Count lines by author within a git repo or a subdirectory of it

lines-by-author.sh

#!/bin/bash

# Shows the number of lines authored by each author in the current repo or the specified subdirectory
lines-by-author() {
  if ! command -v parallel >/dev/null 2>&1; then
    echo "GNU parallel is not installed. Please install it to run this script."
    exit 1
  fi

  # Default to the repo root if no path is provided
  repo_root=$(git rev-parse --show-toplevel) || exit 1
  if [ -z "$1" ]; then
    path=$repo_root
  else
    path=$1
  fi

  # Print the directory of path relative to the repo root
  rel_path=$(realpath --relative-to="$repo_root" "$path")
  echo "Counting lines by author in $rel_path"

  # list tracked files under $path, blame in parallel, tally authors
  git ls-files -z -- "$path" |
    parallel -0 --no-notice -j "$(nproc)" \
      'git blame --line-porcelain {} 2>/dev/null' |
    awk '
      /^author / {
        author = substr($0, 8)
        counts[author]++
      }
      END {
        for (a in counts)
          printf "%d %s\n", counts[a], a
      }
    ' |
    sort -rn
}

lines-by-author "$1"