generate_validate_hashes.py

"""
This tool calculates hashes for ISO files using 7z's hashing capabilities.
Optionally it prompts the user to verify that the hash is valid (usually by checking Google).
TODO:
 - Gather info from user on the ISO file
 - Distribution, Version, Architecture, URL Source(s), Hashes (Computed), Notes (Specifically, How Was Verified?)
[s] skip
[l] mark legacy
[d] set Distro
[v] set Version
[a] set Arch
[n] set Note
[c] recalculate hash
  - if change from known, prompt
  - if no change, let user know
  - if not previously known, add
[C] mark hashes verified
"""

import os
import re
import subprocess

CRC32_HASH_MATCH = r"[0-9A-F]{8}"
CRC64_HASH_MATCH = r"[0-9A-F]{16}"
SHA256_HASH_MATCH = r"[0-9A-F]{64}"
SHA1_HASH_MATCH = r"[0-9A-F]{40}"
BLAKE2sp_HASH_MATCH = r"[0-9A-F]{64}"

PROMPT_FOR_VALID = True
SKIP_LEGACY = True

_7Z_HASH_MATCH = (
    r"^"
    + " ".join(
        [
            CRC32_HASH_MATCH,
            CRC64_HASH_MATCH,
            SHA256_HASH_MATCH,
            SHA1_HASH_MATCH,
            BLAKE2sp_HASH_MATCH,
        ]
    )
    + r"\s+(\d+)\s+(.*)$"
)
_7Z_RE = re.compile(_7Z_HASH_MATCH)

if os.name == "nt":
    import msvcrt

    def getch():
        return msvcrt.getwch()


else:
    import sys, tty, termios

    def getch():
        fd = sys.stdin.fileno()
        old_settings = termios.tcgetattr(fd)
        try:
            tty.setraw(sys.stdin.fileno())
            ch = sys.stdin.read(1)
        finally:
            termios.tcsetattr(fd, termios.TCSADRAIN, old_settings)
        return ch

for root, dirs, files in os.walk("."):
    if SKIP_LEGACY and 'Legacy' in root:
        print('Skipping Legacy directory', root)
        continue
    else: print('Working on directory:', root)
    isos = [x for x in files if x.endswith(".iso")]
    for iso in isos:
        iso_path = os.path.join(root, iso)
        iso_hashes_path = iso_path + ".hashes"
        print("working on", iso_path)
        if os.path.exists(iso_hashes_path):
            print("path exists:", iso_hashes_path)
            with open(iso_hashes_path, "r") as f:
                data = f.read()
                if 'Valid' in data:
                    print('Already hashed & marked for validity, continuing...')
                    print([x for x in data.split('\n') if 'Valid' in x][0])
                    continue
        else:
            data = subprocess.run(
                ["7z", "h", "-scrc*", iso_path], capture_output=True, text=True
            ).stdout
            with open(iso_hashes_path, 'w') as f: f.write(data)

        if PROMPT_FOR_VALID:
            lines = data.split("\n")
            sha1 = [x for x in lines if x.startswith('SHA1')][0].split(' ')[-1]
            sha2 = [x for x in lines if x.startswith('SHA256')][0].split(' ')[-1]
            subprocess.Popen(['clip'], stdin=subprocess.PIPE).communicate(sha2.encode())
            print(data)
            print(iso_path)
            print(sha2)
            print('Are these hashes valid? ([y]es/[n]o/[s]kip]/[d]elete/[q]uit/sha[1]/sha[2]')

            while True:
                answer = getch()
                if answer.startswith('y'):
                    lines += ['Valid: true']
                    print('Marking valid @', iso_hashes_path)
                    break
                if answer.startswith('n'):
                    lines += ['Valid: false']
                    print('Marking invalid @', iso_hashes_path)
                    break
                if answer.startswith('s'):
                    print('Skipping')
                    break
                if answer.startswith('d'):
                    print('Deleting')
                    os.unlink(iso_path)
                    break
                if answer.startswith('q'):
                    print('Quitting')
                    sys.exit()
                if answer.startswith('1'):
                    subprocess.Popen(['clip'], stdin=subprocess.PIPE).communicate(sha1.encode())
                    print('Copied SHA1 to Clipboard')
                if answer.startswith('2'):
                    subprocess.Popen(['clip'], stdin=subprocess.PIPE).communicate(sha2.encode())
                    print('Copied SHA2 to Clipboard')
            
            with open(iso_hashes_path, 'w') as f:
                f.write('\n'.join(lines))