#!/usr/bin/env python3
# /opt/weval-l99/screens-health-check.py v2
# Distinction: UP / SLOW / BROKEN (5xx) / DOWN (timeout/DNS) / PROTECTED (401/403)
import json, os, sys, time, re, tempfile
from concurrent.futures import ThreadPoolExecutor, as_completed
from urllib.request import Request, urlopen
from urllib.error import URLError, HTTPError
from urllib.parse import quote, urlparse, urlunparse

CARTO_FILE = "/var/www/html/cartographie-screens.html"
OUT_FILE = "/var/www/html/api/screens-health.json"
WORKERS = 40
TIMEOUT = 10
SLOW_MS = 8000

def extract_data():
    if not os.path.exists(CARTO_FILE):
        return []
    with open(CARTO_FILE, "r", encoding="utf-8") as f:
        html = f.read()
    m = re.search(r"const DATA = (\[.*?\]);", html, re.DOTALL)
    if not m:
        return []
    try:
        return json.loads(m.group(1))
    except Exception:
        return []

def classify(code, elapsed_ms):
    if code == 0:
        return "DOWN"
    if 200 <= code < 400:
        return "SLOW" if elapsed_ms > SLOW_MS else "UP"
    if code in (401, 403):
        return "PROTECTED"
    if code == 404:
        return "NOT_FOUND"
    if code in (405, 501):
        return "UP"  # method not allowed = page exists but rejected HEAD
    if code == 400:
        return "PROTECTED"  # endpoint exists, expects POST/specific input
    if code == 429:
        return "SLOW"  # rate limited but endpoint alive
    if code == 502:
        return "SLOW"  # upstream transient
    if 500 <= code < 600:
        return "BROKEN"
    return "DOWN"

def check_one(entry):
    url = entry.get("url")
    name = entry.get("name")
    # Percent-encode non-ASCII chars in path (fixes DOWN on URLs like méthodologie, adhérence)
    try:
        p = urlparse(url)
        if any(ord(c) > 127 for c in p.path):
            encoded_path = quote(p.path, safe="/")
            url = urlunparse((p.scheme, p.netloc, encoded_path, p.params, p.query, p.fragment))
    except Exception:
        pass
    t0 = time.time()
    code = 0
    try:
        # First try GET with Range (PHP handles GET better than HEAD; Range avoids loading full body)
        req = Request(url, method="GET", headers={"User-Agent": "WEVIA-HealthCheck/2.0", "Range": "bytes=0-0"})
        try:
            resp = urlopen(req, timeout=TIMEOUT)
            code = resp.status
        except HTTPError as e:
            code = e.code
            # If HEAD rejected (405/501), retry GET with Range header
            if code in (405, 501):
                try:
                    req2 = Request(url, method="GET", headers={"User-Agent": "WEVIA-HealthCheck/2.0", "Range": "bytes=0-0"})
                    resp2 = urlopen(req2, timeout=TIMEOUT)
                    code = resp2.status
                except HTTPError as e2:
                    code = e2.code
                except Exception:
                    pass
    except (URLError, Exception):
        code = 0
    elapsed_ms = int((time.time() - t0) * 1000)
    status = classify(code, elapsed_ms)
    return {"name": name, "url": url, "server": entry.get("server"), "status": status, "code": code, "ms": elapsed_ms}

def main():
    data = extract_data()
    if not data:
        print("NO_DATA")
        sys.exit(1)
    print(f"Checking {len(data)} screens, {WORKERS} workers, timeout {TIMEOUT}s")
    t0 = time.time()
    results = []
    with ThreadPoolExecutor(max_workers=WORKERS) as ex:
        futures = {ex.submit(check_one, e): e for e in data}
        for fut in as_completed(futures):
            try:
                results.append(fut.result())
            except Exception:
                e = futures[fut]
                results.append({"name": e.get("name"), "url": e.get("url"), "server": e.get("server"), "status": "ERROR", "code": 0, "ms": 0})
    elapsed = time.time() - t0
    counts = {}
    for r in results:
        counts[r["status"]] = counts.get(r["status"], 0) + 1
    out = {
        "generated_at": time.strftime("%Y-%m-%dT%H:%M:%S%z"),
        "elapsed_sec": round(elapsed, 1),
        "total": len(results),
        "counts": counts,
        "by_url": {r["url"]: {"status": r["status"], "code": r["code"], "ms": r["ms"]} for r in results}
    }
    tmp = tempfile.NamedTemporaryFile(mode="w", delete=False, dir="/var/www/html/api/", suffix=".tmp", encoding="utf-8")
    json.dump(out, tmp, separators=(",", ":"))
    tmp.close()
    os.replace(tmp.name, OUT_FILE)
    os.chmod(OUT_FILE, 0o644)
    summary = " ".join(f"{k}={v}" for k, v in sorted(counts.items()))
    print(f"DONE in {elapsed:.1f}s: {summary}")
    print(f"Written: {OUT_FILE}")

if __name__ == "__main__":
    main()