#!/usr/bin/env python3
"""Fase 5 (local): paridad entre el fichero capturado y lo que sirve ahora el Joomla local.

Detecta capturas truncadas, paginas de error congeladas y desfases de contenido.
No toca produccion.
"""
import os, re, sys, random, hashlib, subprocess, json
from urllib.parse import urlsplit

BASE = "/home/rafa/joomla-migration/mirror-antiguo"
RUN = open(os.path.join(BASE, "CURRENT_RUN")).read().strip()
DIR = os.path.join(BASE, "runs", RUN)
RAW = os.path.join(DIR, "raw", "antiguo.feadulta.com")
N = int(sys.argv[1]) if len(sys.argv) > 1 else 200

TITLE = re.compile(r"<title[^>]*>(.*?)</title>", re.I | re.S)
SCRIPTS = re.compile(r"<(script|style)[^>]*>.*?</\1>", re.I | re.S)
TAGS = re.compile(r"<[^>]+>")
WS = re.compile(r"\s+")
# bloques que cambian entre peticiones. El contador de visitas de K2 ("Read N times") se
# incrementa con nuestra propia peticion, asi que dos lecturas de la MISMA pagina nunca coinciden:
# se normaliza en vez de contarlo como diferencia.
VOLATILE = re.compile(r"[0-9a-f]{32}|csrf|token", re.I)
HITS = re.compile(r"(Read|Visto|Le[ií]do)\s+\d+\s+(times|veces)", re.I)

def texthash(s):
    s = SCRIPTS.sub(" ", s)
    s = TAGS.sub(" ", s)
    s = WS.sub(" ", s).strip()
    s = VOLATILE.sub("", s)
    s = HITS.sub("HITS", s)
    return hashlib.sha256(s.encode("utf-8", "replace")).hexdigest(), len(s)

def title(s):
    m = TITLE.search(s)
    return WS.sub(" ", m.group(1)).strip() if m else ""

urls = [l.strip() for l in open(os.path.join(BASE, "inventory", "urls-input.txt"))]
random.seed(20260729)
sample = random.sample(urls, min(N, len(urls)))

res = {"muestra": len(sample), "ok_status": 0, "falta_fichero": 0,
       "titulo_igual": 0, "titulo_distinto": 0, "texto_igual": 0, "texto_distinto": 0,
       "diffs": []}

for u in sample:
    path = urlsplit(u).path
    fp = os.path.join(RAW, path.lstrip("/"))
    if path.endswith("/"):
        fp = os.path.join(fp, "index.html")
    if not os.path.exists(fp):
        res["falta_fichero"] += 1
        res["diffs"].append({"url": u, "motivo": "fichero ausente"})
        continue
    res["ok_status"] += 1
    disk = open(fp, encoding="utf-8", errors="replace").read()
    live = subprocess.run(
        ["curl", "-s", "--max-time", "60", "-H", "Host: antiguo.feadulta.com",
         "http://127.0.0.1:8086" + path],
        capture_output=True).stdout.decode("utf-8", "replace")

    td, tl = title(disk), title(live)
    if td == tl:
        res["titulo_igual"] += 1
    else:
        res["titulo_distinto"] += 1
        res["diffs"].append({"url": u, "motivo": "titulo", "mirror": td[:120], "vivo": tl[:120]})

    hd, ld = texthash(disk)
    hl, ll = texthash(live)
    if hd == hl:
        res["texto_igual"] += 1
    else:
        res["texto_distinto"] += 1
        res["diffs"].append({"url": u, "motivo": "texto", "len_mirror": ld, "len_vivo": ll})

out = os.path.join(DIR, "parity-report.json")
json.dump(res, open(out, "w"), indent=2, ensure_ascii=False)
for k in ("muestra", "falta_fichero", "titulo_igual", "titulo_distinto", "texto_igual", "texto_distinto"):
    print(k, "=", res[k])
print("informe:", out)
for d in res["diffs"][:15]:
    print(" ", d)