Kiểm thực thể thương hiệu bằng JSON-LD và llms.txt: script Python khoảng 60 dòng

{
  "name": "Example Studio",
  "url": "https://example.com/",
  "telephone": "+84 912 345 678",
  "sameAs": [
    "https://www.facebook.com/examplestudio",
    "https://www.linkedin.com/company/examplestudio"
  ]
}

import json, re, requests

UA = {"User-Agent": "entity-check/1.0"}

def jsonld_blocks(url):
    html = requests.get(url, headers=UA, timeout=15).text
    blocks = re.findall(
        r'<script[^>]+application/ld\+json[^>]*>(.*?)</script>', html, re.S)
    out = []
    for b in blocks:
        try:
            data = json.loads(b)
        except json.JSONDecodeError:
            print("JSON-LD lỗi cú pháp:", b[:80])
            continue
        items = data.get("@graph", [data]) if isinstance(data, dict) else data
        out.extend(i for i in items if isinstance(i, dict))
    return out

def find_org(items):
    for i in items:
        t = i.get("@type")
        types = t if isinstance(t, list) else [t]
        if {"Organization", "LocalBusiness", "ProfessionalService"} & set(types):
            return i
    return None

def read_llms(base):
    r = requests.get(base.rstrip("/") + "/llms.txt", headers=UA, timeout=15)
    if r.status_code != 200:
        return None
    lines = r.text.splitlines()
    name = next((l[2:].strip() for l in lines if l.startswith("# ")), None)
    summary = next((l[2:].strip() for l in lines if l.startswith("> ")), "")
    phones = re.findall(r"\+?\d[\d .]{8,}\d", r.text)
    return {"name": name, "summary": summary, "phones": phones}

def digits(s):
    return re.sub(r"\D", "", s or "").removeprefix("84").lstrip("0")

def check(base, truth):
    org = find_org(jsonld_blocks(base))
    if not org:
        print("[LỖI] Không thấy Organization trong HTML ban đầu")
        return
    if org.get("name") != truth["name"]:
        print("[LỆCH] name:", org.get("name"), "≠", truth["name"])
    if org.get("url", "").rstrip("/") != truth["url"].rstrip("/"):
        print("[LỆCH] url:", org.get("url"))
    if digits(org.get("telephone")) != digits(truth["telephone"]):
        print("[LỆCH] telephone:", org.get("telephone"))
    missing = set(truth["sameAs"]) - set(org.get("sameAs", []))
    for m in missing:
        print("[THIẾU] sameAs:", m)
    for s in org.get("sameAs", []):
        code = requests.head(s, headers=UA, timeout=15, allow_redirects=True).status_code
        if code >= 400:
            print("[CHẾT] sameAs", code, s)

    llms = read_llms(base)
    if llms is None:
        print("[INFO] Không có llms.txt, bỏ qua")
    else:
        if llms["name"] != truth["name"]:
            print("[LỆCH] llms.txt name:", llms["name"])
        if llms["phones"] and digits(truth["telephone"]) not in map(digits, llms["phones"]):
            print("[LỆCH] llms.txt có số điện thoại khác:", llms["phones"])

if __name__ == "__main__":
    truth = json.load(open("entity.json", encoding="utf-8"))
    check(truth["url"], truth)