backupmanager_ynh/sources/retention.py

import os
import re
from datetime import datetime, timedelta


def apply_retention(job, new_archive_name, backup_dir):
    """Applique la politique de rétention après une sauvegarde réussie."""
    archives = _list_archives_for_job(job, backup_dir)
    if job.retention_mode == "count":
        to_delete = _retention_count(archives, job.retention_value)
    elif job.retention_mode == "daily":
        to_delete = _retention_daily(archives, job.retention_value)
    else:
        return []

    deleted = []
    for archive_filename in to_delete:
        base = os.path.splitext(archive_filename)[0]
        for ext in (".tar", ".info.json"):
            full = os.path.join(backup_dir, base + ext)
            if os.path.exists(full):
                os.remove(full)
                deleted.append(base + ext)
    return deleted


def _list_archives_for_job(job, backup_dir):
    """Liste les archives correspondant à ce job, triées par date (plus ancienne en premier)."""
    from flask import current_app
    instance = current_app.config["INSTANCE_NAME"]

    if job.type == "ynh_app":
        import json
        cfg = json.loads(job.config_json or "{}")
        app_id = cfg.get("app_id", "")
        prefix = f"{instance}_{app_id}_"
    elif job.type == "ynh_system":
        prefix = f"{instance}_system_"
    else:
        prefix = f"{instance}_{job.name.lower().replace(' ', '-')}_"

    archives = []
    for fname in os.listdir(backup_dir):
        if fname.startswith(prefix) and fname.endswith(".tar"):
            archives.append(fname)

    archives.sort(key=_extract_date)
    return archives


def _extract_date(filename):
    match = re.search(r'(\d{8})', filename)
    if match:
        try:
            return datetime.strptime(match.group(1), "%Y%m%d")
        except ValueError:
            pass
    return datetime.min


def _retention_count(archives, keep_n):
    if len(archives) <= keep_n:
        return []
    return archives[: len(archives) - keep_n]


def _retention_daily(archives, days):
    cutoff = datetime.utcnow() - timedelta(days=days)
    to_delete = []
    seen_dates = set()
    for archive in reversed(archives):
        date = _extract_date(archive)
        if date < cutoff:
            to_delete.append(archive)
            continue
        date_key = date.date()
        if date_key in seen_dates:
            to_delete.append(archive)
        else:
            seen_dates.add(date_key)
    return to_delete