Script to walk through a git repository commit by commit and count the lines of code.

gitwalker.bash

#!/bin/bash

REPO_URL=$1
BRANCH=$2

TMP_DIR="./.tmp"
CLOC_EXCLUDED_DIRS="node_modules,vendor"
CLOC_EXCLUDED_LANGS="JSON,Markdown,YAML,SVG,SQL,Dockerfile,make"
OUT_FILE="./results.csv"

which cloc > /dev/null 2>&1 || {
  echo "This script uses cloc to count lines of code. Please install cloc to continue."
  echo "https://github.com/AlDanial/cloc"
  exit
}

which jq > /dev/null 2>&1 || {
  echo "This script uses jq to analyze the json output from cloc. Please install jq to continue."
  echo "https://github.com/stedolan/jq"
  exit
}

[ -z $REPO_URL ] && {
  echo "Please specify the repository URL as first parameter."
  exit 1
}

[ -z $BRNACH ] && {
  BRANCH="master"
}

# ---------------------------------

[ -d $TMP_DIR ] && rm -rf $TMP_DIR

git clone --branch $BRANCH $REPO_URL $TMP_DIR

echo "COMMIT_N,COMMIT_HASH,CODE_LINES,COMMENT_LINES,FILES" > $OUT_FILE

N_COMMITS=$(git -C $TMP_DIR rev-list origin/$BRANCH --count)
I=1
for COMMIT in $(git -C $TMP_DIR rev-list origin/$BRANCH --reverse); do
  git -C $TMP_DIR checkout $COMMIT > /dev/null 2>&1

  echo "[$I/$N_COMMITS] $COMMIT"

  CLOC_RES=$(cloc $TMP_DIR \
    --exclude-dir=$CLOC_EXCLUDED_DIRS \
    --exclude-lang=$CLOC_EXCLUDED_LANGS \
    --json | jq '.SUM')

  N_CODE=$(echo $CLOC_RES | jq ".code")
  N_COMMENT=$(echo $CLOC_RES | jq ".comment")
  N_FILES=$(echo $CLOC_RES | jq ".nFiles")

  echo "$I,$COMMIT,$N_CODE,$N_COMMENT,$N_FILES" >> $OUT_FILE

  I=$(( $I + 1 ))
done