#!/usr/bin/env python3
"""Capture only a completed, byte-identical and CRC-verified Forgejo ZIP."""
import argparse
import hashlib
from pathlib import Path
import secrets
import subprocess
import time
import zipfile


def validate_archive(path):
    with zipfile.ZipFile(path) as archive:
        names = set(archive.namelist())
        if 'forgejo-db.sql' not in names or not any(n.startswith('repos/') for n in names):
            raise ValueError('required archive content missing')
        if archive.testzip() is not None:
            raise ValueError('archive checksum failure')


def capture(namespace, pod, destination):
    k = ['kubectl', 'exec', '--request-timeout=120s', '-n', namespace, pod, '-c', 'gitea', '--']
    def call(args, timeout=150):
        r = subprocess.run(k + args, capture_output=True, timeout=timeout)
        if r.returncode: raise ValueError('capture command failed')
        return r.stdout
    remote = '/tmp/wp0029-backup-' + secrets.token_hex(12)
    completed = False
    try:
        # Completion belongs to this exact process, not a global pgrep or file existence.
        command = f'umask 077; forgejo dump -f {remote}.zip >{remote}.log 2>&1; result=$?; printf "%s" "$result" >{remote}.exit'
        call(['sh','-c', 'nohup sh -c "$1" >/dev/null 2>&1 </dev/null &', 'sh', command])
        for _ in range(180):
            r = subprocess.run(k + ['cat',remote+'.exit'], capture_output=True, timeout=30)
            if r.returncode == 0:
                completed = True
                if r.stdout.strip() != b'0': raise ValueError('dump process failed')
                break
            time.sleep(10)
        else: raise ValueError('dump completion timeout')
        size = int(call(['stat','-c%s',remote+'.zip']).strip())
        expected = call(['sha256sum',remote+'.zip']).split()[0].decode()
        if size <= 0: raise ValueError('empty archive')
        chunk = 4*1024*1024
        with destination.open('xb') as out:
            destination.chmod(0o600)
            for index in range((size+chunk-1)//chunk):
                piece = call(['dd','if='+remote+'.zip','bs='+str(chunk),'skip='+str(index),'count=1'])
                if len(piece) != min(chunk,size-index*chunk): raise ValueError('short archive chunk')
                out.write(piece)
        actual = hashlib.file_digest(destination.open('rb'),'sha256').hexdigest()
        if actual != expected: raise ValueError('archive transfer mismatch')
        validate_archive(destination)
    finally:
        # Do not remove a file while a timed-out producer may still be writing it.
        if completed:
            call(['rm','-f',remote+'.zip',remote+'.log',remote+'.exit'])


def main():
    p=argparse.ArgumentParser(description=__doc__)
    p.add_argument('--namespace',required=True); p.add_argument('--pod',required=True)
    p.add_argument('--output',required=True,type=Path)
    a=p.parse_args()
    try: capture(a.namespace,a.pod,a.output)
    except Exception:
        print('ERROR: Forgejo archive capture or integrity validation failed')
        return 1
    print('Forgejo archive capture and integrity verified')
    return 0

if __name__=='__main__': raise SystemExit(main())