Enhance GEDCOM parsing and update dependencies

[jadsongmatos]

• Improved the __get_indi method in Gedcom class with detailed comments for better understanding of individual parsing.
• Added command-line options for resuming GEDCOM downloads with specified start and end levels.
• Updated requests and fake-useragent dependencies in requirements files.

[Linekio]

Hey, the resume feature seems interesting, but I think you've broken a little too much code.
On the other hand, there are too many comments, and they should be in English. Remember that comments should explain the “why” and not the “what.”
requests_cache seems unnecessary since getmyancestors never makes the same request twice.

[jadsongmatos]

requests_cache is used when I'm creating a second file where it may have relatives in common, it will be in the cache, sorry for the low quality of the code, I haven't finished what I wanted yet, but I thought it would be cool to share how it's going

[gamesguru]

I do not think the contribution is low quality but may need cleaning up.

The request cache in principle is fine. It's good.

Some other changes are maybe unnecessary or risky, but we can evaluate those case-by-case.

We ideally need to add a basic test suite to protect against breaking changes. @Linekio thoughts?

I agree about the comments. We should remove some and use English where possible.

[gamesguru]

requests_cache seems unnecessary since getmyancestors never makes the same request twice.

unless you want to build your tree again, using a larger value for -a or -d. Then it does.

I'm working on aggregating his work and a few other PRs together, should be ready soon.

I decided cache-control and conditional requests were the best middle ground. Hope to offer alternative drivers, like csv/json or maybe postgres if there's interest, but I think SQLite works for now.