saveImages flag is not documented or tested

[dsyme]

1. The saveImages flag came from FsLab's literate scripting. It is not documented or tested.

2. The docs don't explain how to embed image outputs specifically, just some general information on extending display formatting.

[gbaydin]

Just to share what I did to embed image outputs that seems to work for me:

I wrote these functions

/// Returns the file contents as Base64 encoded string
let fileToBase64String fileName =
    let bytes = System.IO.File.ReadAllBytes(fileName)
    System.Convert.ToBase64String(bytes)

/// Given a PNG image file name, returns an HTML image element with the image content included as a Base64 encoded string
let pngToHtml fileName cssStyle =
    sprintf """<img src="data:image/png;base64,%s" style="%s"/>""" (fileName |> fileToBase64String) cssStyle

Then I can do things like

// Generates an image file in the code
batch.saveImage("mnist-samples.png") 

// Embeds the generated image in the fsdocs output
(*** hide ***)
pngToHtml "mnist-samples.png" "width:25%;height:auto;image-rendering:pixelated"
(*** include-it-raw ***)

[github-actions]

🤖 This is an automated response from Repo Assist.

Checking the current codebase, the --saveimages flag is fully implemented in BuildCommand.fs with three modes:

Value	Behaviour
none (default)	No images are saved/downloaded
some	Images are saved for .ipynb (Pynb) and .tex (LaTeX) outputs
all	Images are saved for all output formats, including HTML and Markdown

The CLI help text (since at least v19) already describes this:

--saveimages    Save images referenced in docs (some|none|all).
                If 'some' then image links in formatted results are saved
                for latex and ipynb output docs.

So for embedding images in HTML API docs, you need --saveimages all. The implementation downloads remote images and saves them alongside the generated output.

The documentation gap identified here (2021) still appears valid — the fsdocs documentation site doesn't prominently explain this flag or show examples of how to embed images in literate scripts/API doc comments. The base64 workaround shared by @gbaydin above remains a useful alternative for cases where --saveimages all doesn't cover the specific use case (e.g. images produced dynamically during --eval).

Related: issue #928 tracks a similar request specifically for images in XML doc comments and API documentation pages.

Generated by Repo Assist

To install this workflow, run gh aw add githubnext/agentics/workflows/repo-assist.md@b87234850bf9664d198f28a02df0f937d0447295. View source at https://github.com/githubnext/agentics/tree/b87234850bf9664d198f28a02df0f937d0447295/workflows/repo-assist.md.

[dsyme]

/repo-assist fix the documentation problems

[github-actions]

Pull request created: #1065

Generated by Repo Assist

[github-actions]

🤖 This is an automated response from Repo Assist.

I've created a draft PR with documentation fixes for both issues raised here:

What's covered:

1. --saveimages flag — the commandline.md reference now documents all three values (none / some / all) and explains which output formats each affects (LaTeX + notebook for some; all formats including HTML for all).

2. Embedding image outputs from scripts — two approaches are now documented:

   • Using fsi.AddHtmlPrinter to register a custom printer for a specific chart/plot type (automatically renders via include-it).
   • Using a plain inlinePng helper with (*** include-it-raw ***) for images saved to disk during script evaluation — which is the same approach @gbaydin shared above.

The new content lives in commandline.md (new "Embedding Images" section) and evaluation.fsx (expanded "Using AddPrinter and AddHtmlPrinter" section).

Generated by Repo Assist for issue #683

To install this agentic workflow, run

gh aw add githubnext/agentics/workflows/repo-assist.md@da02dddebe71d9a937665abdcd1f5214dab852a7