nur-muslim-companion/scripts/automation.py

#!/usr/bin/env python3
"""
FalahOS Sales Automation — Runs every 6 hours via cron.
Executes all MCP-driven tasks in a single Python process per MCP
to avoid session expiry issues.

Usage:
    python scripts/automation.py
    python scripts/automation.py --mautic-only
    python scripts/automation.py --gumroad-only
    python scripts/automation.py --ghost-only
"""

import os
import sys
import json
import csv
import time
import requests
from datetime import datetime
from urllib.parse import urljoin

# ─── Config ──────────────────────────────────────────────────
BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
MAUTIC_BASE = os.environ.get("MAUTIC_BASE_URL", "https://mautic.falahos.my").rstrip("/")
GUMROAD_TOKEN = os.environ.get("GUMROAD_ACCESS_TOKEN", "")
GHOST_KEY = os.environ.get("GHOST_ADMIN_API_KEY", "")
METRICS_FILE = os.path.join(BASE_DIR, "daily_metrics.jsonl")
QUEUE_FILE = os.path.join(BASE_DIR, "content_queue.json")
LEADS_FILE = os.path.join(BASE_DIR, "quality_leads.json")

# ─── Helpers ─────────────────────────────────────────────────
def log(msg):
    print(f"[{datetime.now().isoformat()}] {msg}")

def append_metrics(**kwargs):
    kwargs["timestamp"] = time.time()
    kwargs["date"] = datetime.now().isoformat()[:10]
    with open(METRICS_FILE, "a") as f:
        f.write(json.dumps(kwargs) + "\n")

# ─── 1. MAUTIC: Broadcast + Nurture ─────────────────────────
def run_mautic():
    """
    All Mautic operations in a single process to avoid session expiry.
    Uses web session login (not OAuth2 API).
    """
    log("Mautic: starting...")

    # Import server module in same process
    sys.path.insert(0, "/Users/wmj24/.hermes/mcp-servers/mautic-server")
    from server import _web_login, _session, BASE, API_TIMEOUT, _ensure_auth
    import re
    from bs4 import BeautifulSoup

    # Login
    result = _web_login()
    if not result.get("success"):
        log(f"Mautic: login failed — {result.get('error', 'unknown')}")
        return False
    log("Mautic: logged in")

    session = requests.Session()
    if _session.get("cookies"):
        session.cookies.update(_session["cookies"])

    # --- Get segments ---
    r = session.get(urljoin(BASE, "/s/segments"), timeout=API_TIMEOUT, verify=False)
    # Parse segment IDs from page
    segments = {}
    m = re.findall(r'/s/segments/(\d+)[^"]*">([^<]+)<', r.text)
    for seg_id, name in m:
        segments[name.strip()] = seg_id
    log(f"Mautic: found {len(segments)} segments")

    # --- Get contacts in high-value segment ---
    segment_id = None
    for name, sid in segments.items():
        if "high-value" in name.lower() or "islamic" in name.lower():
            segment_id = sid
            break

    if segment_id:
        # Check segment contact count
        r = session.get(urljoin(BASE, f"/s/segments/{segment_id}/contacts"),
                        timeout=API_TIMEOUT, verify=False)
        # Count contacts from page
        contact_count = r.text.count("contact-row") if "contact-row" in r.text else 0
        log(f"Mautic: segment has ~{contact_count} contacts")

    # --- Check Mautic stats ---
    auth_result = _ensure_auth()
    if auth_result.get("success"):
        # Try API stats
        r = session.get(urljoin(BASE, "/s/dashboard"), timeout=API_TIMEOUT, verify=False)
        log("Mautic: dashboard checked")

    # --- Send broadcast to high-value segment ---
    # Rotate through email templates based on day of week
    day_of_week = datetime.now().weekday()
    email_templates = [
        {"name": "Day0-Welcome", "subject": "Islamic Dev Toolkit 🕌"},
        {"name": "Day2-CaseStudy", "subject": "How Gading built quran-api"},
        {"name": "Day5-Tripwire", "subject": "Islamic UI Kit for $10"},
        {"name": "Day8-Objections", "subject": "I'm not a designer — neither were they"},
        {"name": "Day11-Core", "subject": "Complete FalahOS Stack $49"},
        {"name": "Day14-FAQ", "subject": "Quick questions?"},
        {"name": "Day18-Enterprise", "subject": "Building for an organization?"},
    ]
    template = email_templates[day_of_week % len(email_templates)]
    log(f"Mautic: would send '{template['name']}' — '{template['subject']}'")

    # Note: Actual email send requires segment_id + email_id
    # These are created once in mautic_setup.py and referenced here
    log("Mautic: broadcast logic ready (needs segment_id + email_id from setup)")
    return True


# ─── 2. GUMROAD: Cart Recovery + Revenue Check ─────────────
def run_gumroad():
    """Check abandoned carts and recover them. Also check daily revenue."""
    log("Gumroad: starting...")

    if not GUMROAD_TOKEN:
        log("Gumroad: no token set, skipping")
        return False

    headers = {"Authorization": f"Bearer {GUMROAD_TOKEN}"}

    # --- Check abandoned carts ---
    try:
        r = requests.get("https://api.gumroad.com/v2/abandoned_carts",
                         headers=headers, timeout=20)
        if r.status_code == 200:
            carts = r.json().get("abandoned_carts", [])
            recovered = 0
            for cart in carts:
                if not cart.get("recovery_email_sent"):
                    try:
                        r2 = requests.post(
                            f"https://api.gumroad.com/v2/abandoned_carts/{cart['id']}/send_recovery_email",
                            headers=headers, timeout=20
                        )
                        if r2.status_code == 200:
                            recovered += 1
                            log(f"  Recovered cart: {cart['id']} (${cart.get('amount', 0)/100:.2f})")
                    except Exception as e:
                        log(f"  Cart recovery error: {e}")
            log(f"Gumroad: recovered {recovered} carts")
        else:
            log(f"Gumroad: cart check returned {r.status_code}")
    except Exception as e:
        log(f"Gumroad: cart check error — {e}")

    # --- Check daily revenue ---
    try:
        r = requests.get("https://api.gumroad.com/v2/reports/overview",
                         headers=headers, timeout=20)
        if r.status_code == 200:
            data = r.json()
            revenue = data.get("total_revenue", 0)
            orders = data.get("total_sales", 0)
            log(f"Gumroad: revenue=${revenue/100:.2f}, orders={orders}")
            append_metrics(gumroad_revenue=revenue, gumroad_orders=orders)
        else:
            log(f"Gumroad: revenue check returned {r.status_code}")
    except Exception as e:
        log(f"Gumroad: revenue check error — {e}")

    return True


# ─── 3. GHOST: Publish Next Blog Post ──────────────────────
def run_ghost():
    """Publish the next post from content_queue.json."""
    log("Ghost: starting...")

    if not GHOST_KEY:
        log("Ghost: no API key set, skipping")
        return False

    if not os.path.exists(QUEUE_FILE):
        log("Ghost: no content queue found")
        return False

    with open(QUEUE_FILE) as f:
        queue = json.load(f)

    if not queue:
        log("Ghost: queue is empty")
        return False

    key_id, secret = GHOST_KEY.split(":", 1)
    import jwt

    token = jwt.encode(
        {"iat": int(time.time()), "exp": int(time.time()) + 300, "aud": "/v4/admin/"},
        bytes.fromhex(secret), algorithm="HS256", headers={"kid": key_id}
    )

    headers = {
        "Authorization": f"Ghost {token}",
        "Content-Type": "application/json",
    }

    post = queue.pop(0)
    data = {
        "posts": [{
            "title": post["title"],
            "html": post["html"],
            "status": "published",
            "tags": post.get("tags", []),
            "slug": post.get("slug", ""),
        }]
    }

    try:
        r = requests.post(
            "https://ghost.falahos.my/ghost/api/admin/posts/",
            headers=headers, json=data, timeout=20
        )
        if r.status_code == 201:
            log(f"Ghost: published '{post['title']}'")
        else:
            log(f"Ghost: publish failed ({r.status_code}) — {r.text[:200]}")
            # Put it back if it failed
            queue.insert(0, post)
    except Exception as e:
        log(f"Ghost: publish error — {e}")
        queue.insert(0, post)

    with open(QUEUE_FILE, "w") as f:
        json.dump(queue, f, indent=2)

    return True


# ─── 4. ANALYTICS: Snapshot ─────────────────────────────────
def run_analytics():
    """Capture daily metrics snapshot."""
    log("Analytics: capturing snapshot...")

    metrics = {}

    # Try Gumroad revenue
    if GUMROAD_TOKEN:
        try:
            headers = {"Authorization": f"Bearer {GUMROAD_TOKEN}"}
            r = requests.get("https://api.gumroad.com/v2/reports/overview",
                             headers=headers, timeout=10)
            if r.status_code == 200:
                data = r.json()
                metrics["gumroad_revenue"] = data.get("total_revenue", 0)
                metrics["gumroad_orders"] = data.get("total_sales", 0)
        except Exception:
            pass

    append_metrics(**metrics)
    log(f"Analytics: {json.dumps(metrics)}")
    return True


# ─── 5. SCOUT: Weekly Lead Scrape ───────────────────────────
def run_scout_scrape():
    """Scrape 10 new GitHub leads for Islamic tech. Run on Sundays only."""
    log("Scout: weekly lead scrape...")

    sys.path.insert(0, "/Users/wmj24/.hermes/mcp-servers/scout-server/src")
    from app.scrapers.github import scrape_profile
    from app.scrapers.enrichment import enrich_lead

    # Search for Islamic tech contributors
    search_queries = [
        "quran api language:javascript stars:>50",
        "prayer times language:javascript stars:>30",
        "islamic language:typescript stars:>20",
        "halal language:javascript stars:>20",
        "muslim app language:javascript stars:>50",
    ]

    new_leads = []
    for query in search_queries:
        try:
            r = requests.get(
                "https://api.github.com/search/repositories",
                params={"q": query, "per_page": "10", "sort": "stars"},
                headers={"Accept": "application/vnd.github.v3+json"},
                timeout=15
            )
            if r.status_code == 200:
                repos = r.json().get("items", [])
                for repo in repos:
                    owner = repo["owner"]["login"]
                    # Skip already known leads
                    known = [l.get("username") for l in _load_leads()]
                    if owner in known:
                        continue
                    try:
                        profile = scrape_profile(owner)
                        # Only keep if it looks relevant
                        bio = (profile.get("bio") or "").lower()
                        name = (profile.get("full_name") or "").lower()
                        if any(kw in bio + name for kw in ["muslim", "islam", "quran", "prayer", "halal", "arabic"]):
                            enriched = enrich_lead(profile)
                            new_leads.append(enriched)
                            log(f"  New lead: {owner} (score: {enriched.get('lead_score', 0)})")
                    except Exception as e:
                        log(f"  Error scraping {owner}: {e}")
        except Exception as e:
            log(f"  Search error for '{query}': {e}")

    # Save new leads
    if new_leads:
        existing = _load_leads()
        existing.extend(new_leads)
        with open(LEADS_FILE, "w") as f:
            json.dump(existing, f, indent=2)
        _export_csv(existing)
        log(f"Scout: added {len(new_leads)} new leads (total: {len(existing)})")
    else:
        log("Scout: no new leads found this week")

    return True


def _load_leads():
    if os.path.exists(LEADS_FILE):
        with open(LEADS_FILE) as f:
            return json.load(f)
    return []


def _export_csv(leads):
    csv_path = os.path.join(BASE_DIR, "quality_leads.csv")
    with open(csv_path, "w", newline="") as f:
        fieldnames = ["email", "firstname", "lastname", "company", "position",
                      "tags", "source", "profile_url", "website", "lead_score", "notes"]
        writer = csv.DictWriter(f, fieldnames=fieldnames)
        writer.writeheader()
        for lead in leads:
            row = lead.copy()
            row["tags"] = "; ".join(row.get("tags", []))
            writer.writerow({k: row.get(k, "") for k in fieldnames})


# ─── MAIN ────────────────────────────────────────────────────
def main():
    log("=" * 50)
    log("FalahOS Sales Automation Starting")
    log("=" * 50)

    # Always run these
    run_mautic()
    run_gumroad()
    run_ghost()
    run_analytics()

    # Weekly (Sunday = 6)
    if datetime.now().weekday() == 6:
        run_scout_scrape()

    log("=" * 50)
    log("FalahOS Automation Complete")
    log("=" * 50)


if __name__ == "__main__":
    main()