[IO] Enhance IO class by adding support for reading master-slave constraints and geometries from sub-model parts (transition to extend Metis geometries/constraint support))

[loumalouomega]

📝 Description

1. Code Cleanup and Comment Improvements:

   • Fixed minor formatting and consistency issues in the comments, particularly for license information and the PartitioningInfo struct.

2. Expanded PartitioningInfo Struct:

   • Added new fields to the PartitioningInfo struct for handling additional entity types:
     • ConstraintsPartitions and GeometriesPartitions: New fields to track partitions where master-slave constraints and geometries are local, respectively.
     • ConstraintsAllPartitions and GeometriesAllPartitions: New fields to track partitions where master-slave constraints and geometries are present, including ghost partitions.

3. Added Virtual Methods for Sub-Model Part Reading:

   • Introduced several new virtual methods for reading data from sub-model parts:
     • ReadSubModelPartElementsAndConditionsIds: The original method reads element and condition IDs. The newly added overload reads additional IDs for master-slave constraints and geometries.
     • ReadNodalGraphFromEntitiesList: Similar to the previous method, this one reads a nodal graph, and the new overload includes support for master-slave constraints and geometries.
   • Both new methods include a warning message, indicating that while the methods for master-slave constraints and geometries are not yet implemented, they extend the original functionality.

🆕 Changelog

• Improve comments and add new partitioning information in PartitioningInfo struct
• Add methods to read master-slave constraints and geometries from sub-model parts

[loumalouomega]

@RiccardoRossi you can review this first, is a very small case

[RiccardoRossi]

i am not into this ... is @pooyan-dadvand iterating this with you?

[loumalouomega]

i am not into this ... is @pooyan-dadvand iterating this with you?

Yes

[pooyan-dadvand]

I have two general comments on this PR:

1. Do we need all the methods? Please don't add any method just for coherency. To my understanding we need a method to read the connectivity graph.
2. Would you please change all the MasterSlaveConstraints with Constraints?

[loumalouomega]

I have two general comments on this PR:

1. Do we need all the methods? Please don't add any method just for coherency. To my understanding we need a method to read the connectivity graph.

You mean to focus only in the extrictely necessary?

2. Would you please change all the MasterSlaveConstraints with Constraints?

I can some, but other not beacause we already defined in the io.h some methods with that name. I can rename them too if you agree.

[loumalouomega]

I have two general comments on this PR:

1. Do we need all the methods? Please don't add any method just for coherency. To my understanding we need a method to read the connectivity graph.

You mean to focus only in the extrictely necessary?

Yes methods are used.

[pooyan-dadvand]

I understand that this is the substitution of the above method. Then we shall change the above one.....

[loumalouomega]

I keep it the old one because we are just changing this file in this PR, in other case should be removed in the final one

[pooyan-dadvand]

This one again. I definitely suggest you to provide a list of usage in Kratos to see how changing the interface would affect us.

[loumalouomega]

This is used in the final PR:#13435
See applications/MetisApplication/custom_processes/metis_divide_submodelparts_heterogeneous_input_process.cpp

[pooyan-dadvand]

The interface is not tidy at all but it is a continuation of the existing one. For this reason I approve it for now and in the future we can change all of them coherently.