🗄️ Step 8: Database Setup¶

After hardware detection, we need to initialize the database and load baseline measurements.

📊 Create Database Tables¶

Tables are created automatically on first run, but you can verify:

# Check if database exists
ls -la data/experiments.db

# List all tables
sqlite3 data/experiments.db ".tables"

Expected tables:

experiments
hardware_config
environment_config
runs
energy_samples
cpu_samples
interrupt_samples
thermal_samples
orchestration_events
llm_interactions
orchestration_tax_summary
idle_baselines
task_categories

⏱️ Measure Idle Baseline¶

This measures your system's idle power consumption:

python -c "
from core.energy_engine import EnergyEngine
from core.config_loader import ConfigLoader
from core.utils.baseline_manager import BaselineManager

engine = EnergyEngine(ConfigLoader().get_hardware_config())
baseline = engine.measure_idle_baseline(duration_seconds=10, num_samples=3)

bm = BaselineManager()
bm.save(baseline)

print(f'✅ Baseline saved: {baseline.baseline_id}')
print(f'   Package idle power: {baseline.power_watts.get(\"package-0\", 0):.3f} W')
"

Important: Keep your system idle during measurement (no mouse/keyboard/movement, close background apps).

💾 Load Hardware & Environment to Database¶

python scripts/load_configs_to_db.py

This creates config/current_session.json with your:

hw_id (hardware fingerprint)
env_id (environment fingerprint)

✅ Verify Database Setup¶

# Check hardware record
sqlite3 data/experiments.db "SELECT hw_id, cpu_model, hardware_hash FROM hardware_config;"

# Check environment record
sqlite3 data/experiments.db "SELECT env_id, git_commit, env_hash FROM environment_config;"

# Check baseline
sqlite3 data/experiments.db "SELECT baseline_id, package_power_watts FROM idle_baselines ORDER BY timestamp DESC LIMIT 1;"

🔄 Next Step¶

Proceed to Model Configuration to set up your LLM API keys.