aankitdas/resource-optimization-ml

1import sqlite32import pandas as pd3import os4 5print("Setting up SQLite Database...\n")6 7# Create/connect to database8db_path = 'resource_optimization.db'9conn = sqlite3.connect(db_path)10cursor = conn.cursor()11 12print(f"Connected to database: {db_path}\n")13 14# ==================== Load Services ====================15print("Loading services.csv...")16services_df = pd.read_csv('data/services.csv')17services_df.to_sql('services', conn, if_exists='replace', index=False)18print(f"Loaded {len(services_df)} services\n")19 20# ==================== Load Regional Latency ====================21print("Loading regional_latency.csv...")22latency_df = pd.read_csv('data/regional_latency.csv')23latency_df['timestamp'] = pd.to_datetime(latency_df['timestamp'])24latency_df.to_sql('regional_latency', conn, if_exists='replace', index=False)25print(f"Loaded {len(latency_df)} latency records\n")26 27# ==================== Load Traffic Patterns ====================28print("Loading traffic_patterns.csv...")29traffic_df = pd.read_csv('data/traffic_patterns.csv')30traffic_df['timestamp'] = pd.to_datetime(traffic_df['timestamp'])31traffic_df.to_sql('traffic_patterns', conn, if_exists='replace', index=False)32print(f"Loaded {len(traffic_df)} traffic records\n")33 34# ==================== Load Service Placement ====================35print("Loading service_placement.csv...")36placement_df = pd.read_csv('data/service_placement.csv')37placement_df['timestamp'] = pd.to_datetime(placement_df['timestamp'])38placement_df.to_sql('service_placement', conn, if_exists='replace', index=False)39print(f"Loaded {len(placement_df)} placement records\n")40 41# ==================== Create Indexes (for faster queries) ====================42print("Creating indexes for faster queries...")43cursor.execute('CREATE INDEX IF NOT EXISTS idx_service_id ON services(service_id)')44cursor.execute('CREATE INDEX IF NOT EXISTS idx_service_placement_service ON service_placement(service_id)')45cursor.execute('CREATE INDEX IF NOT EXISTS idx_traffic_service ON traffic_patterns(service_id)')46cursor.execute('CREATE INDEX IF NOT EXISTS idx_latency_regions ON regional_latency(region1, region2)')47print("Indexes created\n")48 49conn.commit()50 51# ==================== Verify Data ====================52print("="*60)53print("DATABASE SETUP COMPLETE!")54print("="*60)55 56# Show table info57tables = cursor.execute("SELECT name FROM sqlite_master WHERE type='table'").fetchall()58print(f"\nTables in database ({len(tables)}):")59for table in tables:60    count = cursor.execute(f"SELECT COUNT(*) FROM {table[0]}").fetchone()[0]61    print(f"   • {table[0]}: {count:,} rows")62 63 64conn.close()