GS: Versioning scenario should include a page about directories

[iesahin]

Currently dvc add is told in context of files only in track-a-file-or-directory.md

We need to provide directory examples and describe how they differ from files and how their members and content are tracked with examples.

[iesahin]

I removed this content from S&P scenario, this should be put elsewhere.

---

How are directories cached and tracked?

By default, DVC stores the outputs defined in stage add in the cache
similar to other data. data/prepared directory is defined as an output in
prepare stage, and the hash value is:

grep -A 1 'path: data/prepared' dvc.lock{{execute}}

and this points us to .dvc/cache/f1/b1d214c4cc7a3efdb200410227b975.dir file.

It's a JSON file that lists all individual elements of the directory with
their hash values.

cat .dvc/cache/f1/b1d214c4cc7a3efdb200410227b975.dir | jq{{execute}}

For example we see that the individual hash value of train.tsv as
fcebfd4c6f1645ac4987d39f1c5cf610 and check its content

.dvc/cache/fc/ebfd4c6f1645ac4987d39f1c5cf610{{open}}.

Note also that DVC adds /prepared to .gitignore to prevent output data
files to be committed in Git.

You can review changes in the repository from the Source Control tab of VS Code.

Finally we are registering dvc.yaml, dvc.lock and .gitignore to Git and
complete this step.

git add dvc.yaml dvc.lock data/.gitignore
git commit -m "Configured prepare stage"
```{{execute}}