How to have code output (a figure) on a separate slide

[jlperla]

Description

I am generating a set of beamer slides - with an accompanying jupyter notebook with a fair amount of code examples of plotting. Right now my pattern is the I am doing something like the following:

## Plotting $k_t$ vs. $k_{t+1}$
```{python}
#| eval: false
k_grid = np.linspace(0, 1.5, 100)
fig, ax = plt.subplots()
ax.plot(k_grid, g(k_grid), label="$g(k_t)$")
ax.plot(k_grid, k_grid, label="$k_{t+1}=k_t$")
ax.set_xlabel("$k_t$")
ax.set_ylabel("$k_{t+1}$")
ax.legend()
plt.show()
```

## Plotting $k_t$ vs. $k_{t+1}$
```{python}
#| echo: false
k_grid = np.linspace(0, 1.5, 100)
fig, ax = plt.subplots()
ax.plot(k_grid, g(k_grid), label="$g(k_t)$")
ax.plot(k_grid, k_grid, label="$k_{t+1}=k_t$")
ax.set_xlabel("$k_t$")
ax.set_ylabel("$k_{t+1}$")
ax.legend()
plt.show()
```

Is there a better way to do this that copy/pasting all of the code into two separate blocks? It is fine to me if the output always has the slide name from the previous section.... or even had no title as a worst case.

[mcanouil]

See https://quarto.org/docs/presentations/revealjs/#output-location/

[jlperla]

Thanks. Alas, this is for beamer. Is there an equivalent there? I tried just putting

## Plotting $k_t$ vs. $k_{t+1}$
```{python}
#| output-location: slide
k_grid = np.linspace(0, 1.5, 100)
fig, ax = plt.subplots()
ax.plot(k_grid, g(k_grid), label="$g(k_t)$")
ax.plot(k_grid, k_grid, label="$k_{t+1}=k_t$")
ax.set_xlabel("$k_t$")
ax.set_ylabel("$k_{t+1}$")
ax.legend()
plt.show()
```

And no luck

[mcanouil]

I had no way to know since a complete and reproducible document was not provided 🤷‍♂️
My bad, it was written beamer in the text. Note that sharing a complete document is always more useful than a small snippet of code.

Via the knitr engine, there is the concept of code cell reuse (https://bookdown.org/yihui/rmarkdown-cookbook/reuse-chunks.html) but I don't think it even exist in Jupyter/Python.

[jlperla]

Sorry about that. The full code which compiles. My current workaround:

---
title: "Test Doc"
format: beamer 
execute: 
  cache: true    
echo: true
---

## Some Code

Here is some python code

```{python}
#| eval: false
import matplotlib.pyplot as plt

plt.plot([1, 2, 3, 4])
plt.ylabel('some numbers')
plt.show()
```

## Results

```{python}
#| echo: false
import matplotlib.pyplot as plt

plt.plot([1, 2, 3, 4])
plt.ylabel('some numbers')
plt.show()
```

The revealjs with output-location:slide works great:

---
title: "Test Doc"
format: revealjs 
execute: 
  cache: true    
echo: true
---

## Some Code

Here is some python code

```{python}
#| output-location: slide
import matplotlib.pyplot as plt

plt.plot([1, 2, 3, 4])
plt.ylabel('some numbers')
plt.show()
```

BUt it is revealjs specific. Ignored if you swap that to beamer

I tried various permutaitons on the chunk stuff, but it didn't help. Also, even things like

```{python, eval=FALSE}
import matplotlib.pyplot as plt

plt.plot([1, 2, 3, 4])
plt.ylabel('some numbers')
plt.show()
```

Don't work. You need the manual #| eval: false

Not sure if there are any other possible hacks I am missing?

[mcanouil]

Don't think so or maybe there are "cell magic" in Python/Jupyter. (I am no expert in Jupyter)

[cscheid]

Don't work. You need the manual #| eval: false

This is not a "manual" declaration; that's the proper way to declare these in Quarto. (The knitr style is supported as legacy syntax, but it's not a Quarto feature)