main

ai-engine/scripts/extract_training_data_colab.ipynb

@@ -0,0 +1,166 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": ["# Training Data Extraction — Google Colab\n", "SSH tunnel ile sunucuya bağlanır, DB'den 270K+ maç çeker, Drive'a kaydeder.\n"]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 1. Gerekli paketler\n",
+    "!pip install sshtunnel psycopg2-binary pandas numpy -q\n",
+    "print('Paketler hazır')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 2. Drive bağla\n",
+    "from google.colab import drive\n",
+    "drive.mount('/content/drive')\n",
+    "import os\n",
+    "DRIVE_DIR = '/content/drive/MyDrive/iddaai'\n",
+    "os.makedirs(DRIVE_DIR, exist_ok=True)\n",
+    "print('Drive hazır:', DRIVE_DIR)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 3. SSH private key upload\n",
+    "# Mac'te terminalde şunu çalıştır, çıktıyı kopyala:\n",
+    "#   cat ~/.ssh/id_ed25519\n",
+    "# Aşağıya yapıştır (BEGIN ve END satırları dahil)\n",
+    "\n",
+    "SSH_PRIVATE_KEY = \"\"\"-----BEGIN OPENSSH PRIVATE KEY-----\n",
+    "BURAYA_KEY_ICERIGINI_YAPISTIR\n",
+    "-----END OPENSSH PRIVATE KEY-----\"\"\"\n",
+    "\n",
+    "# Key dosyasına yaz\n",
+    "key_path = '/root/.ssh/id_ed25519'\n",
+    "os.makedirs('/root/.ssh', exist_ok=True)\n",
+    "with open(key_path, 'w') as f:\n",
+    "    f.write(SSH_PRIVATE_KEY.strip() + '\\n')\n",
+    "os.chmod(key_path, 0o600)\n",
+    "print('SSH key hazır')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 4. SSH Tunnel aç + DB bağlantısını test et\n",
+    "from sshtunnel import SSHTunnelForwarder\n",
+    "import psycopg2\n",
+    "\n",
+    "tunnel = SSHTunnelForwarder(\n",
+    "    ('95.70.252.214', 2222),\n",
+    "    ssh_username='haruncan',\n",
+    "    ssh_pkey=key_path,\n",
+    "    remote_bind_address=('localhost', 5432),\n",
+    "    local_bind_address=('localhost', 15432),\n",
+    ")\n",
+    "tunnel.start()\n",
+    "print(f'Tunnel açık: localhost:{tunnel.local_bind_port}')\n",
+    "\n",
+    "conn = psycopg2.connect(\n",
+    "    host='localhost',\n",
+    "    port=15432,\n",
+    "    dbname='iddaai_db',\n",
+    "    user='iddaai_user',\n",
+    "    password='IddaA1_S4crET!',\n",
+    ")\n",
+    "cur = conn.cursor()\n",
+    "cur.execute(\"SELECT COUNT(*) FROM matches WHERE status='FT' AND score_home IS NOT NULL\")\n",
+    "print(f'DB bağlantısı OK — FT maç sayısı: {cur.fetchone()[0]:,}')\n",
+    "conn.close()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 5. extract_training_data.py kodunu Drive'dan veya doğrudan çalıştır\n",
+    "# Önce repo'yu Drive'a kopyala (yoksa)\n",
+    "import subprocess\n",
+    "\n",
+    "REPO_DIR = f'{DRIVE_DIR}/ai-engine'\n",
+    "SCRIPT   = f'{REPO_DIR}/scripts/extract_training_data.py'\n",
+    "\n",
+    "if not os.path.exists(SCRIPT):\n",
+    "    print('Script bulunamadı — ai-engine klasörünü Drive a yükle:')\n",
+    "    print('  Yerel makinede: cp -r /Users/piton/Documents/GitHub/iddaai/iddaai-be/ai-engine ~/Google\\ Drive/MyDrive/iddaai/')\n",
+    "else:\n",
+    "    print('Script hazır:', SCRIPT)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 6. Extraction'ı çalıştır\n",
+    "import sys, os\n",
+    "sys.path.insert(0, REPO_DIR)\n",
+    "\n",
+    "# DB URL'i tunnel üzerinden ayarla\n",
+    "os.environ['DATABASE_URL'] = 'postgresql://iddaai_user:IddaA1_S4crET!@localhost:15432/iddaai_db'\n",
+    "\n",
+    "# Output CSV'yi Drive'a kaydet\n",
+    "OUTPUT_CSV = f'{DRIVE_DIR}/training_data_full.csv'\n",
+    "\n",
+    "# Script'i import et ve main'i çalıştır\n",
+    "import importlib.util\n",
+    "spec = importlib.util.spec_from_file_location('extract', SCRIPT)\n",
+    "mod  = importlib.util.load_from_spec(spec)\n",
+    "spec.loader.exec_module(mod)\n",
+    "\n",
+    "# OUTPUT_CSV'yi override et\n",
+    "mod.OUTPUT_CSV = OUTPUT_CSV\n",
+    "mod.TOP_LEAGUES_PATH = f'{DRIVE_DIR}/qualified_leagues.json'\n",
+    "\n",
+    "mod.main()\n",
+    "print(f'\\nKaydedildi: {OUTPUT_CSV}')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 7. Tunnel kapat\n",
+    "tunnel.stop()\n",
+    "print('Tunnel kapatıldı')\n",
+    "\n",
+    "# Dosya boyutunu kontrol et\n",
+    "size_mb = os.path.getsize(OUTPUT_CSV) / 1024 / 1024\n",
+    "import pandas as pd\n",
+    "df = pd.read_csv(OUTPUT_CSV, nrows=5)\n",
+    "print(f'CSV: {size_mb:.1f} MB')\n",
+    "print(f'Kolonlar: {len(df.columns)}')"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {"display_name": "Python 3", "language": "python", "name": "python3"},
+  "language_info": {"name": "python", "version": "3.10.0"}
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}