[Enhancement]: Artifacts download with correct file extension (.html instead of .csv/.tsv)

[11bluetree]

What features would you like to see added?

I've noticed a bug when downloading CSV and TSV files from artifacts in LibreChat. When clicking the download button, the file is saved with the extension .html regardless of the actual file content type. This makes it less convenient for users as they need to manually rename files after downloading.

Steps to reproduce:

• Create a CSV or TSV artifact in a conversation
• Click on the download button in the top-right corner of the artifact
• Observe that the downloaded file is named "index.html" instead of having the appropriate extension (.csv or .tsv)

This enhancement would improve user experience by eliminating the need for manual file renaming after downloading artifacts.

More details

This issue affects workflow efficiency when working with data files. Currently, after downloading a TSV or CSV file as "index.html", users need to:

1. Manually rename the file to add the correct extension
2. Sometimes have to open the file in a text editor first to confirm its format before renaming

The expected behavior would be:

1. When downloading a CSV artifact, the file should be saved with a .csv extension
2. When downloading a TSV artifact, the file should be saved with a .tsv extension
3. Other file types should similarly have appropriate extensions

A possible implementation approach might be to:

1. Extract the file type from the artifact's MIME type or title
2. Set the appropriate Content-Disposition header with the correct filename and extension when serving the download
3. Alternatively, parse the artifact title for an extension if one is provided (e.g., "sample.tsv" would download as "sample.tsv")

This enhancement would improve user experience by eliminating the need for manual file renaming after downloading artifacts.

Which components are impacted by your request?

No response

Pictures

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

[danny-avila]

This is not a bug, the artifact is technically "HTML" and why downloading it presents itself as such. Otherwise the artifact would not load at all, and there would be no way of knowing what the right extension is.

Something like this is not formally implemented yet, but is planned, where you can generate non-html text-based files like markdown:

#3995

[11bluetree]

@danny-avila thanks!

[gaieges]

and there would be no way of knowing what the right extension is.

Why do you say that? Artifacts come with a mime type, why not just use that and infer the extension?

I have a change that I've made locally and will use, but want to understand your reasoning there before submitting a PR with it @danny-avila