import os,json,hashlib
from datetime import datetime,timezone
import requests
from bs4 import BeautifulSoup

URL="https://cet.neea.edu.cn/"
STATE="data/state.json"

def category(t):
    if "报名" in t:return "📝 报名"
    if "成绩" in t:return "📊 成绩"
    if "准考证" in t:return "🎫 准考证"
    if "考试" in t:return "📝 考试安排"
    return "📢 其他"

def fetch():
    r=requests.get(URL,headers={"User-Agent":"Mozilla/5.0"},timeout=20)
    r.encoding="utf-8"
    soup=BeautifulSoup(r.text,"lxml")
    out=[]
    for a in soup.find_all("a"):
        t=a.get_text(strip=True)
        if t and any(k in t for k in ["四六级","CET","成绩","报名","考试","准考证"]):
            out.append({"title":t,"url":a.get("href"),"category":category(t)})
    return out[:20]

def load():
    if not os.path.exists(STATE): return []
    return json.load(open(STATE,encoding="utf-8")).get("items",[])

def save(items):
    os.makedirs("data",exist_ok=True)
    json.dump({"items":items,"last_check":datetime.now(timezone.utc).isoformat()},open(STATE,"w",encoding="utf-8"),ensure_ascii=False,indent=2)

def h(x):
    return hashlib.sha256((x["title"]+str(x["url"])).encode()).hexdigest()

def push(items):
    token=os.getenv("PUSH_TOKEN")
    if not token:return
    msg="📢 CET四六级新通知\n\n"
    for x in items:
        msg+=f"{x['category']}\n{x['title']}\n{x['url']}\n\n"
    requests.post("https://push.showdoc.com.cn/server/api/push",
        data={"title":"四六级新通知","content":msg,"token":token},timeout=10)

def main():
    current=fetch()
    old={x.get("hash") for x in load()}
    fresh=[]
    saved=[]
    for x in current:
        x["hash"]=h(x)
        saved.append(x)
        if x["hash"] not in old:
            fresh.append(x)
    if fresh: push(fresh)
    save(saved)

if __name__=="__main__":
    main()