Child code-cells?

[mcanouil]

I tried different things, how to programmatically create child code cells and evaluate those child within Quarto? (My use case is a bit more complicated and related to #2088).

The following does not work for instance.
I can write a script that will produce a Quarto partial document to be included, but I'd rather avoid this two-step approach.

---
title: "Untitled"
format: html
keep-md: true
---

```{r}
#| output: asis
#| echo: false
cat(
  "```{r}",
  "#| eval: true",
  "#| echo: true",
  "1 + 1",
  "plot(1)",
  "```",
  sep = "\n"
)
```

It produces the following md file:

---
title: "Untitled"
format: html
keep-md: true
---

```{r}
#| eval: true
#| echo: true
1 + 1
plot(1)
```

while I would have liked to have:

---
title: "Untitled"
format: html
keep-md: true
---

::: {.cell}

```{.r .cell-code}
1 + 1
```

::: {.cell-output .cell-output-stdout}
```
[1] 2
```
:::

```{.r .cell-code}
plot(1)
```

::: {.cell-output-display}
![](test_files/figure-html/unnamed-chunk-2-1.png){width=672}
:::
:::

The actual use case Quarto document

---
title: "Untitled"
format: revealjs
---

```{r}
#| echo: false
#| output: asis
ggplot_code <- c(
  "library(palmerpenguins)",
  "library(ggplot2)",
  "",
  "ggplot(data = penguins) +", 
  "  aes(x = flipper_length_mm, y = body_mass_g) +",
  "  geom_point(aes(color = species, shape = species), size = 3, alpha = 0.8) +", 
  "  scale_color_manual(values = c(\"darkorange\", \"purple\", \"cyan4\")) +", 
  "  labs(",
  "    title = \"Penguin size, Palmer Station LTER\",", 
  "    subtitle = \"Flipper length and body mass for Adelie, Chinstrap and Gentoo Penguins\",", 
  "    x = \"Flipper length (mm)\",",
  "    y = \"Body mass (g)\",", 
  "    color = \"Penguin species\",",
  "    shape = \"Penguin species\"", 
  "  ) +",
  "  theme_minimal() +",
  "  theme(",
  "    legend.position = c(0.2, 0.7),", 
  "    plot.title.position = \"plot\",",
  "    plot.caption = element_text(hjust = 0, face = \"italic\"),", 
  "    plot.caption.position = \"plot\"",
  "  )"
)
offset <- 4
steps <- c(6, 7, 15, 16, 22)
for (i in steps) {
  cat(
    "\n## Slide 1\n",
    ":::: {.columns}",
    '::: {.column width="60%"}',
    "",
    "```{r}",
    "#| eval: false",
    "#| echo: true",
    paste0("#| code-line-numbers: 1-", i - offset + 1),
    paste(ggplot_code[offset:length(ggplot_code)], collapse = "\n"),
    "```",
    "",
    ":::",
    "",
    '::: {.column width="40%"}',
    "",
    "```{r}",
    "#| eval: true",
    "#| echo: false",
    "#| output: true",
    sub("\\+$", "", paste(ggplot_code[1:i], collapse = "\n")),
    "```",
    "",
    ":::",
    "::::",
    sep = "\n"
  )
}
```

[cderv]

What is exactly the aim of this ? Your full .qmd example seems a bit complicated to generated all this programmatically.
Using an include file seems the way to go with Quarto. I think #2088 is more about issues in our supported feature, or regression from recent updates. Anyway, about your direct question see below.

for R document, the knitr way can still be done. With R Markdown it can be possible using recipies from Cookbook which would work for Quarto as they are knitr supported.
Those two recipies are about programmatically create content for the document leveraging

• https://bookdown.org/yihui/rmarkdown-cookbook/child-document.html
• https://bookdown.org/yihui/rmarkdown-cookbook/knit-expand.html

Basically what is missing in your example is the computation engine processing (here knitr processing) that creates the results from your including code chunk inside code chunk.

---
title: "Untitled"
format: html
keep-md: true
---

```{r}
#| echo: false
#| output: asis
child <- knitr::knit_child(text = c(
  "```{r}",
  "#| eval: true",
  "#| echo: true",
  "1 + 1",
  "```"), quiet = TRUE
)
cat(child, sep = "\n")
```
````

Usually, using `child` chunk option with a child document is a cleaner way to go. Include shortcode in Quarto should be the equivalent of that. 

[mcanouil]

It appears flipbookr is not compatible with Reveal.js and won't be (EvaMaeRey/flipbookr#51).

[cderv]

A good project to add compatibility there ! Or in a quarto extensions 😉

[mcanouil]

Indeed, this is something really useful for workshop/tutorial/teaching, although it requires heavy static code analysis to split/evaluate properly all kind of code.

After a really quick search (I only looked at the first result): https://stackoverflow.com/questions/54987129/how-to-programmatically-create-several-new-cells-in-a-jupyter-notebook
So it is possible as well for Jupyter Notebook.

[cderv]

Good topic for a blog post 😉

[mcanouil]

Once I rebranded my website to Quarto 😅
I have materials to wrote posts about Quarto (for now I have a one post/year, I missed to make one on the 6th of May 2022 😱)