Add filtering capabilities

[Sermus]

It would be awersome to have some filtering capabilities while gathering the stats: 1) Ignore certain files by names or extensions 2) Ignore certain paths 3) Ignore ranges of revisions

For example svnstat provided quite an elegant solution to 1 and 2 by giving an opportunity to set ignore list with apache-like wildcards. Motivation for such filtering is that some files in repository could be out of interest (autogenerated, imported 3rdParty data and so on)

[buserror]

I agree. here theres one fairly huge autogenerated XML file that more or less renders the stats useless, as that authors gets something like 300k lines of 'free' code.

[harryjackson]

Here here, an --exclude option would be great. I just tried the tool and it's the best looking reports I've seen so far.

[BrunoSpy]

:+1:

[nsams]

as a workaround I did filter-branch on the repo before running git_stats:

git filter-branch --index-filter 'git rm --cached --ignore-unmatch -r *.svg' HEAD

[paulz]

:+1:

[trentpolack]

Yeah, I'd adore this too. This is the only repo analyzer that's actually quite wonderful, but my game project has oodles of binary files that are getting included on top of everything else (and the binary files are ~120gb, so moving them around or temporarily removing them seems... Worrisome).