# -*- coding: utf-8 -*-
"""What log history actually exists, per site. Listing only, nothing downloaded.

Paper 09 was parked on the assumption that the logs had rotated away. Before
planning around that, this asks each host what it holds: which archives exist,
how large they are, and how far back they go. It connects over FTPS, lists two
directories, and disconnects. It downloads nothing, writes nothing and prints no
credential.

    python log_inventory.py            every site with a stored credential
    python log_inventory.py broikos.gr collectorcenter.gr
"""
import ftplib
import glob
import io
import os
import re
import sys

HTDOCS = r'C:\xampp\htdocs'
LOG_DIRS = ('/logs', '/access-logs', '/tmp/awstats')


def sites():
    out = []
    for path in sorted(glob.glob(os.path.join(HTDOCS, '*', 'ftp.md'))):
        out.append(os.path.basename(os.path.dirname(path)))
    return out


def creds(site):
    """Host, user and password out of a note written by hand, in three formats.

    These files were written for a person, not a parser: one uses a scheme and
    Username/Password lines, one uses url/user/pass, one uses a markdown table.
    A parser that understands only the first reports the other sites as having no
    credential, which is a wrong answer about access rather than a missing file.
    """
    text = io.open(os.path.join(HTDOCS, site, 'ftp.md'),
                   encoding='utf-8', errors='replace').read()

    def field(*names):
        for n in names:
            m = re.search(r'(?im)^\s*\|?\s*%s\s*[:|]\s*\|?\s*([^\s|]+)' % n, text)
            if m:
                return m.group(1).strip('*` ')
        return None

    host = None
    m = re.search(r'https?://([^\s/]+)', text)
    if m:
        host = m.group(1)
    if not host:
        host = field('host', 'server', 'ftp host')
    if not host:
        m = re.search(r'(?m)^\s*([a-z0-9-]+\.[a-z.]{2,})\s*$', text)
        host = m.group(1) if m else site
    user = field('username', 'user', 'ftp user', 'login')
    pwd = field('password', 'pass', 'ftp password')
    if not (host and user and pwd):
        return None
    return host, user, pwd


def listing(ftp, path):
    lines = []
    try:
        ftp.retrlines('LIST %s' % path, lines.append)
    except ftplib.all_errors:
        return []
    out = []
    for line in lines:
        parts = line.split(maxsplit=8)
        if len(parts) < 9 or parts[8] in ('.', '..'):
            continue
        out.append((parts[8], int(parts[4]) if parts[4].isdigit() else 0,
                    ' '.join(parts[5:8])))
    return out


def inspect(site):
    c = creds(site)
    if not c:
        print('%-22s no usable credential in ftp.md' % site)
        return
    host, user, pwd = c
    try:
        ftp = ftplib.FTP_TLS(timeout=40)
        ftp.connect(host, 21)
        ftp.login(user, pwd)
        ftp.prot_p()
    except Exception as exc:                                    # noqa: BLE001
        print('%-22s connect failed: %s' % (site, str(exc)[:70]))
        return
    found = {}
    for d in LOG_DIRS:
        entries = listing(ftp, d)
        if entries:
            found[d] = entries
    ftp.quit()

    if not found:
        print('%-22s no log directory reachable' % site)
        return
    for d, entries in found.items():
        archives = [e for e in entries if e[0].endswith(('.gz', '.txt'))]
        total = sum(e[1] for e in archives)
        months = sorted({re.search(r'([A-Z][a-z]{2}-\d{4})', e[0]).group(1)
                         for e in archives
                         if re.search(r'([A-Z][a-z]{2}-\d{4})', e[0])})
        print('%-22s %-14s %3d file(s), %6.1f MB%s'
              % (site, d, len(archives), total / 1e6,
                 ('  months: %s .. %s' % (months[0], months[-1])) if months else ''))


def main(argv):
    targets = argv[1:] or sites()
    print('%-22s %-14s %s' % ('site', 'directory', 'what is there'))
    print('-' * 78)
    for s in targets:
        inspect(s)
    print('\nlisting only. Nothing was downloaded and nothing was changed.')
    return 0


if __name__ == '__main__':
    raise SystemExit(main(sys.argv))