Quarto Book, retaining objects?

[follhim]

Description

Hello, I'm a researcher who constantly sends files of the latest code I run to a colleague. I send them one .HTML file, with a table of contents, which is an easy way to navigate the plethora of content in the file.

However, as I continue to work on this project, I see that this HTML file is getting to messy. I want to split my .qmd file into three separate sections, one is the 1_preparation.qmd, 2_setup.qmd, and 3_analysis.qmd

The only problem is that when I render this project as a book, it will get mad at me because the 2_setup.qmd doesn't have access to any of the objects I created in 1_preparation.qmd (and likewise for 3_analysis.qmd).

Now, I know there are some workarounds to this, I could save every object I create to an .Rds file, and then load that file into each .qmd file. But that's so much work.

Does anyone know how to send a single file, HTML, with multiple .qmd files without having to save every object to the workspace?

Thank you!

In someway, this is a continuation of #431

[mcanouil]

Use the include shortcode.

Books are multi page HTML documents so I don't see how it could fit your needs here.

[cderv]

For reference, this is the doc part for includes: https://quarto.org/docs/authoring/includes.html

@follhim This was introduced in quarto as a way to split long document into several files. Mainly to be able to reuse content in several files, but you can also use it to split a single file in several

Example : main.qmd

---
title: main doc
format: html
---

{{< include _1_preparation.qmd >}}

{{< include _2_setup.qmd >}}

{{< include _3_analysis.qmd >}}

This is basically "copy pasting" the content of the .qmd inside the main doc, so they should not contain any YAML header themself.

And I am using _ prefix so that they are not considered documents to be rendered in the project (https://quarto.org/docs/projects/quarto-projects.html#render-targets)

If you are looking for book / website output, but want to use objects between all .qmd, then it requires using persistent storage in some ways. Either as you mentioned saving .Rds, or using any tool that allows that (databases, workflow tool like targets, explicit knitr caching and loading, ...)

This is because Quarto renders .qmd file R computation using a background R session which is not persistent accross renders. Each file renders has its own independent R session. So it is required to use tools and methods that would allow you to 'reuse' data between two R sessions, meaning saving and loading.

This can't be avoided - though I am thinking about making this easier for that in quarto R package maybe. (i.e. offering wrapping function to save / load data - but that exists in various packages already so IDK).

Search this discussion forum for other discussion on this topic too (like #431 as you found)