linguist is skipping over files

[kcantrel]

I have a GitHub repo (currently private) that has lots of files from various languages. I noticed within GitHub it is only mentioning HCL. I opened a support case with GitHub and they suggested installing github-linguist locally, cloning the repo and running linguist manually. Well, I did all that, and sure enough it gets the same results:

$ github-linguist -b
100.00% 17542      HCL

HCL:
EKS/FSxN-as-PVC-for-EKS/terraform/ec2.tf
EKS/FSxN-as-PVC-for-EKS/terraform/eks-cluster.tf
EKS/FSxN-as-PVC-for-EKS/terraform/fsx.tf
EKS/FSxN-as-PVC-for-EKS/terraform/kubernetes.tf
EKS/FSxN-as-PVC-for-EKS/terraform/outputs.tf
EKS/FSxN-as-PVC-for-EKS/terraform/security-groups.tf
EKS/FSxN-as-PVC-for-EKS/terraform/variables.tf
EKS/FSxN-as-PVC-for-EKS/terraform/versions.tf
EKS/FSxN-as-PVC-for-EKS/terraform/vpc.tf

This issue is, there are a lot more than those 9 files in that repo:

$ find . -type f | wc -l
469

And, of course, with various languages. For example:

$ find . -type f -name '*.py' | wc -l
7
$ find . -type f -name '*.ts' | wc -l
36
$ find . -type f -name '*.scss' | wc -l
45

There is no .gitattributes file.

The private repo pretty much has the same content of this repo: https://github.com/NetApp/FSx-ONTAP-samples-scripts repo. The new repo is to replace that one since I get this issue resolved. As you can tell linguist correctly identifies the various language types in that repo so I don't understand why it is struggling with the new one.

So far GitHuh support seems to be stumped. I'm curious if anyone here has any idea what could cause this? And/or any suggestions on how to diagnose it. Unfortunately, the linguist command doesn't seem to have any "debug" options.

[lildude]

It's impossible to determine why without seeing the repo. Your demo repo isn't exhibiting the problem so I can't diagnose a problem that I can't see and speculating will only drag things out and likely causes confusion.

If you add me as a collaborator on your repo, I can take a look.

My current suspicion is your directory structure isn't the same as your demo repo and most of your files are in a directory linguist excludes because it considers it documentation or vendored code.

[kcantrel]

I totally agree. I added you. Let me know if you get kicked off. Thanks for looking at the issue.

[kcantrel]

Oh, and reading your third sentence, we did move a lot of items under a directory named "Samples". Looks like that will match the last rule of the documentation excludes. I guess I need to find a way to override that. Any pointers will be appreciated. Thanks!

[kcantrel]

So, I guess I will need some help with what to put into a .gitattributes file so that linguist will still process files in a Samples directory. I tried:

[Ss]amples?/ -linguist-documentation
[Ss]amples?/ linguist-detectable
[Ss]amples?/** -linguist-documentation
[Ss]amples?/** linguist-detectable

I did commit the .gitattributes file in my local repo (i.e. it didn't push it). It is in the root directory of the project. Any suggestions will be greatly appreciated. Thanks again.

[kcantrel]

Okay, I figured it out. In the .gitattribute file, you don't put the ? character after the [Ss]amples like there is in the documentation.yaml file, and you have to end with **. The following line worked:

[Ss]amples/** -linguist-documentation