`engine: julia` errors

[dmbates]

Bug description

I installed pre-release 1.5.29 to try the engine: julia facility on a Quarto-based book with Julia code chunks at https://github.com/JuliaMixedModels/EmbraceUncertainty . See the db/juliarunner branch of the repository. I consistently get errors like

Running [37/40] at line 742:  dsm01trace = fit(
Running [38/40] at line 779:  only(dsm01.b)
Running [39/40] at line 803:  caterpillar!(
Running [40/40] at line 818:  qqcaterpillar!(
ERROR: TypeError: output.text.map is not a function

Stack trace:
    at mdOutputStream (file:///Applications/quarto/bin/quarto.js:32951:37)
    at mdFromCodeCell (file:///Applications/quarto/bin/quarto.js:32858:29)
    at jupyterToMarkdown (file:///Applications/quarto/bin/quarto.js:32287:34)
    at eventLoopTick (ext:core/01_core.js:153:7)
    at async Object.execute (file:///Applications/quarto/bin/quarto.js:41272:25)
    at async renderExecute (file:///Applications/quarto/bin/quarto.js:77765:27)
    at async renderFileInternal (file:///Applications/quarto/bin/quarto.js:77933:43)
    at async renderFiles (file:///Applications/quarto/bin/quarto.js:77801:17)
    at async renderProject (file:///Applications/quarto/bin/quarto.js:78187:25)
    at async Command.actionHandler (file:///Applications/quarto/bin/quarto.js:82777:32)

when trying to render a chapter (intro.qmd in this case).

I have been trying to create a smaller example illustrating this problem but have been unable to do so.

Steps to reproduce

No response

Expected behavior

No response

Actual behavior

No response

Your environment

• IDE: VSCode: version 1.88
• julia> versioninfo()
  Julia Version 1.10.2
  Commit bd47eca2c8a (2024-03-01 10:14 UTC)
  Build Info:
  Official https://julialang.org/ release
  Platform Info:
  OS: macOS (arm64-apple-darwin22.4.0)
  CPU: 8 × Apple M1 Pro
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-15.0.7 (ORCJIT, apple-m1)
  Threads: 6 default, 0 interactive, 3 GC (on 6 virtual cores)

(@quarto) pkg> st
Status ~/.julia/environments/quarto/Project.toml
[4c0109c6] QuartoNotebookRunner v0.9.1

Quarto check output

% quarto check
Quarto 1.5.29
[✓] Checking versions of quarto binary dependencies...
      Pandoc version 3.1.11: OK
      Dart Sass version 1.70.0: OK
      Deno version 1.41.0: OK
      Typst version 0.11.0: OK
[✓] Checking versions of quarto dependencies......OK
[✓] Checking Quarto installation......OK
      Version: 1.5.29
      Path: /Applications/quarto/bin

[✓] Checking tools....................OK
      TinyTeX: v2024.04
      Chromium: 869685

[✓] Checking LaTeX....................OK
      Using: TinyTex
      Path: /Users/dmbates/Library/TinyTeX/bin/universal-darwin
      Version: 2024

[✓] Checking basic markdown render....OK

[✓] Checking Python 3 installation....OK
      Version: 3.10.13 (Conda)
      Path: /Users/dmbates/.julia/conda/3/bin/python3
      Jupyter: 5.7.2
      Kernels: julia-1.11, julia-1.10, julia-1.9, julia-1.12, python3

[✓] Checking Jupyter engine render....OK

[✓] Checking R installation...........OK
      Version: 4.3.2
      Path: /Library/Frameworks/R.framework/Resources
      LibPaths:
        - /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/library
      knitr: 1.45
      rmarkdown: 2.26

[✓] Checking Knitr engine render......OK

[dmbates]

@jkrumbiegel It doesn't appear that the error is coming directly from the QuartoNotebookRunner.jl code but I thought I should ping you in case this has come up for others.

[cscheid]

ERROR: TypeError: output.text.map is not a function

This is a good hint, I believe. For this specific notebook, output.text is somehow not an array of strings, which is what mdOutputStream expects.

[cscheid]

The error is coming from

https://github.com/quarto-dev/quarto-cli/blob/main/src/execute/julia.ts#L164

@jkrumbiegel will be able to say more here, hopefully.

[jkrumbiegel]

It looks like the ipynb that QuartoNotebookRunner creates does not match what the quarto code expects in this case. Specifically, quarto looks at some cell output of type stream whose field output.text is not an array as Carlos mentioned above. To start debugging, I would render the notebook directly with QuartoNotebookRunner into an ipynb file like

using QuartoNotebookRunner

QuartoNotebookRunner.render(the_qmd_file; output="output.ipynb")

And then look at the json data underlying that ipynb file (you could also upload it somewhere for us), find the place where an output.text of a cell is not an array (I'd guess it's a string as null and undefined give different error messages when you call .map on them). Then you'd know which cell is the culprit, and you can reduce your script to just that cell. That should give an MWE.

[dmbates]

In the examples I see, output.text is always a string, not an array of strings, if I am looking at the right sections. A simple example is

        {
            "id": "72",
            "cell_type": "code",
            "metadata": {
            },
            "source": [
                "show(only(dsm01.β) .+ [-2, 2] * only(dsm01.stderror))"
            ],
            "outputs": [
                {
                    "output_type": "stream",
                    "name": "stdout",
                    "text": "[1492.110894544422, 1562.8891054555756]"
                }
            ],
            "execution_count": 1
        },

I do get some Unicode characters in the stdout stream when there are warnings. For example

       {
            "id": "78",
            "cell_type": "code",
            "metadata": {
            },
            "source": [
                "#| code-fold: true\n",
                "#| fig-cap: Caterpillar plot of prediction intervals for dsm01 random effects\n",
                "#| label: fig-caterpillar_dsm01\n",
                "#| warning: false\n",
                "caterpillar!(\n",
                "  Figure(resolution=(800, 250)),\n",
                "  ranefinfo(dsm01, :batch),\n",
                ")"
            ],
            "outputs": [
                {
                    "output_type": "stream",
                    "name": "stdout",
                    "text": "\u001b[33m\u001b[1m┌ \u001b[22m\u001b[39m\u001b[33m\u001b[1mWarning: \u001b[22m\u001b[39mFound `resolution` in the theme when creating a `Scene`. The `resolution` keyword for `Scene`s and `Figure`s has been deprecated. Use `Figure(; size = ...` or `Scene(; size = ...)` instead, which better reflects that this is a unitless size and not a pixel resolution. The key could also come from `set_theme!` calls or related theming functions.\n\u001b[33m\u001b[1m└ \u001b[22m\u001b[39m\u001b[90m@ Makie ~/.julia/packages/Makie/ND0gA/src/scenes.jl:220\u001b[39m\n"
                },

Would that be a problem?

[cscheid]

Looks like it's always a string: https://nbformat.readthedocs.io/en/latest/format_description.html#stream-output

[cscheid]

Maybe some formats (maybe even Quarto 😬 ) output a list of strings, but we should try to handle both cases.

@jkrumbiegel I can make a fix here, it looks simple enough. I'll ask for a quick review from you

[cscheid]

I have a PR up at #9346, but I'm a little suspicious of what's happening here, and I think it's Quarto's fault. We need to investigate thoroughly.

[jkrumbiegel]

Looks like it can be both given the schema file we're using to validate our outputs.

https://github.com/PumasAI/QuartoNotebookRunner.jl/blob/f95ee95e27de27f5f33758d05cffe2ba499e015d/test/schema/nbformat.v4.schema.json#L362

https://github.com/PumasAI/QuartoNotebookRunner.jl/blob/f95ee95e27de27f5f33758d05cffe2ba499e015d/test/schema/nbformat.v4.schema.json#L460-L465

cc @MichaelHatherly