script to print out git diff line changes per specific files category

readme.md

This script reads git diff and filters lines based on file category. This allows to see line changes for only source files or test or doccumentation.

If you add this script as git alias you'll be able to call it like git count source

$ ./git-diff-count-lines.sh count source
in 5 files: added: 13, removed: 142, total: 155, diff: -129

$ ./git-diff-count-lines.sh count source --nums
5 13 142 155 -129

$ git count source configs
in 11 files: added: 202, removed: 304, total: 506, diff: -102

$ git count list
source:     5    +13   -142   =155   :129
tests:      0     +0     -0     =0     :0
code:       5    +13   -142   =155   :129
configs:    6   +189   -162   =351    :27
assets:     0     +0     -0     =0     :0
docs:       0     +0     -0     =0     :0

Categories can be configured in count-lines.ini, to include line write awk regex for it, to exclude prepend ! to regex.

[category]
include = /(dir1|dir2)\/.*\.(ext1|ext2)/
exclude = !/^unwanted.file/

count-lines.ini

[code]
code  = /(src|styles)\/.*\.(jsx?|tsx?|[sp]?css|vue|html)/
# exclude
specs = !/\.specs?\./
tests = !/\s*tests\//

[tests]
specs = /\.specs?\./
tests = /^tests\//

[source]
code  = /(src|styles)\/.*\.(jsx?|tsx?|[sp]?css|vue|html)/
specs = /\.specs?\./
tests = /\s*tests\//

[configs]
dot-files    = /^\..*|\/\..*/
dot-config   = /.*\.config\..*/
config-exts  = /\.(json|yml|cgf|rc|toml|ini|properties)/
build        = /^build\//
# exclude
package-lock = !/package-lock\.json|yarn\.lock/

[assets]
asset-folder = /^assets\/.*/
asset-exts   = /\.(svg|ico|jpg|jpeg|png|webp|webm|tiff|woff|otf|ttf)/

[docs]
docs-folder = /^docs\/.*/
md_files    = /\.md/

git-diff-count-lines.sh

#!/usr/bin/env bash

function log() {
    local content
    if [[ $# != 0 ]] || [[ -t 0 ]]; then
        content="${@}"
    else
        content="$( < /dev/stdin )"
    fi
    # direct to stdout
    echo "$content" >&2
}

function  count-lines() {
    local config_text=$(gawk '{
        FS="=";
        catMatch = match($0, /^\[(.*)\]/, _ca)
        if (catMatch != 0) {
            cat = _ca[1]
        }
        valMatch = match($0, /^\s*([a-zA-Z0-9_\-]+)\s*=\s*(!?)\/(.*)\/\s*$/, _va)
        if (valMatch != 0) {
            printf("%s %s %s %s\n", cat, _va[1], _va[3], (_va[2]=="!" ? "-" : "+"));
        }
    }' "${BASH_SOURCE%/*}/count-lines.ini")

    declare -A unique_cats
    while read -r category other; do
#        log "cat ${category}"
        unique_cats[$category]=1
    done <<< "$config_text"

    local categories=( "${!unique_cats[@]}" )

    local include_patterns=()
    local exclude_patterns=()
    local numbers_only=0
    local list_all=0
    local verbose=0
    local sub_args=""

    local filter_categories=()
    while [[ $# -gt 0 ]]; do
        opt="$1";
        shift;              #expose next argument
        case "$opt" in
            "--incl" ) include_patterns+=("($1)"); shift;;
            "--excl" ) exclude_patterns+=("($1)"); shift;;
            "--nums" ) numbers_only=1;;
            "--verb" ) verbose=1;;

            *)
                if [[ " ${categories[@]} " =~ " ${opt} " ]]; then
                    filter_categories+=( "$opt" )
                else
                    log "Unknown category '$opt'"
                    return 1
                fi
            ;;
       esac
    done

    if [ ${#filter_categories[@]} -eq 0 ]; then
        list_all=1
    fi

    local changes
    if [[ -t 0 ]]; then # terminal
        changes=$(git diff --numstat)
    else # pipe
        changes="$( < /dev/stdin )"
    fi

    if [[ ${list_all} == 1 ]]; then
        for cat in "${categories[@]}"; do
            local args=""
            [[ $verbose = 1 ]] && args="$args --verb";
            local nums="$(echo "$changes" | count-lines --nums "$cat" $args)"
            [[ $? != 0 ]] && return $?;
	        echo "$cat: $nums" | awk '{
	            printf("%-8s %4d %6s %6s %6s %6s\n", $1, $2, "+"$3, "-"$4, "="$5, ":"($6>0?$6:-$6))
            }'

        done
        return
    fi

    while read -r category section pattern action; do
        if [[ " ${filter_categories[@]} " =~ " ${category} " ]]; then
#            log "$category $pattern $action"
            [[ "$action" == "+" ]] && include_patterns+=("($pattern)");
            [[ "$action" == "-" ]] && exclude_patterns+=("($pattern)");
        fi
    done <<< "$config_text"

    local match_pattern="$( ( IFS=$'|'; echo "${include_patterns[*]}" ))"
    local unmatch_pattern="$( ( IFS=$'|'; echo "${exclude_patterns[*]}" ))"

    if [[ ! -z "${unmatch_pattern// }" ]]; then
        changes=$(echo "$changes" | awk "\$3 !~ /$unmatch_pattern/")
    fi

    if [[ ! -z "${match_pattern// }" ]]; then
        changes=$(echo "$changes" | awk "\$3 ~ /$match_pattern/")
    else
        changes=""
    fi

    local awk_line='{
            if (length($0) != 0) files += 1;
            add += $1;
            rem += $2
        }'
    local awk_printf_str
    if [[ ${numbers_only} == 1 ]]; then
        awk_printf_str="%d %d %d %d %d\n"
    else
        awk_printf_str="in %d files: added: %d, removed: %d, total: %d, diff: %d\n"
    fi
    local awk_end="END {\
            tot = add + rem; dif = add - rem;\
            printf \"$awk_printf_str\", files, add, rem, tot, dif\
        }"
    local awk_cmd="$awk_line $awk_end"

    if [[ ${verbose} == 1 ]]; then
        echo "$changes" | awk '{print "  " $0}' | log
    fi

    echo "$changes" | awk "$awk_cmd"

}



function main() {
    local cmd="$1";
    shift
    case "$cmd" in
        "count" ) count-lines $@;;
        "list-counts" ) count-lines list;;
        *) echo >&2 "Unknown command: $cmd"; exit 1;;
    esac
}

main $@