added vale to CLI

[sanika-n]

Summary

I added Vale, a prose linter to CLI as requested in "Add prose linter to CI #1572 " issue. So, each time a user pushes/ creates a PR, the CLI automatically runs, and its suggestions and warnings can be viewed. Currently, I have configured it to only check on .md files, but that can be changed according to use case by altering the " .vale.ini "file.

Screenshots

Currently after running on the existing repo, vale made a lot of recommendation and suggestions. I intentionally created a "bad" .md file titled TESTVALE.md just to test it, the results are attached below:


It gave some errors and warnings on the existing files as well. I have attached a log to that below.
4_Run Vale.txt

Closes #1572.

[EwoutH]

Thanks for the PR! I appreciate

It's currently over a thousands lines of code spread over 45 files, which is quite challenging to review. I also prefer not to add such amounts of code to our codebase. Are all the files critical for this functionality to function? Anything you might be able to cleanup?

Edit: Going quickly though the log, the warning Prefer 'personal digital assistant' over 'agent'. or (Agent) appears a lot. This is probably an LLM artifact.

Also, could Vale be integrated into our pre-commit config?

Edit 2: It seems Vale supports pre-commit!

• feat: Add pre-commit support errata-ai/vale#558

[sanika-n]

@EwoutH, I updated the version of vale and also suppressed vale to allow the use of "agent". I had actually cloned code from this
repo to -> https://github.com/errata-ai/Microsoft/tree/master/Microsoft to serve as the style for vale and hence the large amount of files, I tried to find a work around, but the only option seems to be to use it as a submodule...
I did integrate vale to pre-commit and it does fail as not all md files are following vale's rules.

However, the reason for failure in the checks above, (pre-commit.ci -pr) says that it failed because it wasn't able to find vale,
I am not able to figure out why that is happening

[EwoutH]

Maybe it’s easier to start with a minimal implementation, using only pre-commit, and then extend from there.

[sanika-n]

@EwoutH, Hi, I think I figured it out it is working now:)

[quaquel]

With 45 files changed, this PR has become impossible to review. I see a lot of value in adding this, however, would it be possible to have only this one change? Next, in follow up pr's we can go throught the code base to fix all issues that are flagged after vale has been activated.

[sanika-n]

Sure, so from what I understand you would like me to cut down on the number of files right? I think one thing that I could is change the style from Microsoft to Write-Good (https://github.com/errata-ai/write-good/tree/master/write-good) this only has 9 yml files...Would that be fine?

[quaquel]

My bad, I had not carefully checked the PR. It is not 45 files changed, but 1 file changed (pre-commit) and 44 new files required for vale. Can you briefly explain the options and why you went with the Microsoft style?

[sanika-n]

Sure, so officially these are the styles that vale recognizes:

(https://github.com/errata-ai/packages)

I chose Microsoft because it has a wider range of error catching abilities and can focus on a lot more writing elements compared to some other options. Write-good is alright too but it is more basic and won't be able to catch errors that Microsoft can.