#!/usr/bin/env python3
"""Ambil isi 3 artikel hc67 (via r.jina.ai; host DPI-block ke situs klien)."""
import json
import re
import subprocess
import sys
import time

URLS = {
    "threads": "https://studiosoft.id/konten-edukasi-kesehatan-kepercayaan-pasien/",
    "fb": "https://studiosoft.id/7-aplikasi-mental-health-anak-usia-4-10-tahun-yang-harus-dicoba-orang-tua/",
    "li": "https://studiosoft.id/cara-memilih-digital-agency-surabaya/",
}


def jina(url, tries=4):
    for i in range(tries):
        p = subprocess.run(["curl", "-s", "--max-time", "200", "https://r.jina.ai/" + url],
                           capture_output=True, text=True)
        raw = p.stdout or ""
        if len(raw) > 500:
            return raw
        time.sleep(4)
    return ""


out = {}
for k, u in URLS.items():
    raw = jina(u)
    out[k] = raw
    print("=== %s -> %d chars (%s)" % (k, len(raw), u))
    open("/tmp/hc67_%s.md" % k, "w").write(raw)

json.dump({k: len(v) for k, v in out.items()}, open("/tmp/hc67_fetch.json", "w"))
print("saved /tmp/hc67_{threads,fb,li}.md")
