Thanks for testing, if you could directly apply the changes that's fine for me, then we can finalize the PR.

[joergbrech]

Looks good! Thanks again!