Including code snippets in Markdown

[TikhonJelvis]

I found a package called pandoc-include-code that, well, does exactly what it says it does.

With this package, we can define named snippets in a .py file:

# start snippet iterate
from typing import Callable, Iterator, TypeVar
X = TypeVar('X')


def iterate(step: Callable[[X], X], start: X) -> Iterator[X]:
    state = start

    while True:
        yield state
        state = step(state)
# end snippet iterate

then include them in .md files with the following syntax:

...an arbitrary domain $\mathcal{D}$.

```{.python include=./chapter4/__init__.py snippet=iterate}
```

The above function take...

This way, we can keep all the snippets for a chapter in an executable (and testable!) Python file, with updates to the code files automatically reflected in the generated PDF.

I did this to a couple of snippets in chapter4.py as an example. I'm not sure this is the best way to organize the code and tests, but it seems like a good start. We can always iterate on this as we go along, and I would be comfortable forking and editing the pandoc-include-code package if we want different or more complicated behavior.

We could even include snippets from the rl/*.py files directly, but I'm not sure that's a good idea—the code might need extra details and complexity over time that isn't needed in the book itself.

[TikhonJelvis]

Once this is merged, it should work from any new Nix shells you open for this project. The change depends on #60 (which I'm going to merge shortly), so the first time you do it, the Nix shell might take a while to start while it's downloading and installing packages in the background.

I haven't tested this on macOS, so there might be some bugs on your computer!

[TikhonJelvis]

Fixed a typo here. It worked fine on your computer because macOS doesn't care about cases in filenames, but Linux is case sensitive so it caused an error for me.