The Curiosity

I've been writing code for years, accumulating numerous repositories on GitHub that I rarely revisit. One day, a thought struck me: "Just how many lines of code have I actually written?"

It's not about measuring productivity or comparing myself to others. The number itself doesn't really matter. It's more like flipping through an old photo album—you're not trying to achieve anything, just curious about the journey.

The Tool: cloc

I discovered cloc, a command-line utility that analyzes source code and produces statistics:

• Number of files per language
• Lines of code, comments, and blanks per language
• Clean, formatted output

Example output looks like this:

-------------------------------------------------------------------------------
Language                     files          blank        comment           code
-------------------------------------------------------------------------------
Python                         45           1234            567          8901
JavaScript                     32            876            234          5678
...                                
-------------------------------------------------------------------------------

The problem? All my code lives on GitHub. Cloning repositories one by one to run cloc locally would be tedious and time-consuming.

The Solution: Automate Everything

I wrote a bash script that handles the entire workflow:

1. Fetches all repositories from my GitHub account using the gh CLI
2. Syncs each repository locally (clones new ones, pulls updates for existing)
3. Runs cloc on the entire directory, excluding build artifacts and dependencies

Prerequisites

Install the required tools on your system:

macOS:

brew install gh jq cloc git

Ubuntu/Debian:

sudo apt update && sudo apt install gh jq cloc git

Windows (WSL):

wsl --install

Then within Ubuntu:

sudo apt update && sudo apt install gh jq cloc git

Important: Authenticate with GitHub CLI before running:

gh auth login

This opens a browser for one-time OAuth authorization.

The Script

#!/usr/bin/env bash

GITHUB_USER="your_username"   # Replace with your GitHub username
STORAGE_DIR="$HOME/projects_backup"

# Initialize storage directory
mkdir -p "$STORAGE_DIR"
cd "$STORAGE_DIR" || exit

echo "Fetching all repositories for $GITHUB_USER..."

# Retrieve repository URLs and process each one
gh repo list "$GITHUB_USER" --limit 500 --json nameWithOwner,url \
  | jq -r '.[].url' | while IFS= read -r repo_url; do
    local_name=$(basename "$repo_url")
    if [ -d "$local_name" ]; then
        echo "Updating existing: $local_name"
        cd "$local_name" && git pull --quiet && cd "$STORAGE_DIR"
    else
        echo "Cloning new repository: $local_name"
        git clone --depth 1 "$repo_url" --quiet
    fi
done

echo ""
echo "Computing code statistics (excluding vendor, node_modules, dist, build)..."
echo ""

cloc --exclude-dir=node_modules,vendor,dist,build,public,target,.git \
     --exclude-lang=JSON,SVG "$STORAGE_DIR"

The --depth 1 flag performs shallow clones, keeping disk usage minimal since we only need the latest code.

Sample Results

Running the script produces output like:

-------------------------------------------------------------------------------
Language                     files          blank        comment           code
-------------------------------------------------------------------------------
PHP                           2334          14079          60356         613174
Vuejs Component                232           2273           1207          29480
JavaScript                     198           2769           2813          16292
CSS                             13           1566            182          13460
YAML                            15           3345              1          12314
HTML                            43            507            262          11776
Markdown                        52           1881             11           4773
Blade                           27            386             52           3942
Go                               4             80             65            887
...                                
-------------------------------------------------------------------------------
SUM:                          3303          27162          65119         708414
-------------------------------------------------------------------------------

Simplifying for Local Projects

If your code is already on disk, skip the GitHub synchronization entirely. Just navigate to any directory and run:

cloc .

That's it—no scripts needed.

Summary

• cloc provides quick, detailed code statistics across multiple languages
• The gh CLI enables programmatic access to GitHub repositories
• Combining these tools with git creates a fully automated statistics pipeline
• While the metric itself is meeningless, it's surprisingly nostalgic to see