WiSearch/ui/webengine.py

from flask import Flask, render_template, request, redirect, session, url_for, flash, Response
import psycopg2
from psycopg2 import extensions, extras
import requests
import re
import functools
import json
import urllib3
import os

urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)
extensions.register_type(extensions.UNICODE)
extensions.register_type(extensions.UNICODEARRAY)

app = Flask(__name__)
app.secret_key = 'wisearch_ultra_secret_key'
CONFIG_FILE = 'config.json'

def load_config():
    if os.path.exists(CONFIG_FILE):
        with open(CONFIG_FILE, 'r', encoding='utf-8') as f:
            return json.load(f)
    return {}

def save_config(cfg):
    with open(CONFIG_FILE, 'w', encoding='utf-8') as f:
        json.dump(cfg, f, indent=2, ensure_ascii=False)

def get_db_connection():
    cfg = load_config()
    db = cfg.get('database', {})
    conn = psycopg2.connect(
        host=db.get('host', '192.168.131.124'),
        port=db.get('port', '5432'),
        user=db.get('user', 'postgres'),
        password=db.get('password', 'postpass'),
        dbname=db.get('dbname', 'wisearch'),
        sslmode=db.get('sslmode', 'disable')
    )
    conn.set_client_encoding('UTF8')
    return conn

# ΑΥΤΟΜΑΤΗ ΔΗΜΙΟΥΡΓΙΑ ΝΕΩΝ ΠΙΝΑΚΩΝ (Blacklist & Submitted URLs)
def init_db():
    try:
        conn = get_db_connection(); cur = conn.cursor()
        cur.execute("CREATE TABLE IF NOT EXISTS blacklisted_domains (id SERIAL PRIMARY KEY, domain TEXT UNIQUE)")
        cur.execute("CREATE TABLE IF NOT EXISTS submitted_urls (id SERIAL PRIMARY KEY, url TEXT, submitted_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP)")
        conn.commit(); cur.close(); conn.close()
    except Exception as e: print("DB Init Error:", e)

init_db()

# --- Decorators ---
def login_required(f):
    @functools.wraps(f)
    def decorated_function(*args, **kwargs):
        if 'logged_in' not in session: return redirect(url_for('login'))
        return f(*args, **kwargs)
    return decorated_function

def admin_required(f):
    @functools.wraps(f)
    def decorated_function(*args, **kwargs):
        if session.get('user_role') != 'admin':
            flash('Δεν έχετε δικαιώματα πρόσβασης.')
            return redirect(url_for('admin_dashboard'))
        return f(*args, **kwargs)
    return decorated_function

# --- AI Helper ---
def expand_query_with_ai(original_query):
    if not original_query: return []
    cfg = load_config()
    creds = cfg.get('cf_credentials', [])
    
    # ΝΕΟ ΠΙΟ ΕΞΥΠΝΟ PROMPT: Απαγορεύει γενικούς όρους και αγνοεί τα παιχνίδια/ονόματα
    system_prompt = """You are a search assistant for a local network. 
    Return ONLY a JSON array of up to 2 related terms/synonyms. 
    If the query is a specific game (like minecraft, csgo), an app, a brand, or a proper noun, return an empty array []. 
    NEVER return generic words like 'Network', 'Interface', 'Server', or 'Internet'. 
    Return ONLY valid JSON. No markdown, no explanations."""
    
    payload = {
        "messages": [
            {"role": "system", "content": system_prompt}, 
            {"role": "user", "content": f"Query: {original_query}"}
        ]
    }
    
    for cred in creds:
        try:
            api_url = f"https://api.cloudflare.com/client/v4/accounts/{cred['account_id']}/ai/run/{cred['model']}"
            r = requests.post(api_url, json=payload, headers={"Authorization": f"Bearer {cred['api_token']}"}, timeout=2.5)
            if r.status_code == 200:
                content = re.sub(r'```json|```', '', r.json()['result']['response']).strip()
                return [w for w in json.loads(content) if len(str(w)) > 1]
        except: continue
    return []

# --- Public Routes ---
@app.route('/')
def home():
    return render_template('index.html')

# --- ΝΕΟ: AI Chat Endpoint για τη σελίδα αποτελεσμάτων ---
@app.route('/api/ai_chat', methods=['POST'])
def ai_chat():
    data = request.get_json()
    query = data.get('query', '')
    user_msg = data.get('message', '')
    context_data = data.get('context', [])
    
    # Φτιάχνουμε ένα κείμενο με τα top αποτελέσματα για να τα διαβάσει το AI
    ctx_str = "\n".join([f"- {item['title']}: {item['snippet']}" for item in context_data])
    
    cfg = load_config()
    creds = cfg.get('cf_credentials', [])
    if not creds:
        return {"response": "Δεν έχουν ρυθμιστεί τα AI Accounts στο Admin Panel."}
        
    # Το ΝΕΟ, πολύ πιο αυστηρό prompt
    system_prompt = f"""Είσαι ο 'WiSearch AI', ένας βοηθός της μηχανής αναζήτησης του δικτύου AWMN. 
Ο χρήστης έψαξε για: '{query}'.
Αποτελέσματα αναζήτησης:
{ctx_str}

ΑΥΣΤΗΡΕΣ ΟΔΗΓΙΕΣ:
1. Απαντας ΠΑΝΤΑ ΑΥΣΤΗΡΑ στην γλωσσα που σε ρωτησαν (Θα καταλαβαινεις την γλωσσα απο τις 4 προτες λεξεις)
3. ΜΗΝ παπαγαλίζεις/επαναλαμβάνεις τις οδηγίες (π.χ. μην γράφεις "Γεια σας, σύντομα και φιλικά" ή "Ελπίζω να βοήθησα").
4. Δώσε κατευθείαν την απάντηση/περίληψη, βασισμένος ΜΟΝΟ στα αποτελέσματα. Αν δεν ξέρεις, πες "Δεν προκύπτει από τα αποτελέσματα"."""
    
    payload = {
        "messages": [
            {"role": "system", "content": system_prompt},
            {"role": "user", "content": user_msg}
        ]
    }
    
    for cred in creds:
        try:
            api_url = f"https://api.cloudflare.com/client/v4/accounts/{cred['account_id']}/ai/run/{cred['model']}"
            r = requests.post(api_url, json=payload, headers={"Authorization": f"Bearer {cred['api_token']}"}, timeout=10)
            if r.status_code == 200:
                return {"response": r.json()['result']['response'].strip()}
        except:
            continue
            
    return {"response": "Παρουσιάστηκε σφάλμα κατά την επικοινωνία με το AI."}

@app.route('/submit_url', methods=['POST'])
def submit_url():
    url = request.form.get('url', '').strip()
    if url:
        try:
            conn = get_db_connection(); cur = conn.cursor()
            cur.execute("INSERT INTO submitted_urls (url) VALUES (%s)", (url,))
            conn.commit(); cur.close(); conn.close()
            flash('Το URL υποβλήθηκε επιτυχώς για σκανάρισμα!')
        except: pass
    return redirect(request.referrer or '/')

@app.route('/api/autocomplete')
def autocomplete():
    q = request.args.get('q', '').strip()
    if len(q) < 2: return Response("[]", mimetype='application/json')
    try:
        conn = get_db_connection(); cur = conn.cursor(cursor_factory=extras.RealDictCursor)
        cur.execute("SELECT search_query FROM activity_logs WHERE search_query ILIKE %s GROUP BY search_query ORDER BY COUNT(*) DESC LIMIT 5", (f"{q}%",))
        res = [r['search_query'] for r in cur.fetchall()]
        cur.close(); conn.close()
        return Response(json.dumps(res), mimetype='application/json')
    except: return Response("[]", mimetype='application/json')

@app.route('/cached/<int:record_id>')
def cached_view(record_id):
    try:
        conn = get_db_connection(); cur = conn.cursor(cursor_factory=extras.RealDictCursor)
        cur.execute("SELECT url, full_content, last_crawled FROM web_health WHERE id = %s", (record_id,))
        res = cur.fetchone()
        cur.close(); conn.close()
        if res:
            html = f"<div style='background:#ffeb3b; color:#000; padding:15px; font-family:sans-serif;'><strong>WiSearch Cached View</strong><br>Αυτή είναι η αποθηκευμένη μορφή της σελίδας <b>{res['url']}</b> όπως ήταν στις {res['last_crawled']}. Το πραγματικό site μπορεί να είναι offline.</div><hr><div style='padding:20px; font-family:sans-serif; line-height:1.6;'>{res['full_content']}</div>"
            return Response(html, mimetype='text/html')
    except: pass
    return "Not found", 404

@app.route('/go')
def go_to_url():
    target_url, record_id = request.args.get('url'), request.args.get('id')
    if target_url and record_id:
        try:
            conn = get_db_connection(); cur = conn.cursor()
            cur.execute("UPDATE web_health SET clicks = COALESCE(clicks, 0) + 1 WHERE id = %s", (record_id,))
            conn.commit(); cur.close(); conn.close()
        except: pass
        return redirect(target_url)
    return redirect('/')

@app.route('/proxy_media')
def proxy_media():
    url = request.args.get('url')
    if not url: return "No URL provided", 400
    try:
        r = requests.get(url, verify=False, timeout=5)
        if r.status_code == 200: return Response(r.content, mimetype=r.headers.get('Content-Type', 'image/jpeg'))
    except: pass
    return "Failed", 404

@app.route('/search')
def search():
    query = request.args.get('q', '').strip()
    tab = request.args.get('tab', 'web')
    category_filter = request.args.get('category', '')
    page = request.args.get('page', 1, type=int)
    offset = (page - 1) * 10
    if not query: return redirect('/')

    try:
        conn = get_db_connection(); cur = conn.cursor()
        cur.execute("INSERT INTO activity_logs (ip_address, user_id, search_query) VALUES (%s, %s, %s)", (request.remote_addr, session.get('user_id'), query))
        conn.commit(); cur.close(); conn.close()
    except: pass

    sponsored = []
    try:
        conn = get_db_connection(); cur = conn.cursor(cursor_factory=extras.RealDictCursor)
        cur.execute("SELECT * FROM sponsored_links WHERE is_active = TRUE AND keywords ILIKE %s", (f"%{query}%",))
        sponsored = cur.fetchall(); cur.close(); conn.close()
    except: pass

    patterns = expand_query_with_ai(query)
    results, image_results, video_results = [], [], []
    has_next = False
    available_categories = []

    try:
        conn = get_db_connection()
        cur = conn.cursor(cursor_factory=extras.RealDictCursor)
        search_term = f"%{query}%"

        if tab == 'web':
            cat_sql = """
                SELECT DISTINCT CASE WHEN dr.hostname ~* '^[a-z0-9-]+\\.awmn$' THEN 'Core Services' ELSE wh.category END as category
                FROM web_health wh
                JOIN dns_records dr ON wh.dns_record_id = dr.id
                WHERE dr.hostname NOT IN (SELECT domain FROM blacklisted_domains)
                AND (unaccent(wh.page_title) ILIKE unaccent(%s) 
                       OR dr.hostname ILIKE %s 
                       OR unaccent(wh.full_content) ILIKE unaccent(%s)
            """
            cat_params = [search_term, search_term, search_term]
            for p in patterns:
                cat_sql += " OR unaccent(wh.full_content) ILIKE unaccent(%s)"
                cat_params.append(f"%{p}%")
            cat_sql += ")"
            
            cur.execute(cat_sql, tuple(cat_params))
            available_categories = [r['category'] for r in cur.fetchall() if r['category']]

            sql = """
                WITH RankedResults AS (
                    SELECT wh.id, dr.hostname, wh.url, wh.page_title, 
                           CASE WHEN dr.hostname ~* '^[a-z0-9-]+\\.awmn$' THEN 'Core Services' ELSE wh.category END as category,
                           wh.ai_summary, wh.activity_status, wh.pagerank_score, wh.clicks,
                           COALESCE(SUBSTRING(wh.full_content, 1, 400), '') AS snippet,
                           (SELECT media_url FROM media_files WHERE source_url = wh.url AND media_type = 'image' LIMIT 1) as thumbnail_url,
                           (
                               CASE WHEN dr.hostname ~* '^[a-z0-9-]+\\.awmn$' THEN 50000 ELSE 0 END 
                               + COALESCE(wh.pagerank_score, 1.0) 
                               + (COALESCE(wh.clicks, 0) * 0.5)
                               + (COALESCE(EXTRACT(EPOCH FROM wh.last_crawled), 0) / 100000.0)
                           ) as total_score,
                           ROW_NUMBER() OVER(
                               PARTITION BY dr.hostname 
                               ORDER BY (
                                   CASE WHEN dr.hostname ~* '^[a-z0-9-]+\\.awmn$' THEN 50000 ELSE 0 END 
                                   + COALESCE(wh.pagerank_score, 1.0) 
                                   + (COALESCE(wh.clicks, 0) * 0.5)
                                   + (COALESCE(EXTRACT(EPOCH FROM wh.last_crawled), 0) / 100000.0)
                               ) DESC
                           ) as rn
                    FROM web_health wh
                    JOIN dns_records dr ON wh.dns_record_id = dr.id
                    WHERE dr.hostname NOT IN (SELECT domain FROM blacklisted_domains)
                    AND (unaccent(wh.page_title) ILIKE unaccent(%s) 
                           OR dr.hostname ILIKE %s 
                           OR unaccent(wh.full_content) ILIKE unaccent(%s)
            """
            params = [search_term, search_term, search_term]
            for p in patterns:
                sql += " OR unaccent(wh.full_content) ILIKE unaccent(%s)"
                params.append(f"%{p}%")
            sql += ")"
            
            if category_filter:
                sql += " AND (CASE WHEN dr.hostname ~* '^[a-z0-9-]+\\.awmn$' THEN 'Core Services' ELSE wh.category END) = %s "
                params.append(category_filter)

            sql += """
                )
                SELECT * FROM RankedResults WHERE rn = 1 ORDER BY total_score DESC LIMIT 11 OFFSET %s
            """
            params.append(offset)
            cur.execute(sql, tuple(params))
            fetched = cur.fetchall()
            if fetched:
                results = list(fetched)
                if len(results) > 10: has_next = True; results = results[:10]

        elif tab in ['images', 'videos']:
            m_type = 'image' if tab == 'images' else 'video'
            sql_media = "SELECT m.* FROM media_files m JOIN web_health wh ON m.source_url = wh.url JOIN dns_records dr ON wh.dns_record_id = dr.id WHERE m.media_type = %s AND dr.hostname NOT IN (SELECT domain FROM blacklisted_domains) AND (m.alt_text ILIKE %s OR m.source_url ILIKE %s"
            params_media = [m_type, search_term, search_term]
            for p in patterns:
                sql_media += " OR m.alt_text ILIKE %s"
                params_media.append(f"%{p}%")
            sql_media += ") LIMIT 51 OFFSET %s"
            params_media.append(offset)
            
            cur.execute(sql_media, tuple(params_media))
            fetched_media = cur.fetchall()
            if fetched_media:
                media_list = list(fetched_media)
                if len(media_list) > 50: has_next = True; media_list = media_list[:50]
                if tab == 'images': image_results = media_list
                else: video_results = media_list

        cur.close(); conn.close()
    except Exception as e: print(f"SEARCH ERROR: {e}")

    return render_template('results.html', results=results, image_results=image_results, video_results=video_results,
                           sponsored=sponsored, query=query, page=page, has_next=has_next, tab=tab, 
                           smart_query=", ".join(patterns), category=category_filter, available_categories=available_categories)

# --- Auth & Admin Routes ---
@app.route('/login', methods=['GET', 'POST'])
def login():
    if request.method == 'POST':
        u, p = request.form.get('username'), request.form.get('password')
        conn = get_db_connection(); cur = conn.cursor(cursor_factory=extras.RealDictCursor)
        cur.execute("SELECT * FROM admin_users WHERE username = %s AND password_hash = %s", (u, p))
        user = cur.fetchone(); cur.close(); conn.close()
        if user:
            session.update({'logged_in': True, 'user_id': user['id'], 'username': u, 'user_role': user.get('role', 'user')})
            return redirect(url_for('admin_dashboard'))
        flash('Λάθος στοιχεία πρόσβασης')
    return render_template('login.html')

@app.route('/logout')
def logout():
    session.clear()
    return redirect(url_for('home'))

@app.route('/admin')
@login_required
def admin_dashboard():
    conn = get_db_connection(); cur = conn.cursor(cursor_factory=extras.RealDictCursor)
    if session.get('user_role') == 'admin':
        cur.execute("SELECT s.*, u.username as owner_name FROM sponsored_links s LEFT JOIN admin_users u ON s.user_id = u.id ORDER BY s.id DESC")
    else:
        cur.execute("SELECT * FROM sponsored_links WHERE user_id = %s ORDER BY id DESC", (session.get('user_id'),))
    sponsored = cur.fetchall()
    
    analytics = {'top_queries': [], 'categories': [], 'blacklists': []}
    if session.get('user_role') == 'admin':
        cur.execute("SELECT search_query, COUNT(*) as count FROM activity_logs GROUP BY search_query ORDER BY count DESC LIMIT 10")
        analytics['top_queries'] = cur.fetchall()
        
        cur.execute("SELECT category, COUNT(*) as count FROM web_health WHERE category IS NOT NULL GROUP BY category")
        analytics['categories'] = cur.fetchall()
        
        cur.execute("SELECT * FROM blacklisted_domains ORDER BY id DESC")
        analytics['blacklists'] = cur.fetchall()

    cur.close(); conn.close()
    
    cfg = load_config()
    cf_creds = cfg.get('cf_credentials', []) if session.get('user_role') == 'admin' else []
    ai_prompt = cfg.get('ai_prompt', '') if session.get('user_role') == 'admin' else ''
    
    return render_template('admin_dashboard.html', sponsored=sponsored, cf_creds=cf_creds, ai_prompt=ai_prompt, analytics=analytics)

@app.route('/admin/sponsored/add', methods=['POST'])
@login_required
def add_sponsored():
    user_id = session.get('user_id')
    data = (request.form.get('title'), request.form.get('url'), request.form.get('description'), request.form.get('keywords'), user_id)
    try:
        conn = get_db_connection(); cur = conn.cursor()
        cur.execute("INSERT INTO sponsored_links (title, url, description, keywords, user_id) VALUES (%s, %s, %s, %s, %s)", data)
        conn.commit(); cur.close(); conn.close()
    except: pass
    return redirect(url_for('admin_dashboard'))

@app.route('/admin/sponsored/delete/<int:link_id>')
@login_required
def delete_sponsored(link_id):
    conn = get_db_connection(); cur = conn.cursor()
    if session.get('user_role') == 'admin': cur.execute("DELETE FROM sponsored_links WHERE id = %s", (link_id,))
    else: cur.execute("DELETE FROM sponsored_links WHERE id = %s AND user_id = %s", (link_id, session.get('user_id')))
    conn.commit(); cur.close(); conn.close()
    return redirect(url_for('admin_dashboard'))

@app.route('/admin/cf/add', methods=['POST'])
@login_required
@admin_required
def add_cf_cred():
    cfg = load_config(); creds = cfg.get('cf_credentials', [])
    new_cred = {"comment": request.form.get('comment', ''), "account_id": request.form.get('account_id', ''), "api_token": request.form.get('api_token', ''), "model": request.form.get('model', '')}
    if new_cred["account_id"]: creds.append(new_cred); cfg['cf_credentials'] = creds; save_config(cfg)
    return redirect(url_for('admin_dashboard'))

@app.route('/admin/cf/delete/<int:index>')
@login_required
@admin_required
def delete_cf_cred(index):
    cfg = load_config(); creds = cfg.get('cf_credentials', [])
    if 0 <= index < len(creds): del creds[index]; cfg['cf_credentials'] = creds; save_config(cfg)
    return redirect(url_for('admin_dashboard'))

@app.route('/admin/cf/prompt', methods=['POST'])
@login_required
@admin_required
def update_ai_prompt():
    cfg = load_config(); cfg['ai_prompt'] = request.form.get('ai_prompt', ''); save_config(cfg)
    return redirect(url_for('admin_dashboard'))

@app.route('/admin/blacklist/add', methods=['POST'])
@login_required
@admin_required
def add_blacklist():
    domain = request.form.get('domain', '').strip()
    if domain:
        try:
            conn = get_db_connection(); cur = conn.cursor()
            cur.execute("INSERT INTO blacklisted_domains (domain) VALUES (%s) ON CONFLICT DO NOTHING", (domain,))
            conn.commit(); cur.close(); conn.close()
        except: pass
    return redirect(url_for('admin_dashboard'))

@app.route('/admin/blacklist/delete/<int:b_id>')
@login_required
@admin_required
def delete_blacklist(b_id):
    conn = get_db_connection(); cur = conn.cursor()
    cur.execute("DELETE FROM blacklisted_domains WHERE id = %s", (b_id,))
    conn.commit(); cur.close(); conn.close()
    return redirect(url_for('admin_dashboard'))

@app.route('/admin/logs')
@login_required
@admin_required
def admin_logs():
    conn = get_db_connection(); cur = conn.cursor(cursor_factory=extras.RealDictCursor)
    cur.execute("SELECT l.*, u.username FROM activity_logs l LEFT JOIN admin_users u ON l.user_id = u.id ORDER BY l.timestamp DESC LIMIT 100")
    logs = cur.fetchall(); cur.close(); conn.close()
    return render_template('admin_logs.html', logs=logs)

@app.route('/admin/users')
@login_required
@admin_required
def admin_users():
    conn = get_db_connection(); cur = conn.cursor(cursor_factory=extras.RealDictCursor)
    cur.execute("SELECT id, username, role FROM admin_users ORDER BY id ASC")
    users = cur.fetchall(); cur.close(); conn.close()
    return render_template('admin_users.html', users=users)

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5000, debug=True)