WIP: Add spanish translation, Literate.jl backend

[miguelraz]

This PR is a WIP with at least 2 goals:

1. Add a spanish translation of the soon-to-be 1.0 DataFrames tutorials
2. Add Literate.jl based notebooks for easier source control

Rationale:

1. Spanish translation I will add as I go, and can rebase if needed.
2. Because I used a very handy ipynyb -> .jl Literate.jl converter, I already have the files that Literate.jl would use to produce the current .ipynb notebooks. It seems silly to use them for Spanish and not make them available to the rest of the tutorials (as a Portuguese translation was also suggested, this will reduce shared work.)

[bkamins]

This is very nice. I understand that this is still WIP (but having checked this it could just be merged - right?, as in the PR you just provide the .jl files that would reproduce what we have now - right?)

[storopoli]

I will fork miguelraz:spanish-tutorials as storopoli:portuguese-tutorials and start from the commit a200be1 (add literate jl files) and work out the portuguese translation. Since I am working at the University during the week I will do it during this week. So I'll probably have it done by Wednesday. Thank you!

[miguelraz]

PHEW! All done! I will ping some spanish speaking friends to look over it.

[pdeffebach]

Should we delete this? Since it will be a 1.0 compatible anyway so there's no need to talk about deprecations.

[miguelraz]

Yeah there's a couple of pre DataFrames 1.0 kinks that can be cleaned up. I'm waiting to run the notebooks and test that everything works to filter many of these out. I still don't know much about DataFrames so I don't know what is/isn't design yet.

[bkamins]

@miguelraz: please let me know when this is good to merge

[miguelraz]

I have to fix some errors on the main source files:
There's some missing exports like randstring -> Random.randstring in 01_constructors and DataFrame(rand(3)) breaking Tables.jl AbstractRow interface.

[bkamins]

OK. Thank you!

[storopoli]

Also there is some funky stuff with srand (at least in my Julia 1.6.0)

[bkamins]

Where is srand used? There is no such function AFAICT.

[storopoli]

Here is one example @bkamins, but they are everywhere...

[bkamins]

This should be Random.seed!(1234). Where did you get srand from? Here is the source (from my notebooks on master):

using Random
Random.seed!(1234)
x[:, shuffle(names(x))]

[storopoli]

We can easily replace with a regex in both portuguese and spanish. I just followed the .jl converted by @miguelraz

[bkamins]

As commented on Slack it seems that @miguelraz did not use master for doing the conversion. The fix should be easy hopefully.

[storopoli]

another one @bkamins (it shows 16 times in the files)

[miguelraz]

Yup, it's my bad @storopoli - I forked the master on github but had a stale repo locally which I made the notebooks from. I'm changing the src/*.jl files now to update them manually.

[bkamins]

@miguelraz - I am back to this PR. Is it ready to merge or there are still some issues to resolve.
I recall that there was some problem with handling of lines of code that produced an error - how do we stand with this?

[miguelraz]

Still issues. I'll get to it today and ping you.

[bkamins]

Sure - no rush. I am porting other tutorials to DataFrames.jl 1.1, so I have my pipeline filled with work :).

[bkamins]

To avoid blocking users I have updated the Jupyter Notebooks to 1.1.1 release of DataFrames.jl. Fortunately the changes are minor:

• Notebook 4, comment under cell 6 about memory mapping
• Notebook 4, instructions for JDF.jl
• Notebook 4, exclude one JSON timing in reading plot as it is very slow
• Notebook 6, added an example of subset
• Notebook 7, changes in comparing of categorical vs underlying type; and change of get to unwrap