Improving connection commands testing

[germa89]

Wrapping connection commands to upload the file
Adding tests

[germa89]

Locally

Locally (Windows 11, MAPDL v22.2) only the CADs Catia and Parasolid (x_t) fails.

CICD

In the CICD it seems a different story, maybe the docker images do not include the tools (executables) to convert those CAD formats.

v23.2 and newer

We get the following error:

========================= PyMAPDL Pytest short summary =========================
[FAILED] test_readin_sat - E                     Specified library does not exist.
[FAILED] test_readin_catiav5 - E           Ignore these messages by setting 'ignore_errors'=True
[FAILED] test_readin_x_t - E                     Specified library does not exist.

SAT files

test_readin_sat's Specified library does not exist. error is:

 raise MapdlRuntimeError(
                        f"\n\nError in instance {self.name}\n\n" + error_message
                    )
E                   ansys.mapdl.core.errors.MapdlRuntimeError: 
E                   
E                   Error in instance GRPC_127.0.0.1:21000
E                   
E                   *** ERROR ***                           CP =       0.000   TIME= 00:00:00
E                    An error was detected while operating on a shared library:              
E                     PATH = /ansys_inc/ansys/lib/linx64/libac4.so                           
E                     CMD = ~SATIN                                                           
E                     FNC = AnsysConnectCad                                                  
E                     Specified library does not exist.

Solution in the following CAD format.

Parasolid (x_t)

test_readin_x_t one is:

>                   raise MapdlRuntimeError(
                        f"\n\nError in instance {self.name}\n\n" + error_message
                    )
E                   ansys.mapdl.core.errors.MapdlRuntimeError: 
E                   
E                   Error in instance GRPC_127.0.0.1:21000
E                   
E                   *** ERROR ***                           CP =       0.000   TIME= 00:00:00
E                    An error was detected while operating on a shared library:              
E                     PATH = /ansys_inc/ansys/lib/linx64/libac4.so                           
E                     CMD = ~PARAIN                                                          
E                     FNC = AnsysConnectCad                                                  
E                     Specified library does not exist.

It seems the missing library libac4.so is only 59 kb... (on windows) so I think we should get it back in the docker images.

Catia

The error is different:

>           raise MapdlInvalidRoutineError(text)
E           ansys.mapdl.core.errors.MapdlInvalidRoutineError: *** WARNING ***                         CP =       0.000   TIME= 00:00:00
E            ~CAT5IN is not a recognized BEGIN command, abbreviation, or macro.      
E            
E           
E           Ignore these messages by setting 'ignore_errors'=True

However, that command should work as BEGIN command (it does with the full local installation).
So I think the Catia converter command is loaded dynamically (or sort of). If the library is not available, the command is not loaded. I do not know the size of the missing library in this case. Maybe is it the same library as the above?

I will need some confirmation on this @koubaa @kmkoshy @FredAns

In 23.1 and older

I could only test 23.1 but I will do also older versions later.

SAT

It is read, but probably no geometry is read in. For some reason, no error is catch. Maybe I'm only skipping volumes? Pinging @mikerife

Parasolid (x_t)

The conversion seems to fail File file.anf does not exist.:

>           raise MapdlCommandIgnoredError(text)
E           ansys.mapdl.core.errors.MapdlCommandIgnoredError: QUERY = COMMON,,IGES_TOL,,INT,5
E             
E               ANSYS Connection for Parasolid 
E             
E                 Part name:   ''/.
E                 Output file: .anf
E                 Entities:    SOLIDS
E                 File type:   ANF
E             
E            Failed to open ''/..x_t for reading
E           
E            *** WARNING ***                         CP =       0.000   TIME= 00:00:00
E            File file.anf does not exist.                                           
E             The /INPUT command is ignored.                                         
E           
E           
E            ***** ROUTINE COMPLETED *****  CP =         0.000
E           
E           
E              *****MAPDL VERIFICATION RUN ONLY*****
E                DO NOT USE RESULTS FOR PRODUCTION
E           
E                     ***** MAPDL ANALYSIS DEFINITION (PREP7) *****
E           
E           Ignore these messages by setting 'ignore_errors'=True

Maybe the configuration is not correct?? Pinging @mikerife for feedback.

Catia

Same as v23.1 and above. Command is missing.

Final decision

Do we want to include those tools to convert CADs in the MAPDL images? I guess we should.... Pinging @koubaa @dts12263 @mikerife for feedback.

[codecov]

Codecov Report

Merging #2311 (0e167e7) into main (a2d1951) will increase coverage by 0.23%.
Report is 1 commits behind head on main.
The diff coverage is 95.16%.

@@            Coverage Diff             @@
##             main    #2311      +/-   ##
==========================================
+ Coverage   81.79%   82.02%   +0.23%     
==========================================
  Files          45       45              
  Lines        8435     8464      +29     
==========================================
+ Hits         6899     6943      +44     
+ Misses       1536     1521      -15     

[koubaa]

Do we want to include those tools to convert CADs in the MAPDL images?

Maybe we can have two tiers of images, one "full" and one "minimal", and run most tests on minimal but a small set of tests on the full one. The minimal one could be much smaller (no mumps, no plugins, etc)

[germa89]

Do we want to include those tools to convert CADs in the MAPDL images?

Maybe we can have two tiers of images, one "full" and one "minimal", and run most tests on minimal but a small set of tests on the full one. The minimal one could be much smaller (no mumps, no plugins, etc)

This is interesting. Let's see what @dts12263 thinks

[germa89]

This PR has been totally defeated.... I end up introducing tests that fail consistently.

Hopefully in the future I can fix some. I will need to do changes in the docker containers.

[germa89]

LGTM.

[pyansys-ci-bot]

✅ Approving this PR because germa89 said so in here 😬