crunch.sh

#!/bin/bash

# Usage:
# ./crunch.sh

FILE_LIST=/tmp/dirless_file_list

rg --files | grep -v '.lock$' > $FILE_LIST

TOTAL_FILES=$(cat $FILE_LIST | wc -l)
TOTAL_LINES=$(cat $FILE_LIST | tr '\n' '\0' | xargs -0 cat | wc -l)
file_percentage () { echo "$(echo "scale=1; 100 * $1 / $TOTAL_FILES" | bc)%"; }
line_percentage () { echo "$(echo "scale=1; 100 * $1 / $TOTAL_LINES" | bc)%"; }
divide () { echo "scale=1; $1 / $2" | bc; }
format_number () { printf "%'d" $1; }

crunch () {
  files=$(cat $FILE_LIST | grep $1 $2 | wc -l)
  echo '    files:' $(format_number $files) "($(file_percentage $files))"

  lines=$(cat $FILE_LIST | grep $1 $2 | tr '\n' '\0' | xargs -0 cat | wc -l)
  echo -n '    lines:' $(format_number $lines) "($(line_percentage $lines))"

  echo " (avg $(divide $lines $files) per file)"
}

#######################################################################################################################
# CONFIGURE THE DIFFERENT SECTIONS YOU WANT HERE
sections=('\.css$' '\.html$' '\.js$' '\.py$' '\.csv$' '\.json$' '\.txt$' '\.sql')
#######################################################################################################################

for pattern in ${sections[@]}; do
  echo ___________________________________________
  echo $pattern
  crunch $pattern
done

echo ===========================================
echo NONE OF THE ABOVE PATTERNS
for i in ${!sections[@]}; do
  # modify the array to wrap each pattern in parens (a bit dodgy)
  sections[$i]="\(${sections[$i]}\)"
done
join_by () { local d=$1; shift; echo -n "$1"; shift; printf "%s" "${@/#/$d}"; }
crunch $(join_by '\|' ${sections[@]}) -v

# print all the files that didn't match any of the section patterns with this
# cat $FILE_LIST | grep -v $(join_by '\|' ${sections[@]})

echo ===========================================
echo OVERALL
crunch '.'