add new section in HMM tutorial

[JasonPekos]

This adds a section to the HMMs tutorial that has a much faster and more concise version of the model (four lines!) using HiddenMarkovModels.jl.

There's a subsection after that which shows how to recover the state estimates with the Viterbi algorithm.

---

This closes:

• Add an implementation of forward-backward algorithm for HMM models #330

[github-actions]

Preview the changes: https://turinglang.org/docs/pr-previews/508
Please avoid using the search feature and navigation bar in PR previews!

[yebai]

Thanks @JasonPekos. It looks good!

[penelopeysm]

Suggested change

	```{julia}
	#| code-fold: true
	#| code-summary: "HMM Plotting Functions"
	p = plot(xlim=(0, 30), ylim=(-1, 5), size=(500, 250))
	```{julia}
	p = plot(xlim=(0, 30), ylim=(-1, 5), size=(500, 250))

[JasonPekos]

I really don't like how this vector — which contains no information not present in the plot, and which is not part of anyones actual workflow (in practice, people just import data via csvs etc.) — takes up so much space on the page.

that is why I originally introduced the folder code blocks (I agree that the other use especially is excessive, but I am more torn here).

I am considering doing something like

y = [i for i in vcat(fill(1.0, 6), fill(2.0, 6), fill(3.0, 7), fill(2.0, 4), fill(1.0, 7))]

but ofc this is potentially more confusing — thoughts?

[penelopeysm]

Hmm, the vector isn't this code block though – it's the one above. I do see your point though! I'm sorry if I sound too conservative here about code folding :) It's mainly because I'd like to preserve a consistent user experience across the docs. But if there are good places we can apply folding to in the docs I'm super happy to start using it.

For this particular vector, would this work maybe?

y = [fill(1.0, 6)..., fill(2.0, 6)..., fill(3.0, 7)..., fill(2.0, 4)..., fill(1.0, 7)...]

The other alternative would just be to put multiple values on one line:

y = [
    1.0, 1.0, 1.0, ...,
    ...,
]

though I recognise JuliaFormatter doesn't like that. Although we aren't explicitly checking code style in this repo, there is a page outlining the style guide so I guess we should try to practise what we preach 😄

[penelopeysm]

Could we do setprogress!(false) instead of using Quarto's output: false?

[penelopeysm]

As above – if we do setprogress!(false) above can we enable the output as usual?

[JasonPekos]

I think we could, but do we want to? I assume most users will want the progress updates in their REPL when working through the tutorial, but they just aren't a good fit for the published quarto format.

As long as we are not actually recommending users use setprogress!(false) (which I think we aren't?), it seems appropriate to just disable this in the chunk options for the rendered docs.

[penelopeysm]

In fact (I totally missed this earlier) but there is already a call to setprogress!(false) earlier in the page (which is hidden). Many other pages do call setprogress!(false) so it wouldn't be too out of place here, so I feel like it should be alright to just remove the output: false lines.

I'd be happy to add a clarification about setprogress! on the get started page in a separate PR :)

[penelopeysm]

(By the way, I'm assuming that the output you wanted to suppress was indeed the progress logging. Do correct me if that's not the case)