build: update ansys-dpf-core

[germa89]

As the title.

Related to #2953

[ansys-reviewer-bot]

Thanks for opening a Pull Request. If you want to perform a review write a comment saying:

@ansys-reviewer-bot review

[germa89]

Just testing.

[germa89]

Hopefully this has been addressed in the new version: ansys/pydpf-core#1438

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 86.99%. Comparing base (3635824) to head (05950e6).
Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3082      +/-   ##
==========================================
+ Coverage   86.94%   86.99%   +0.05%     
==========================================
  Files         187      187              
  Lines       14622    14623       +1     
==========================================
+ Hits        12713    12722       +9     
+ Misses       1909     1901       -8     

[germa89]

Still broken. @PProfizi

[PProfizi]

Hi @germa89,

Looking at the error, it seems you get the error linked to the change in operator naming for NMISC and SMISC.
I think you are already aware that sadly there was a breaking change on those. I just would like to understand why you would get this error here. What server version is this pipeline using?
I see a Docker image for 22.2 being used, isn't that a bit old?

[germa89]

Hi @germa89,

Looking at the error, it seems you get the error linked to the change in operator naming for NMISC and SMISC. I think you are already aware that sadly there was a breaking change on those. I just would like to understand why you would get this error here. What server version is this pipeline using? I see a Docker image for 22.2 being used, isn't that a bit old?

Hi @PProfizi

Yes, it is a bit old. But I didn't have the time to figure it out how to deploy a newer version of DPF in our CICD. I know the docker registry I'm using hasn't been updated in 2 years. I'm happy to get some guidance on how to use newer versions of DPF in our CICD.

[PProfizi]

Hi @germa89,
Looking at the error, it seems you get the error linked to the change in operator naming for NMISC and SMISC. I think you are already aware that sadly there was a breaking change on those. I just would like to understand why you would get this error here. What server version is this pipeline using? I see a Docker image for 22.2 being used, isn't that a bit old?

Hi @PProfizi

Yes, it is a bit old. But I didn't have the time to figure it out how to deploy a newer version of DPF in our CICD. I know the docker registry I'm using hasn't been updated in 2 years. I'm happy to get some guidance on how to use newer versions of DPF in our CICD.

Hi @germa89, do you necessarily need to go through the Docker registry?
We do have a GitHub action to install DPF, which you can see used here.

[germa89]

Hi @germa89,
Looking at the error, it seems you get the error linked to the change in operator naming for NMISC and SMISC. I think you are already aware that sadly there was a breaking change on those. I just would like to understand why you would get this error here. What server version is this pipeline using? I see a Docker image for 22.2 being used, isn't that a bit old?

Hi @PProfizi
Yes, it is a bit old. But I didn't have the time to figure it out how to deploy a newer version of DPF in our CICD. I know the docker registry I'm using hasn't been updated in 2 years. I'm happy to get some guidance on how to use newer versions of DPF in our CICD.

Hi @germa89, do you necessarily need to go through the Docker registry? We do have a GitHub action to install DPF, which you can see used here.

I do not need to go through the docker route. However, any proposed solution should work on github runners outside Ansys network. Do the action works in that way?

[PProfizi]

Hi @germa89,
Looking at the error, it seems you get the error linked to the change in operator naming for NMISC and SMISC. I think you are already aware that sadly there was a breaking change on those. I just would like to understand why you would get this error here. What server version is this pipeline using? I see a Docker image for 22.2 being used, isn't that a bit old?

Hi @PProfizi
Yes, it is a bit old. But I didn't have the time to figure it out how to deploy a newer version of DPF in our CICD. I know the docker registry I'm using hasn't been updated in 2 years. I'm happy to get some guidance on how to use newer versions of DPF in our CICD.

Hi @germa89, do you necessarily need to go through the Docker registry? We do have a GitHub action to install DPF, which you can see used here.

I do not need to go through the docker route. However, any proposed solution should work on github runners outside Ansys network. Do the action works in that way?

Yes, we use it on GitHub runners outside of the Ansys network. It just needs access via a secret to a token to our DPF server repo.

[germa89]

@PProfizi thank you a lot for this! I will start to work on this after next week.

[germa89]

Note to myself:
https://learn.microsoft.com/en-us/powershell/scripting/install/install-ubuntu?view=powershell-7.5

Installing PowerShell on Ubuntu - PowerShell

Information about installing PowerShell on Ubuntu

[germa89]

@PProfizi

I'm getting issues on this code:

dpf.core.make_tmp_dir_server(dpf.SERVER)

if dpf.SERVER.local_server:
    model = dpf.Model(rst_path)
else:
    server_file_path = dpf.upload_file_in_tmp_folder(rst_path)
    model = dpf.Model(server_file_path)

The purpose of this code is to check DPF is local or is on a container.

Can you suggest any adjustments?

[PProfizi]

@PProfizi DPF actions require to have powershell installed even on ubuntu? That's a bummer...

Hi @germa89
I think GH Ubuntu runners can run pwsh scripts by default.
We do not install it explicitly in our own pipelines, and nobody reported this as an issue before.
If you want we can try modifying the action to use something else.

[PProfizi]

@PProfizi

I'm getting issues on this code:

dpf.core.make_tmp_dir_server(dpf.SERVER)

if dpf.SERVER.local_server:
    model = dpf.Model(rst_path)
else:
    server_file_path = dpf.upload_file_in_tmp_folder(rst_path)
    model = dpf.Model(server_file_path)

The purpose of this code is to check DPF is local or is on a container.

Can you suggest any adjustments?

Hi @germa89, what issue are you getting?
This should work, this is similar to what we do here for example files.

[germa89]

@PProfizi DPF actions require to have powershell installed even on ubuntu? That's a bummer...

Hi @germa89 I think GH Ubuntu runners can run pwsh scripts by default. We do not install it explicitly in our own pipelines, and nobody reported this as an issue before. If you want we can try modifying the action to use something else.

Because we do tests in our own ubuntu docker image (we need MAPDL installed), we do not install DPF. What about a docker image I can pull and connect to?

I rather not install more stuff on the image/container. And the CICD is already in place for connecting to a DPF docker container.

[PProfizi]

@PProfizi DPF actions require to have powershell installed even on ubuntu? That's a bummer...

Hi @germa89 I think GH Ubuntu runners can run pwsh scripts by default. We do not install it explicitly in our own pipelines, and nobody reported this as an issue before. If you want we can try modifying the action to use something else.

Because we do tests in our own ubuntu docker image (we need MAPDL installed), we do not install DPF. What about a docker image I can pull and connect to?

I rather not install more stuff on the image/container. And the CICD is already in place for connecting to a DPF docker container.

Ok so either you need a docker image for DPF, which we do not ship as of now, maybe contact @ansjhajjar for more info.
Or you should still be able to use the action if you install Powershell, no? Again, we can see how to switch from Powershell to something else.

[germa89]

Ok so either you need a docker image for DPF, which we do not ship as of now, maybe contact @ansjhajjar for more info.

I think the docker image is the most convenient, less setup, less mess up.

Why did we stop updating the docker image in here https://github.com/ansys/pydpf-core/pkgs/container/dpf-core ??
I guess there is a real need for it? there are a lot of downloads (+37k). But it might be just PyMAPDL.. probably not.

Can we update that docker image? @ansjhajjar @PProfizi @cbellot000

By the way, where is that container is being built? Where is the recipe?

Or you should still be able to use the action if you install Powershell, no? Again, we can see how to switch from Powershell to something else.

I did install powershell in our containers, but still facing issues regarding the action and the script Ans.Dpf.Grpc.sh itself because we are running "in-docker" container (github runner spawn our container, and then it runs everything from there). That is why I think the docker container will be easier.

[germa89]

I think I kind of find out where the recipe is. In TFS/DPF... file dpf_generate_docker.yml.

[PProfizi]

I think I kind of find out where the recipe is. In TFS/DPF... file dpf_generate_docker.yml.

AFAIK the image and its recipe are completely outdated though.