Using Faker for anonymizing data.

[nautics889]

At first, thanks to @gventuri for merging #826 so soon.
These patch contained some refactoring of Anonymizer class, and when I was working on it, i've found it quite strange that we're generating mocked data ourself (like e-mails, card numbers, etc). I mean, instead we can use Faker.
Moreover, we already have Faker in the project dependencies list (and not development one, btw).

1. Shouldn't we use Faker in Anonymizer, for generating anonymous data for PandasAI requests?
2. Should we generate some use Faker for our tests to create fixtures? (I bet we should)

---

P. S. decided to create a discussion topic rather than an issue, I feel it more like suggestion, not a feature request

[gventuri]

Hi @nautics889, thanks a lot for the contributions :)

1. Yeah, I think we weren't using faker yet when we wrote the anonymizer, but I'd totally agree we should use faker instead!
2. Agreed on this point as well!