find_large.py

import os
import subprocess
from pathlib import Path

def get_large_files():
    # Get all objects
    out = subprocess.check_output(['git', 'rev-list', '--objects', '--all']).decode('utf-8', errors='ignore')
    objects = {}
    for line in out.strip().split('\n'):
        parts = line.split(maxsplit=1)
        if len(parts) > 1:
            objects[parts[0]] = parts[1]
            
    # Find pack files
    pack_dir = Path('.git/objects/pack')
    idx_files = list(pack_dir.glob('*.idx'))
    
    if not idx_files:
        print("No pack files found.")
        return
        
    # Run verify-pack on all idx files
    verify_out = []
    for idx in idx_files:
        try:
            res = subprocess.check_output(['git', 'verify-pack', '-v', str(idx)]).decode('utf-8', errors='ignore')
            verify_out.extend(res.split('\n'))
        except:
            pass
            
    # Parse sizes
    sizes = []
    for line in verify_out:
        parts = line.split()
        if len(parts) >= 3 and parts[2].isdigit():
            sizes.append((parts[0], int(parts[2])))
            
    # Sort and print top 20
    sizes.sort(key=lambda x: x[1], reverse=True)
    for obj, size in sizes[:20]:
        print(f"{size/1024/1024:.2f} MB: {objects.get(obj, obj)}")

if __name__ == '__main__':
    get_large_files()