# -*- coding: utf-8 -*-
"""stale_source_monitor.py — GOV-F1.16a

Standalone script that probes every SRC URL for liveness and content drift,
logging results to 03_Registers/SITE.csv and WARNING to daily_audit_log.md.

Run manually or on a schedule. NOT invoked during build_all.py.
"""
import os, sys, csv, hashlib, time, urllib.request

HERE = os.path.dirname(os.path.abspath(__file__))
ROOT = os.path.dirname(HERE)
SRC_CSV = os.path.join(ROOT, "03_Registers", "SRC.csv")
SITE_CSV = os.path.join(ROOT, "03_Registers", "SITE.csv")
LOG = os.path.join(HERE, "daily_audit_log.md")

def main():
    if not os.path.exists(SRC_CSV):
        print("SRC.csv not found — run build_all.py first")
        sys.exit(1)

    # Ensure SITE.csv exists with headers
    if not os.path.exists(SITE_CSV):
        with open(SITE_CSV, "w", newline="", encoding="utf-8") as f:
            csv.writer(f).writerow(["date_checked", "url", "status_code", "response_time_ms", "changed_flag"])

    rows = []
    with open(SRC_CSV, "r", encoding="utf-8") as f:
        r = csv.DictReader(f)
        for row in r:
            url = row.get("Source URL", "").strip()
            if url.startswith("http"):
                rows.append(row)

    today = time.strftime("%Y-%m-%d")
    new_records = []
    warnings = []

    for row in rows:
        url = row["Source URL"]
        try:
            req = urllib.request.Request(url, method="HEAD")
            start = time.time()
            resp = urllib.request.urlopen(req, timeout=30)
            elapsed_ms = int((time.time() - start) * 1000)
            status = resp.getcode()
            # Content drift: hash first 1KB of body (fall back to HEAD headers if body empty)
            try:
                body = resp.read(1024)
            except Exception:
                body = b""
            content_hash = hashlib.sha256(body).hexdigest()[:16]
            changed = False  # Would need persistent storage of previous hash to detect change
        except urllib.error.HTTPError as e:
            elapsed_ms = 0
            status = e.code
            content_hash = ""
        except Exception as e:
            elapsed_ms = 0
            status = "ERR: %s" % type(e).__name__
            content_hash = ""

        new_records.append([today, url, status, elapsed_ms, "N/A"])

        if str(status) != "200":
            msg = "WARNING: %s returned %s" % (url, status)
            warnings.append(msg)
            print(msg)

    # Append to SITE.csv
    with open(SITE_CSV, "a", newline="", encoding="utf-8") as f:
        w = csv.writer(f)
        for rec in new_records:
            w.writerow(rec)

    # Write warnings to audit log
    if warnings:
        with open(LOG, "a", encoding="utf-8") as f:
            f.write("\n## Stale-source monitor %s\n" % today)
            for w in warnings:
                f.write("- %s\n" % w)

    print("Checked %d URLs. Results appended to %s" % (len(new_records), SITE_CSV))

if __name__ == "__main__":
    main()