Deploymentcenter/scripts/deploy.py

#!/usr/bin/env python3
import os
import json
import ftplib
import sys
import hashlib
from pathlib import Path

SCRIPT_DIR = Path(__file__).resolve().parent
PROJECT_DIR = SCRIPT_DIR.parent
CONFIG_FILE = SCRIPT_DIR / 'deploy_config.json'
CACHE_FILE = SCRIPT_DIR / '.deploy_cache.json'

IGNORE_PATTERNS = {
    '.git',
    '.gitignore',
    'scripts',
    '.deploy_cache.json',
    'Serverdaten.txt',
    'Serverdaten.txt.bak',
    'bin',
    'obj',
    # Laufzeitdaten gehoeren dem Server, nicht dem Arbeitsplatz. var/.htaccess
    # wird dennoch uebertragen, damit das Verzeichnis existiert und gesperrt ist.
    'log',
    '__pycache__',
    'client-dotnet',
    # Gebaute Installer-Binaries. Sie gehoeren nach /installer/ und werden
    # von upload_installer.py dorthin gebracht - ueber diesen Weg landeten
    # sonst ~100 MB zusaetzlich unter /artifacts/ im Webroot.
    'artifacts',
}

def load_config():
    if not CONFIG_FILE.exists():
        print(f"Error: Config file {CONFIG_FILE} does not exist.")
        sys.exit(1)
    with open(CONFIG_FILE, 'r', encoding='utf-8') as f:
        return json.load(f)

def load_cache():
    if CACHE_FILE.exists():
        try:
            with open(CACHE_FILE, 'r', encoding='utf-8') as f:
                return json.load(f)
        except Exception:
            return {}
    return {}

def save_cache(cache):
    with open(CACHE_FILE, 'w', encoding='utf-8') as f:
        json.dump(cache, f, indent=2)

def compute_hash(filepath):
    h = hashlib.md5()
    with open(filepath, 'rb') as f:
        while chunk := f.read(8192):
            h.update(chunk)
    return h.hexdigest()

def ensure_remote_dir(ftp, remote_path):
    dirs = [d for d in remote_path.strip('/').split('/') if d]
    current = ''
    for d in dirs:
        current += '/' + d
        try:
            ftp.cwd(current)
        except ftplib.error_perm:
            try:
                ftp.mkd(current)
                print(f"Created remote directory: {current}")
            except Exception as e:
                print(f"Warning: Could not create directory {current}: {e}")

def should_ignore(rel_path):
    parts = Path(rel_path).parts
    if not parts:
        return False
    for part in parts:
        if part in IGNORE_PATTERNS:
            return True
        if part == '.htaccess':
            continue
        if part.startswith('.'):
            return True
    return False

def collect_changes(cache):
    """Ermittelt alle Dateien, die sich seit dem letzten Deployment geaendert haben."""
    files_to_upload = []
    for root, dirs, files in os.walk(PROJECT_DIR):
        rel_dir = os.path.relpath(root, PROJECT_DIR)
        if rel_dir == '.':
            rel_dir = ''

        dirs[:] = [d for d in dirs if not should_ignore(os.path.join(rel_dir, d))]

        for f in files:
            rel_file = os.path.normpath(os.path.join(rel_dir, f)).replace('\\', '/')
            if should_ignore(rel_file):
                continue

            full_path = os.path.join(root, f)
            file_hash = compute_hash(full_path)

            if cache.get(rel_file) != file_hash:
                files_to_upload.append((rel_file, full_path, file_hash))

    files_to_upload.sort(key=upload_priority)
    return files_to_upload


def upload_priority(entry):
    """
    Reihenfolge des Uploads.

    Abhaengigkeiten zuerst: waere public/index.php schon oben, waehrend
    src/bootstrap.php noch fehlt, liefe die Seite in der Zwischenzeit in einen
    Fehler. Die .htaccess im Wurzelverzeichnis kommt zuletzt, damit neue Routen
    erst greifen, wenn ihre Zielskripte vorhanden sind.
    """
    rel_file = entry[0]

    if rel_file == '.htaccess':
        rank = 4
    elif rel_file.startswith('public/'):
        rank = 3
    elif rel_file.startswith(('src/', 'config/', 'sql/', 'var/')):
        rank = 1
    else:
        rank = 2

    return (rank, rel_file)


def dry_run():
    """Zeigt an, was uebertragen wuerde, ohne eine Verbindung aufzubauen."""
    cache = load_cache()
    changes = collect_changes(cache)

    if not changes:
        print("Keine Aenderungen - der Server ist auf dem aktuellen Stand.")
        return

    total = sum(os.path.getsize(p) for _, p, _ in changes)
    print(f"{len(changes)} Datei(en) wuerden uebertragen ({total / 1024:.1f} KB):\n")
    for rel_file, full_path, _ in changes:
        marker = 'NEU' if rel_file not in cache else '   '
        print(f"  {marker}  {rel_file}  ({os.path.getsize(full_path) / 1024:.1f} KB)")


def deploy():
    config = load_config()
    cache = load_cache()
    new_cache = dict(cache)

    print(f"Connecting to FTP {config['host']}...")
    try:
        if config.get('secure', False):
            ftp = ftplib.FTP_TLS()
            ftp.connect(config['host'], config.get('port', 21))
            ftp.login(config['user'], config['pass'])
            ftp.prot_p()
        else:
            ftp = ftplib.FTP()
            ftp.connect(config['host'], config.get('port', 21))
            ftp.login(config['user'], config['pass'])
        print("Logged in successfully.")
    except Exception as e:
        print(f"FTP Connection failed: {e}")
        sys.exit(1)

    files_to_upload = collect_changes(cache)

    if not files_to_upload:
        print("No changed files to upload. Remote is up to date!")
        ftp.quit()
        return

    print(f"Found {len(files_to_upload)} file(s) to upload:")
    for rel_file, full_path, file_hash in files_to_upload:
        print(f" -> {rel_file}")

    uploaded = 0
    try:
        for rel_file, full_path, file_hash in files_to_upload:
            remote_file_path = f"/{rel_file}"
            remote_dir = os.path.dirname(remote_file_path).replace('\\', '/')
            if remote_dir and remote_dir != '/':
                ensure_remote_dir(ftp, remote_dir)

            ftp.cwd('/')
            print(f"Uploading {rel_file} ...")
            with open(full_path, 'rb') as f:
                ftp.storbinary(f"STOR {remote_file_path}", f)

            new_cache[rel_file] = file_hash
            uploaded += 1
    finally:
        # Der Cache wird auch bei einem Abbruch geschrieben. Sonst beginnt der
        # naechste Lauf wieder bei null und ueberträgt alles erneut.
        save_cache(new_cache)
        try:
            ftp.quit()
        except Exception:
            pass

    print(f"\nDeployment abgeschlossen: {uploaded} von {len(files_to_upload)} Datei(en) uebertragen.")

if __name__ == '__main__':
    if '--dry-run' in sys.argv or '-n' in sys.argv:
        dry_run()
    else:
        deploy()