mazda/nasa_archive.py

import hashlib
import logging
import os
import re
import shutil
from datetime import datetime, timedelta
from pathlib import Path

logger = logging.getLogger("nasa")


def file_get_hash(filename: str) -> str:
    with open(filename, "r") as frh:
        data = frh.read()
        return calculate_sha256(data)


def calculate_sha256(data: str) -> str:
    return hashlib.sha256(data.encode()).hexdigest()


def archive_files(export_dir: str):
    last_week = (datetime.now() - timedelta(days=6)).timestamp()
    for file in Path(export_dir).glob("*.json"):
        if file.stat().st_ctime < last_week:
            file.unlink()

    archive_path = Path(export_dir + "/Archiv")
    for file in Path(export_dir + "/temp").glob("*.json"):
        p = re.search(r"NASA_\d{5}_(20\d{4})_", file.name)
        if not p:
            continue
        period = p[1]
        year = period[:4]
        dest_folder = archive_path / year / period
        os.makedirs(dest_folder, exist_ok=True)
        file_hash = file_get_hash(file)

        if has_identical_file(dest_folder, file_hash):
            file.unlink()
            continue
        shutil.copy(file, archive_path.parent / file.name)
        file.rename(dest_folder / file.name)


def has_identical_file(target: Path, file_hash: str) -> bool:
    for archived_file in Path(target).glob("*.json"):
        if file_get_hash(archived_file) == file_hash:
            return True
    return False