HOALedgerIQ_Website/agents/marketing-seo/scripts/seo-agent-v2.py

#!/usr/bin/env python3
"""
Marketing-SEO Agent v2 - With GA4 Integration
24/7 Monitoring: Site Uptime + Traffic Analytics
"""
import json
import time
import urllib.request
from datetime import datetime
from pathlib import Path
from google.analytics.data import BetaAnalyticsDataClient
from google.analytics.data_v1beta.types import RunReportRequest, DateRange, Metric

WORKSPACE = Path(__file__).parent.parent
LOG_DIR = WORKSPACE / "logs"
STATE_FILE = WORKSPACE / "state" / "agent-state.json"
CONFIG_DIR = WORKSPACE / "config"
GA_CREDS = CONFIG_DIR / "ga-credentials.json"
GA_PROPERTY = "526394825"

SITES = [
    "https://www.hoaledgeriq.com",
    "https://app.hoaledgeriq.com"
]
MONITOR_INTERVAL = 3600

LOG_DIR.mkdir(parents=True, exist_ok=True)

def log(msg):
    ts = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
    line = f"[{ts}] {msg}"
    print(line)
    log_file = LOG_DIR / f"seo-agent-{datetime.now().strftime('%Y%m%d')}.log"
    with open(log_file, 'a') as f:
        f.write(line + '\n')

def check_site(url):
    """Check if site is up"""
    start = time.time()
    try:
        req = urllib.request.Request(url, headers={"User-Agent": "SEO-Agent/1.0"})
        with urllib.request.urlopen(req, timeout=15) as r:
            return r.getcode() == 200, r.getcode(), round(time.time() - start, 2)
    except Exception as e:
        return False, str(e), None

def get_ga4_data():
    """Get GA4 traffic data"""
    try:
        client = BetaAnalyticsDataClient.from_service_account_json(str(GA_CREDS))
        request = RunReportRequest(
            property=f"properties/{GA_PROPERTY}",
            date_ranges=[DateRange(start_date="1daysAgo", end_date="today")],
            metrics=[
                Metric(name="sessions"),
                Metric(name="activeUsers"),
                Metric(name="newUsers"),
                Metric(name="bounceRate"),
                Metric(name="averageSessionDuration")
            ]
        )
        response = client.run_report(request)
        
        if response.rows:
            r = response.rows[0]
            return {
                "sessions": int(r.metric_values[0].value),
                "users": int(r.metric_values[1].value),
                "new_users": int(r.metric_values[2].value),
                "bounce_rate": float(r.metric_values[3].value),
                "avg_duration": float(r.metric_values[4].value)
            }
    except Exception as e:
        return {"error": str(e)}
    return {"sessions": 0, "users": 0, "new_users": 0}

def send_alert(title, message, severity="warning"):
    """Send Telegram alert"""
    log(f"🔔 ALERT [{severity}]: {title}")
    try:
        tg_msg = f"🔔 *SEO Alert: {title}*\n\n{message}\n\n⏰ {datetime.now().strftime('%H:%M')}"
        subprocess.run(["openclaw", "message", "send", "--text", tg_msg], 
                      capture_output=True, timeout=10)
    except:
        pass

def hourly_check():
    """Hourly monitoring: Sites + GA4"""
    log("=== Hourly Site + Traffic Check ===")
    
    site_status = {}
    for site in SITES:
        is_up, status, time_ms = check_site(site)
        site_status[site] = {"up": is_up, "status": status, "time_ms": time_ms}
        if is_up:
            log(f"✅ {site}: UP ({status}) - {time_ms}s")
        else:
            log(f"❌ {site}: DOWN ({status})")
            send_alert(f"SITE DOWN: {site}", f"Status: {status}", "critical")
    
    # GA4 traffic
    ga = get_ga4_data()
    if "error" not in ga:
        log(f"📊 GA4 Traffic: {ga.get('sessions',0)} sessions, {ga.get('users',0)} users")
    else:
        log(f"⚠️ GA4 Error: {ga.get('error')}")
    
    return {"sites": site_status, "ga4": ga}

def main():
    log("🚀 Marketing-SEO Agent v2 Started")
    log(f"Sites: {', '.join(SITES)}")
    log(f"GA4 Property: {GA_PROPERTY}")
    
    last_check = 0
    
    while True:
        now = datetime.now()
        now_ts = int(now.timestamp())
        
        if now_ts - last_check >= MONITOR_INTERVAL:
            hourly_check()
            last_check = now_ts
        
        time.sleep(60)

if __name__ == "__main__":
    main()

def daily_rank_check():
    """Daily check - if any keywords break into top 100, alert"""
    import re
    
    state_file = WORKSPACE / "state" / "rank-data.json"
    if not state_file.exists():
        return
    
    with open(state_file) as f:
        data = json.load(f)
    
    positions = data.get('positions', {})
    
    # Check if any are now ranked (non-null)
    ranked = sum(1 for p in positions.values() if p is not None)
    total = len(positions)
    
    if ranked > 0:
        log(f"📈 Rank Progress: {ranked}/{total} keywords now ranking")
        
        # Alert on new rankings
        report = "🎉 *RANKING PROGRESS!*\n\n"
        for kw, pos in positions.items():
            if pos:
                report += f"• {kw}: #{pos}\n"
        
        send_alert("New Rankings Detected!", report, "info")
    else:
        log(f"📊 SEO Status: ({ranked}/{total} keywords in top 100 - baseline phase)")

def get_monthly_milestone():
    """Return current SEO milestone based on launch date"""
    launch = datetime(2026, 3, 22)  # Launch date
    now = datetime.now()
    days_live = (now - launch).days
    
    if days_live < 30:
        return "Month 1: Focus on technical SEO + content creation"
    elif days_live < 90:
        return "Month 2-3: Target long-tail keywords, build backlinks"
    elif days_live < 180:
        return "Month 4-6: Optimize for primary keywords"
    else:
        return "Phase 2: Established - maintenance + expansion"

# Add to daily check