Skip to content

Example 08: Batch Run Application

Run analytics across multiple buildings in parallel.

What You'll Learn

  • ✅ Batch process multiple buildings
  • ✅ Parallel execution
  • ✅ Aggregate results
  • ✅ Handle qualification failures

Workflow

graph TD
    A[Load App] --> B[Discover Buildings]
    B --> C{For Each Building}
    C --> D[Check Qualification]
    D -->|Qualified| E[Run Analysis]
    D -->|Not Qualified| F[Skip]
    E --> G[Collect Results]
    F --> G
    G --> H[Aggregate Results]

    style A fill:#e3f2fd
    style C fill:#fff9c4
    style E fill:#c8e6c9
    style H fill:#c8e6c9

Complete Example

from pathlib import Path
from hhw_brick import apps

# Load application
app = apps.load_app("secondary_loop_temp_diff")

# Find all Brick models
brick_models = list(Path("tests/fixtures/Brick_Model_File").glob("*.ttl"))

print(f"Found {len(brick_models)} buildings")

# Batch run with parallel processing
results = []
qualified_count = 0

for model in brick_models:
    # Check qualification
    qualified, details = app.qualify(str(model))

    if qualified:
        qualified_count += 1
        # Run analysis
        result = app.analyze(
            brick_model_path=str(model),
            timeseries_path=f"timeseries/{model.stem}.csv",
            output_dir="results/"
        )
        results.append(result)

print(f"\n✓ Batch analysis complete!")
print(f"  Qualified: {qualified_count} / {len(brick_models)}")
print(f"  Analyzed: {len(results)} buildings")

Output: 

Found 10 buildings

Processing buildings:
  ✓ building_29: Qualified, analyzed
  ✗ building_34: Not qualified (missing flow sensor)
  ✓ building_53: Qualified, analyzed
  ✓ building_55: Qualified, analyzed
  ...

✓ Batch analysis complete!
  Qualified: 8 / 10
  Analyzed: 8 buildings
  Results saved to: results/

Aggregated Results

Combine results from multiple buildings:

import pandas as pd

# Load all result files
all_results = []
for result_file in Path("results/").glob("*.csv"):
    df = pd.read_csv(result_file)
    df['building'] = result_file.stem
    all_results.append(df)

# Combine
combined = pd.concat(all_results, ignore_index=True)
combined.to_csv("results/combined_results.csv", index=False)

print(f"✓ Combined {len(all_results)} buildings")
print(f"  Total records: {len(combined)}")

Run Example

python examples/08_batch_run_application.py

Performance Tips

  • ⚡ Use parallel processing for large datasets
  • ⚡ Pre-check qualification to avoid wasted computation
  • ⚡ Process timeseries data in chunks

Next Steps

📂 Source Code: examples/08_batch_run_application.py