Remove similar URLs, particularly integer checks and images

[ameenmaali]

Add more deduplication logic with -s switch for similar URLs. Issue at #6

[larskraemer]

Two small suggestions, both for get_path_components:

• You're erasing the /'s of the path completely. This will make test.com/a/b and test.com/ab the same URL in this mode. Maybe just keep the /'s in to avoid this.
• For performance reasons, you might want to make url_path and token both string_views and use a strategy similar to the one in parse to cut up the path. Changing a string_view is much faster than erase-ing from a string

[ameenmaali]

Nice catch @larskraemer! - I addressed the first one. I'll merge this in, and rework it a bit to use string_view in the next couple days.