Lines of code per git author

[RocketJannis]

It would be cool, if there was an option to show lines of code per Git author using git blame.

I found this command somewhere some time ago that does exactly that (note that it filters for java and kt files here):

git ls-tree --name-only -z -r HEAD|egrep -z -Z -E '\.(java|kt)$' |xargs -0 -n1 git blame --line-porcelain|grep "^author "|sort|uniq -c|sort -nr

The output looks like this:

  22916 author jannis
   2076 author other name
   1576 author another name
   1552 author 3rd name

The problem is that this also counts blank lines and comments which cloc already solves.
Feel free to just close if this is out of scope for cloc. Thank you for reading and the best line counter I know :)

[AlDanial]

This is an interesting exercise. My approach would be to use cloc's ability to write results as SQL, then create an additional table that maps authors to files, something like this:

   # first do regular count
   cloc --sql=this_project.sql .
 
   # then create a table matching authors to files
   echo "create table authors( name text, file text);" > add_authors.sql
   echo "begin transaction;" >> add_authors.sql
   git ls-tree -r HEAD --name-only | while read file; do
       if [ -f "$file" ]; then
           author=$(git blame --line-porcelain "$file" | grep "^author " | cut -d' ' -f2- | sort | uniq -c | sort -nr | head -n 1 | sed -E 's/^[[:space:]]*[0-9]+[[:space:]]+//')
           echo "insert into authors values('$author', '$file');"
       fi
   done >> add_authors.sql
   echo "commit;" >> add_authors.sql
 
   # finally create the database
   cat this_project.sql add_authors.sql | sqlite3 this_project.db

Then craft a select statement to aggregate results by authors. I don't have a fully working example (yet) but figured I'd throw my work-in-progress ideas out there.

[Safari77]

I spent two hours on this after morning coffee :)

$ git-cloc.py kernel/*.c
AUTHOR                                        FILES     CODE  DISTRIBUTION                          
-------------------------------------------------------------------------------------------------
Linus Torvalds                                   44     3463  █▋···························· 5.75%
Tejun Heo                                        18     3058  █▌···························· 5.08%
Eric W. Biederman                                31     2816  █▍···························· 4.68%
Thomas Gleixner                                  31     2365  █▏···························· 3.93%
Al Viro                                          26     2249  █····························· 3.74%
Christian Brauner                                20     1855  ▉····························· 3.08%
Oleg Nesterov                                    20     1622  ▊····························· 2.70%
Masami Hiramatsu                                 11     1463  ▋····························· 2.43%
Paul E. McKenney                                 20     1440  ▋····························· 2.39%
Peter Zijlstra                                   32     1219  ▌····························· 2.03%
Frederic Weisbecker                              23     1195  ▌····························· 1.99%
Richard Guy Briggs                                8     1066  ▌····························· 1.77%
Christoph Hellwig                                22      925  ▍····························· 1.54%
David Howells                                    21      898  ▍····························· 1.49%
Dave Young                                        5      858  ▍····························· 1.43%
Paul Moore                                        6      856  ▍····························· 1.42%
Eric Paris                                        9      833  ▍····························· 1.38%
Baoquan He                                       11      749  ▎····························· 1.24%
Sebastian Andrzej Siewior                        29      711  ▎····························· 1.18%
Alexander Graf                                    4      685  ▎····························· 1.14%
Coiby Xu                                          4      679  ▎····························· 1.13%
Mathieu Desnoyers                                 9      613  ▎····························· 1.02%
Jens Axboe                                       10      531  ▎····························· 0.88%
Alexey Dobriyan                                  28      518  ▎····························· 0.86%
Kees Cook                                        29      512  ▎····························· 0.85%
Andrey Konovalov                                  3      465  ▏····························· 0.77%
[output truncated]

Parallel processing by default.

$ time git-cloc.py kernel/reboot.c --per-file --email|tr @ x
AUTHOR           EMAIL                         FILE                    BLANK  COMMENT     CODE
-----------------------------------------------------------------------------------------------
Aaro Koskinen    aaro.koskinenxnokia.com       kernel/reboot.c             2        0       12
Ahmad Fatoum     a.fatoumxpengutronix.de       kernel/reboot.c            12        0       93
Andrey Smirnov   andrew.smirnovxgmail.com      kernel/reboot.c             4        0       22
Benjamin Bara    benjamin.baraxskidata.com     kernel/reboot.c             0        0        4
Chuansheng Liu   chuansheng.liuxintel.com      kernel/reboot.c             0        7        1
Dave Young       dyoungxredhat.com             kernel/reboot.c             0        0        1
Dmitry Osipenko  dmitry.osipenkoxcollabora.com kernel/reboot.c            51      112      214
Dongmin Lee      ldmldm05xgmail.com            kernel/reboot.c             1        6        7
Fabio Estevam    festevamxdenx.de              kernel/reboot.c             0        0       11
Guenter Roeck    linuxxroeck-us.net            kernel/reboot.c             3       62       15
Joel Granados    joel.granadosxkernel.org      kernel/reboot.c             0        0        1
Joel Stanley     joelxjms.id.au                kernel/reboot.c            10        6       31
Kai-Heng Feng    kai.heng.fengxcanonical.com   kernel/reboot.c             2        4       10
Kees Cook        keesxkernel.org               kernel/reboot.c             0        0        3
Matteo Croce     mcrocexmicrosoft.com          kernel/reboot.c            37        4      188
Matti Vaittinen  mazziesaccountxgmail.com      kernel/reboot.c             8       23       14
Mike Rapoport    rpptxkernel.org               kernel/reboot.c             0        0        1
Paul E. McKenney paulmckxkernel.org            kernel/reboot.c             0        0        1
Petr Mladek      pmladekxsuse.com              kernel/reboot.c             0        0        1
Pingfan Liu      kernelfansxgmail.com          kernel/reboot.c             0        0        3
Robin Holt       holtxsgi.com                  kernel/reboot.c            54       99      230
Shawn Guo        shawn.guoxlinaro.org          kernel/reboot.c             0        0        1
Thomas Gleixner  tglxxlinutronix.de            kernel/reboot.c             0        1        0
Thomas Weißschuh linuxxweissschuh.net          kernel/reboot.c             0        5        1
Vivek Goyal      vgoyalxredhat.com             kernel/reboot.c             0        0        1
YueHaibing       yuehaibingxhuawei.com         kernel/reboot.c             1        0       23
tangmeng         tangmengxuniontech.com        kernel/reboot.c             0        0        4
zhangguopeng     zhangguopengxkylinos.cn       kernel/reboot.c             0        0        4

real	0m0,390s
user	0m0,309s
sys	0m0,096s

https://github.com/Safari77/git-cloc

[AlDanial]

@Safari77 that's an interesting program but I wonder why the need to filter inputs by extension? I had to comment out lines 346 and 347 to get complete results:

344     for f in all_files:
345         if not f: continue
346    #     if not f.endswith(active_extensions):
347    #         continue            

In some cases files with multiple contributors, eg README.md, do show multiple names but others like cloc only show one name despite git blame showing several names. Still, a good start.

[Safari77]

@Safari77 that's an interesting program but I wonder why the need to filter inputs by extension? I had to comment out lines 346 and 347 to get complete results:

344 for f in all_files:
345 if not f: continue
346 # if not f.endswith(active_extensions):
347 # continue

In some cases files with multiple contributors, eg README.md, do show multiple names but others like cloc only show one name despite git blame showing several names. Still, a good start.

It by default processes only some common "code" extensions.

$ git-cloc.py --extensions all README.md |head
AUTHOR                      FILES     CODE  DISTRIBUTION                          
----------------------------------------------------------------------------------
Al Danial                       1     1719  █████████████████▋⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱  59.13%
AlDanial                        1      574  █████▉⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱  19.75%
Your Name                       1      267  ██▊⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱   9.18%
Brahim Djoudi                   1      209  ██▏⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱   7.19%
MichaelDimmitt                  1       35  ▎⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱   1.20%
Ave                             1       25  ▎⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱   0.86%
Zev Eisenberg                   1        9  ▏⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱   0.31%
Jakub Wilk                      1        6  ▏⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱⸱   0.21%