Which Git commit stats are easy to pull

Question

Previously I have enjoyed TortoiseSvn's ability to generate simple commit stats for a given SVN repository. I wonder what is available in Git and am particularly interested in :

• Number of commits per user
• Number of lines changed per user
• activity over time (for instance aggregated weekly changes)

Any ideas?

Answer 1

The best tool so far I identfied is gitinspector. It give the set report per user, per week etc

You can install like below with npm

npm install -g gitinspector

Details to get the links are below

https://www.npmjs.com/package/gitinspector
https://github.com/ejwa/gitinspector/wiki/Documentation
https://github.com/ejwa/gitinspector

example commands are

gitinspector -lmrTw
gitinspector --since=1-1-2017

etc

Answer 2

Note that, if your repo is on GitHub, you now (May 2013) have a new set of GitHub API to get interesting statistics.
See "File CRUD and repository statistics now available in the API"

That would include:

• Contributors
• Commit Activity
• Code Frequency
• Participation
• Punch Card

Answer 3

Here is a simple ruby script that I used to get author, lines added, lines removed, and commit count from git. It does not cover commits over time.

Note that I have a trick where it ignores any commit that adds/removes more than 10,000 lines because I assume that this is a code import of some sort, feel free to modify the logic for your needs. You can put the below into a file called gitstats-simple.rb and then run

git log --numstat --pretty='%an' | ruby gitstats-simple.rb

contents of gitstats-simple.rb

#!/usr/bin/ruby

# takes the output of this on stdin: git log --numstat --prety='%an'

map = Hash.new{|h,k| h[k] = [0,0,0]}
who = nil
memo = nil
STDIN.read.split("\n").each do |line|
  parts = line.split
  next if parts.size == 0
  if parts[0].match(/[a-z]+/)
    if who && memo[0] + memo[1] < 2000
      map[who][0] += memo[0]
      map[who][1] += memo[1]
      map[who][2] += 1
    end
    who = parts[0]
    memo = [0,0]
    next
  end
  if who
    memo[0]+=line[0].to_i
    memo[1]+=parts[1].to_i
  end
end

puts map.to_a.map{|x| [x[0], x[1][0], x[1][1], x[1][2]]}.sort_by{|x| -x[1] - x[2]}.map{|x|x.inspect.gsub("[", "").gsub("]","")}.join("\n")

Answer 4

I've written a small shell script that calculates merge statistics (useful when dealing with a feature-branch-based workflow). Here's an example output on a small repository:

[$]> git merge-stats
% of Total Merges               Author  # of Merges  % of Commits
            57.14     Daniel Beardsley            4          5.63
            42.85        James Pearson            3         30.00

Answer 5

You can use gitlogged gem (https://github.com/dexcodeinc/gitlogged) to get activities by author and date. This will give you report like this:

gitlogged 2016-04-25 2016-04-26

which returns the following output

################################################################

Date: 2016-04-25

Yunan (4):
      fix attachment form for IE (#4407)
      fix (#4406)
      fix merge & indentation attachment form
      fix (#4394) unexpected after edit wo

gilang (1):
      #4404 fix orders cart


################################################################
################################################################

Date: 2016-04-26

Armin Primadi (2):
      Fix document approval logs controller
      Adding git tool to generate summary on what each devs are doing on a given day for reporting purpose

Budi (1):
      remove validation user for Invoice Processing feature

Yunan (3):
      fix attachment in edit mode (#4405) && (#4430)
      fix label attachment on IE (#4407)
      fix void method (#4427)

gilang (2):
      Fix show products list in discussion summary
      #4437 define CApproved_NR status id in order


################################################################

Answer 6

First, you don't have to pull anything (as in network pull), because you have the whole repository and the whole history locally. I'm pretty sure there are tools that will give you statistics, but sometimes you can just be creative with the command lines. For instance, this (just out of my head) will give you the number of commits per user:

git log --pretty=format:%ae \
| gawk -- '{ ++c[$0]; } END { for(cc in c) printf "%5d %s\n",c[cc],cc; }'

Other statistics you asked for may need more thought put into it. You may want to see the tools available. Googling for git statistics points to the GitStats tool, which I have no experience with and even less idea of what it takes to get it run on windows, but you can try.