fix(ci): include mermaid diagrams as png and svg (technical)

For a while, we have suffered from random timeouts when rendering the book.
The log-output from GHA during the "Render book" stage just ended with:

|.............................................| 100% output file: advanced_technical_aspects_of_mlr3.knit.md

Error: The operation was canceled.

after hitting the maximum runtime of 6 hours.
(the success rate was around 50/50).

When rendering with the --execute-debug flag
more log-output was given.
The log output at the end of rendering the technical chapter (pdf) was:

|
|.............................................| 100%

output file: advanced_technical_aspects_of_mlr3.knit.md

[knitr engine]: writing results
[knitr engine]: exiting
Error: The operation was canceled.

for the other chapters when rendering to pdf, the output was

|............................................| 100%

output file: preprocessing.knit.md

[knitr engine]: writing results
[knitr engine]: exiting
[knitr engine]: postprocess
[knitr engine]: writing results
[knitr engine]: exiting

--> something with the postprocessing went wrong and the bug was identified to be in the technical chapter

The problem was NOT the large-scale benchmarking chapter

quarto-dev/quarto-actions#45 (comment)

.github/workflows/book-weekly.yml

@@ -57,4 +57,4 @@ jobs:
           restore-keys: ${{ runner.os }}-openml-
 
       - name: Render book
-        run: quarto render book/ --cache-refresh
+        run: quarto render book/ --cache-refresh --execute-debug

.github/workflows/build-book.yml

@@ -56,7 +56,7 @@ jobs:
           restore-keys: ${{ runner.os }}-openml-
 
       - name: Render book
-        run: quarto render book/ --cache-refresh
+        run: quarto render book/ --cache-refresh --execute-debug
 
       - name: Add .lycheeignore
         run: cp .lycheeignore book/_book/.lycheeignore

book/chapters/chapter10/advanced_technical_aspects_of_mlr3.qmd

@@ -372,66 +372,77 @@ In contrast, if we parallelize the inner benchmark (@fig-parallel-inner) then th
 future::plan(list("sequential", "multisession"))
 ```
 
-```{mermaid}
-%%| fig-width: 5.5
-%%| fig-responsive: false
-%%| label: fig-parallel-outer
-%%| fig-cap: CPU utilization for four CPUs while parallelizing the outer five-fold CV with a sequential two-fold CV inside. Jobs are labeled as [iteration outer]-[iteration inner].
-gantt
-    title CPU Utilization
-    dateFormat  s
-    axisFormat %S
-    section C1
-    Iteration 1-1           :done, 0, 4s
-    Iteration 1-2           :done, 4, 4s
-    Iteration 5-1           :done, 8, 4s
-    Iteration 5-2           :done, 12, 4s
-
-    section C2
-    Iteration 2-1           :done, 0, 4s
-    Iteration 2-2           :done, 4, 4s
-    Idle                    :8, 8s
-
-    section C3
-    Iteration 3-1           :done, 0, 4s
-    Iteration 3-2           :done, 4, 4s
-    Idle                    :8, 8s
-
-    section C4
-    Iteration 4-1           :done, 0, 4s
-    Iteration 4-2           :done, 4, 4s
-    Idle                    :8, 8s
-```
-
-```{mermaid}
-%%| fig-width: 5.5
-%%| fig-responsive: false
-%%| label: fig-parallel-inner
-%%| fig-cap: CPU utilization for four cores while parallelizing the inner benchmarking (consisting of two holdout evaluations) with a sequential five-fold CV outside. Jobs are labeled as [iteration outer]-[iteration inner].
-
-gantt
-    title CPU Utilization
-    dateFormat  s
-    axisFormat %S
-    section C1
-    Iteration 1-1           :done, 0, 4s
-    Iteration 2-1           :done, 4, 4s
-    Iteration 3-1           :done, 8, 4s
-    Iteration 4-1           :done, 12, 4s
-    Iteration 5-1           :done, 16, 4s
-
-    section C2
-    Iteration 1-2           :done, 0, 4s
-    Iteration 2-2           :done, 4, 4s
-    Iteration 3-2           :done, 8, 4s
-    Iteration 4-2           :done, 12, 4s
-    Iteration 5-2           :done, 16, 4s
-
-    section C3
-    Idle                    :0, 20s
-
-    section C4
-    Idle                    :0, 20s
+```{r, cpu-utilization-1}
+#| echo: false
+#| fig-width: 5.5
+#| label: fig-parallel-outer
+#| fig-cap: CPU utilization for four CPUs while parallelizing the outer five-fold CV with a sequential two-fold CV inside. Jobs are labeled as [iteration outer]-[iteration inner].
+include_multi_graphics("cpu_utilization_1")
+```
+
+
+```{r, include = FALSE}
+#| fig-responsive: false
+#gantt
+#    title CPU Utilization
+#    dateFormat  s
+#    axisFormat %S
+#    section C1
+#    Iteration 1-1           :done, 0, 4s
+#    Iteration 1-2           :done, 4, 4s
+#    Iteration 5-1           :done, 8, 4s
+#    Iteration 5-2           :done, 12, 4s
+#
+#    section C2
+#    Iteration 2-1           :done, 0, 4s
+#    Iteration 2-2           :done, 4, 4s
+#    Idle                    :8, 8s
+#
+#    section C3
+#    Iteration 3-1           :done, 0, 4s
+#    Iteration 3-2           :done, 4, 4s
+#    Idle                    :8, 8s
+#
+#    section C4
+#    Iteration 4-1           :done, 0, 4s
+#    Iteration 4-2           :done, 4, 4s
+#    Idle                    :8, 8s
+```
+
+
+```{r, cpu-utilization-2}
+#| echo: false
+#| fig-width: 5.5
+#| label: fig-parallel-inner
+#| fig-cap: CPU utilization for four cores while parallelizing the inner benchmarking (consisting of two holdout evaluations) with a sequential five-fold CV outside. Jobs are labeled as [iteration outer]-[iteration inner].
+include_multi_graphics("cpu_utilization_2")
+```
+
+```{r, include = FALSE}
+#| fig-responsive: false
+#gantt
+#    title CPU Utilization
+#    dateFormat  s
+#    axisFormat %S
+#    section C1
+#    Iteration 1-1           :done, 0, 4s
+#    Iteration 2-1           :done, 4, 4s
+#    Iteration 3-1           :done, 8, 4s
+#    Iteration 4-1           :done, 12, 4s
+#    Iteration 5-1           :done, 16, 4s
+#
+#    section C2
+#    Iteration 1-2           :done, 0, 4s
+#    Iteration 2-2           :done, 4, 4s
+#    Iteration 3-2           :done, 8, 4s
+#    Iteration 4-2           :done, 12, 4s
+#    Iteration 5-2           :done, 16, 4s
+#
+#    section C3
+#    Idle                    :0, 20s
+#
+#    section C4
+#    Idle                    :0, 20s
 ```
 
 In this example, both possibilities for parallelization are not exploiting the full potential of the four cores.