import asyncio
import os
import asyncpg
import smtplib
import re # Ny import for tekst-vasking
from datetime import datetime
from email.mime.text import MIMEText
from email.mime.multipart import MIMEMultipart
from playwright.async_api import async_playwright
try:
    from playwright_stealth import stealth_async as apply_stealth
except ImportError:
    from playwright_stealth import stealth as apply_stealth

from apscheduler.schedulers.asyncio import AsyncIOScheduler
from dotenv import load_dotenv

load_dotenv()

DB_URL = "postgresql://teeoff_admin:teeoff_secret_password@db:5432/teeoff"

def clean_text(text):
    """Fjerner spesialtegn og normaliserer tekst for sammenligning"""
    return re.sub(r'[^a-zA-Z0-9æøåÆØÅ]', '', text).lower()

def interpret_status(text, keyword=None):
    t_raw = text.lower()
    
    if keyword:
        # Fuzzy match: Vi renser både tekst og søkeord for bindestreker/mellomrom
        k_clean = clean_text(keyword)
        
        # Hvis vi ikke finner søkeordet engang i renset form, gi opp
        if k_clean not in clean_text(t_raw):
            return "NOT_FOUND"
            
        # Hvis vi finner det, prøv å isolere teksten rundt det originale ordet
        # Vi leter etter det originale keywordet først
        parts = re.split(re.escape(keyword), t_raw, flags=re.IGNORECASE)
        if len(parts) > 1:
            t_raw = parts[1][:150]
        else:
            # Fallback hvis keywordet er delt av HTML-tagger (f.eks 18 <strong>hull</strong>)
            t_raw = t_raw[-200:] # Bruk slutten av teksten hvis ordet er vanskelig å isolere

    if any(word in t_raw for word in ["stengt", "lukket", "frost", "snø", "is", "closed", "stenger"]):
        return "stengt"
    if any(word in t_raw for word in ["vintergreen", "vintergrønn", "vinter"]):
        return "aapen_med_vintergreener"
    if any(word in t_raw for word in ["snart", "åpner kl"]):
        return "aapner_snart"
    if any(word in t_raw for word in ["åpen", "åpent", "aapen", "open"]):
        return "aapen"
    return "ukjent"

def send_report(changes, warnings):
    if not changes and not warnings: return
    subject = f"TeeOff Banestatus Rapport - {datetime.now().strftime('%d.%m.%Y')}"
    body = "BANESTATUS RAPPORT\n" + "="*30 + "\n\n"
    if changes: body += "✅ OPPDATERINGER:\n" + "\n".join(changes) + "\n\n"
    if warnings: body += "⚠️ MERKNADER / ADVARSLER:\n" + "\n".join(warnings) + "\n"
    
    msg = MIMEMultipart(); msg['From'] = os.getenv("SMTP_USER"); msg['To'] = os.getenv("EMAIL_TO"); msg['Subject'] = subject
    msg.attach(MIMEText(body, 'plain'))
    try:
        with smtplib.SMTP_SSL(os.getenv("SMTP_SERVER"), int(os.getenv("SMTP_PORT"))) as server:
            server.login(os.getenv("SMTP_USER"), os.getenv("SMTP_PASS"))
            server.send_message(msg)
        print("✅ Rapport sendt på e-post.")
    except Exception as e: print(f"❌ E-post feil: {e}")

async def run_daily_scraping():
    print(f"🚀 Starter sjekk {datetime.now().strftime('%H:%M:%S')}...")
    conn = await asyncpg.connect(DB_URL)
    facilities = await conn.fetch("SELECT id, name, scrape_status_url, scrape_status_selector FROM facilities WHERE scrape_status_url IS NOT NULL")
    
    changes, warnings = [], []

    async with async_playwright() as p:
        browser = await p.chromium.launch(headless=True)
        context = await browser.new_context()

        for f in facilities:
            page = await context.new_page()
            try: await apply_stealth(page)
            except: pass

            try:
                print(f"🔍 Besøker {f['name']}...")
                await page.goto(f['scrape_status_url'], timeout=60000, wait_until="networkidle")
                
                # Vent på at innholdet skal lande
                await asyncio.sleep(3)

                element = await page.query_selector(f['scrape_status_selector'])
                if not element:
                    warnings.append(f"❌ {f['name']}: Fant ikke elementet '{f['scrape_status_selector']}'")
                    continue
                
                full_text = await element.inner_text()
                await conn.execute("UPDATE facilities SET status_updated_at = CURRENT_DATE WHERE id = $1", f['id'])

                courses = await conn.fetch("SELECT id, name, status, scrape_keyword FROM courses WHERE facility_id = $1", f['id'])
                for c in courses:
                    new_status = interpret_status(full_text, c['scrape_keyword'])
                    
                    if new_status == "NOT_FOUND":
                        warnings.append(f"❓ {f['name']} ({c['name']}): Fant ikke søkeordet '{c['scrape_keyword']}' på siden.")
                        continue

                    old_status = c['status'] or "ukjent"
                    if new_status != old_status and new_status != "ukjent":
                        await conn.execute("UPDATE courses SET status = $1 WHERE id = $2", new_status, c['id'])
                        changes.append(f"🔹 {f['name']} ({c['name']}): {old_status.upper()} ➔ {new_status.upper()}")
                        print(f"✅ Oppdatert status for {f['name']} - {c['name']}")
                    else:
                        print(f"   - {c['name']}: Ingen endring ({new_status.upper()})")

            except Exception as e:
                warnings.append(f"🔥 {f['name']}: Feil: {str(e)[:100]}")
            finally:
                await page.close()
        await browser.close()
    
    await conn.close()
    send_report(changes, warnings)
    print("🏁 Ferdig.")

if __name__ == "__main__":
    asyncio.run(run_daily_scraping())