#!/usr/bin/env python3
"""Cek fakta artikel konsultan pajak - tahap 2: baca teks pasal yang akan dikutip."""
import re, zlib, urllib.request, urllib.error

UA = {"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/124 Safari/537.36"}


def get(u, n=6000000):
    try:
        r = urllib.request.Request(u, headers=UA)
        with urllib.request.urlopen(r, timeout=120) as x:
            return x.status, x.read(n)
    except urllib.error.HTTPError as e:
        return e.code, b""
    except Exception as e:
        return type(e).__name__, b""


def teks(b):
    t = b.decode("utf-8", "replace")
    t = re.sub(r"<(script|style)[^>]*>.*?</\1>", " ", t, flags=re.S | re.I)
    t = re.sub(r"<[^>]+>", " ", t)
    t = re.sub(r"&nbsp;?", " ", t).replace("&amp;", "&").replace("&quot;", '"')
    return re.sub(r"\s+", " ", t).strip()


def pdf_text(b):
    hasil = []
    for c in re.findall(rb"stream\r?\n(.*?)endstream", b, re.S):
        try:
            d = zlib.decompress(c)
        except Exception:
            continue
        for m in re.finditer(rb"\[(.*?)\]\s*TJ", d, re.S):
            for s in re.finditer(rb"\(([^()\\]*)\)", m.group(1)):
                hasil.append(s.group(1).decode("latin-1", "replace"))
        for m in re.finditer(rb"\(([^()\\]{2,})\)", d):
            hasil.append(m.group(1).decode("latin-1", "replace"))
    return re.sub(r"\s+", " ", " ".join(hasil))


def potong(t, pat, sebelum=60, sesudah=700, maks=2):
    n = 0
    for m in re.finditer(pat, t, re.I):
        a, z = max(0, m.start() - sebelum), min(len(t), m.start() + sesudah)
        print(f"  >{t[a:z]}")
        n += 1
        if n >= maks:
            break
    if n == 0:
        print("  [TIDAK DITEMUKAN]", pat)


def buka(url, label, pola):
    st, b = get(url)
    print(f"\n=== {label} -> HTTP {st} {len(b)}B")
    if st != 200 or not b:
        return ""
    t = pdf_text(b) if b[:4] == b"%PDF" else teks(b)
    print(f"    teks {len(t)} char | awal: {t[:150]}")
    for p in pola:
        potong(t, p)
    return t


# 1) Konsolidasi UU KUP (teks berlaku setelah UU HPP) - Pasal 25, 26, 27
tk = buka("https://bats-consulting.com/konsolidasi-uu-kup", "Konsolidasi UU KUP", [
    r"Pasal 25\s*\(1\)", r"Pasal 26\s*\(1\)", r"Pasal 27\s*\(1\)"])
open("/tmp/kup.txt", "w").write(tk)

# 2) UU 14/2002 Pengadilan Pajak - Pasal 23 (gugatan), Pasal 36 (banding), Pasal 33
tp = buka("https://jdih.kemenkeu.go.id/api/download/fulltext/2002/14TAHUN2002UU.htm", "UU 14/2002", [
    r"Pasal 23\b", r"Pasal 36\b", r"Pasal 33\b", r"Pasal 34\b"])
open("/tmp/pp.txt", "w").write(tp)

# 3) PMK 111/PMK.03/2014 Konsultan Pajak
tm = buka("https://jdih.kemenkeu.go.id/api/download/5a00d249-d68c-427e-b688-0906eefa5ed0/PMK%20111%202014.pdf",
          "PMK 111/2014", [r"Pasal 1\b", r"Pasal 2\b", r"Pasal 3\b", r"Pasal 4\b", r"Pasal 5\b"])
open("/tmp/pmk111.txt", "w").write(tm)
