#!/usr/bin/env python3
"""TSC — ganti sisa rujukan domain lama (termasuk bentuk escaped di data blok) ke situs sekarang.

Kasus yang tertangkap: beranda masih memuat gambar latar dari
https://klinikdnisidoarjo.com/wp-content/uploads/2022/08/procedures-section-bg.jpg
padahal domain itu sudah tidak bisa diakses (DNS mati). Salinan gambarnya ada di situs baru.
"""
import base64, importlib.util, json, os, re, sys, time, urllib.error, urllib.request

WS = "/home/mantapman/.hermes/profiles/atun/workspace"
BACKUP = f"{WS}/tsc_backup_domain2"
os.makedirs(BACKUP, exist_ok=True)
spec = importlib.util.spec_from_file_location("cfg", f"{WS}/cfg_tsc.py")
m = importlib.util.module_from_spec(spec)
spec.loader.exec_module(m)
CFG = m.CFG
DOM = CFG["domain"]
AUTH = base64.b64encode(f"{CFG['user']}:{CFG['pass']}".encode()).decode()
UA = {"User-Agent": "Mozilla/5.0"}

# host saja yang diganti, supaya path gambar tetap utuh
POLA = [
    (re.compile(r"klinikdnisidoarjo\.com", re.I), "www.theskincodeclinic.com"),
    (re.compile(r"klinikdnikrian\.com", re.I), "www.theskincodeclinic.com"),
]


def rest(path, method="GET", data=None):
    h = {"Authorization": f"Basic {AUTH}", **UA}
    body = None
    if data is not None:
        body = json.dumps(data).encode()
        h["Content-Type"] = "application/json"
    r = urllib.request.Request(f"https://{DOM}{path}", data=body, headers=h, method=method)
    for _ in range(3):
        try:
            with urllib.request.urlopen(r, timeout=180) as x:
                return x.status, json.loads(x.read())
        except urllib.error.HTTPError as e:
            return e.code, e.read().decode()[:200]
        except Exception:
            time.sleep(4)
    return 0, "gagal"


go = "--go" in sys.argv
total = 0
for jenis in ("posts", "pages"):
    page = 1
    while True:
        st, d = rest(f"/wp-json/wp/v2/{jenis}?per_page=100&page={page}&status=publish&context=edit"
                     f"&_fields=id,slug,link,content")
        if not isinstance(d, list) or not d:
            break
        for x in d:
            raw = x["content"]["raw"]
            baru = raw
            n = 0
            for pola, ganti in POLA:
                n += len(pola.findall(baru))
                baru = pola.sub(ganti, baru)
            if not n:
                continue
            total += n
            print(f"  #{x['id']:<6} {jenis:<6} {n}× | {x['link'][:60]}")
            if go:
                open(f"{BACKUP}/{jenis}_{x['id']}.txt", "w").write(raw)
                st2, _ = rest(f"/wp-json/wp/v2/{jenis}/{x['id']}", "POST", {"content": baru})
                if st2 != 200:
                    print(f"        ❌ HTTP {st2}")
                time.sleep(0.3)
        if len(d) < 100:
            break
        page += 1

print(f"\ntotal rujukan domain lama diganti: {total}")
if not go:
    print("(dry run — pakai --go)")