[Q&A]: Will quarto html parameter 'embed-resources' include js-files needed to run webR-extension chunks when using channel type "automatic"?

[mskyttner]

What's your question?

I read here about 'embed-resources' not always being able to include .js-files. Often my goal is to have a standalone .html rendered from quarto, that can be sent as one file to recipients or web servers / buckets. To understand whether the extension works seemlessly with the 'embed-resources' flag is my goal / reason for asking the question.

I also wonder if there is a way to have regular chunks write data that webR chunks can read over http(s) ie pastebin style data access... what would be a good pattern/strategy without installing too many webR packages?

[coatless]

@mskyttner

Regarding the first question, I do not believe that we'll see the embed-resources option work for the quarto-webr as I'm not sure we can embed the WASM files. If the 'service-worker' communication channel is being used, no matter what we'll need to have at least two external files present.

For the second question, could you clarify a bit more what you mean? Is it reflecting persistent storage of input values or?

[mskyttner]

Here is some clarification for the second question: if a "build time" chunk embeds a blob of data in the HTML document using data URIs, could a "run time" webr-r chunk read this data?

Here is an illustration (.qmd and .html in a zip) and more attempts at clarification.

webr_read_embedded.zip

The .qmd in a gist: https://gist.github.com/mskyttner/2b5f01180a91f1cfd3117ed19497c332

Screenshot:

[coatless]

For the equation render issue, this is because the document YAML uses the embed-resources: true key. This has nothing to do with quarto-webr.

Regarding including embedded document data, the webr-r code cell only handles R code. You would have to write some custom JS inside of a raw HTML code cell to load the exported document data into webR's virtual file system (VFS) system, e.g. webR.FS.writeFile() & webR.FS.readFile(). From there, it could be loaded using regular R code. Or, you could load the data directly into the global environment while placing it in the VFS system.

[coatless]

I'm going to close out the question as being answered. If you end up writing the raw code cell, feel free to let me know!

I would suggest at this point creating a repository that contains the data sets, c.f.

https://github.com/coatless/raw-data

Enable the "GitHub Pages" option with default settings so that each data set is available at a URL like:

https://<username>.github.io/<repository>/<dataset>.csv

Then, using the read.csv() inside of a {webr-r} code cell to read in the data, e.g.

```{webr-r}
df = read.csv("https://coatless.github.io/raw-data/pima.csv")
```

[coatless]

With #168, you can use embed-resources if you are not using the service-worker communication channel.