HHW Brick: Heating Hot Water System Brick Schema Toolkit

A Python package for converting heating hot water system data to Brick Schema models with comprehensive validation and portable analytics.

📚 Documentation | 🚀 Getting Started | 💡 Examples | 🐛 Issues

---

Description

This package provides tools for converting building hot water system data to Brick Schema models and running portable analytics applications.

Core Contributions:

• CSV-to-Brick Converter: Automated conversion from tabular BMS data to Brick Schema 1.4 RDF models
• Multi-Level Validators: Ontology, point count, equipment count, and structural pattern validation
• Portable Analytics: Building-agnostic applications that use SPARQL to auto-discover required sensors

Key Benefits:

• Interoperability: Standardized semantic models work across different BMS platforms
• Portability: Write analytics once, run on any qualified building without recoding
• Quality Assurance: Comprehensive validation ensures model correctness

The package supports five hot water system types (condensing boilers, non-condensing boilers, generic boilers, district hot water, district steam) and has been tested on 216 real buildings.

---

Usage

Installation

For Users (when published to PyPI):

pip install hhw-brick

For Development (current method):

# Clone the repository
git clone https://github.com/CenterForTheBuiltEnvironment/HHW_brick.git
cd HHW_brick

# Install in editable mode
pip install -e .

The -e flag installs the package in editable mode, so changes to the source code are immediately reflected without reinstalling.

System Requirements:

• Python 3.8 or higher
• All dependencies are automatically installed (see pyproject.toml)

Input Data: For sample input data format, see: https://doi.org/10.5061/dryad.t4b8gtj8n

---

📚 Documentation

📖 View Full Documentation →

For comprehensive guides, tutorials, and API reference, please visit our documentation site:

• Getting Started Guide
• User Guide - Conversion
• User Guide - Validation
• User Guide - Applications
• Examples

---

Basic Workflow

The typical workflow consists of three steps: conversion, validation, and application.

1. Convert CSV to Brick Model

from hhw_brick import CSVToBrickConverter

converter = CSVToBrickConverter()
graph = converter.convert_to_brick(
    metadata_csv="metadata.csv",
    vars_csv="vars_available_by_building.csv",
    building_tag="105",
    output_path="building_105.ttl"
)
print(f"Generated {len(graph)} RDF triples")

2. Validate the Brick Model

from hhw_brick import BrickModelValidator, GroundTruthCalculator

# Generate ground truth from input CSV (expected counts)
calculator = GroundTruthCalculator()
ground_truth = calculator.calculate(
    metadata_csv="metadata.csv",
    vars_csv="vars_available_by_building.csv",
    output_csv="ground_truth.csv"
)

# Validate generated model
validator = BrickModelValidator(ground_truth_csv_path="ground_truth.csv")

# Ontology validation
result = validator.validate_ontology("building_105.ttl")
print(f"Valid: {result['valid']}")

# Point count validation
point_result = validator.validate_point_count("building_105.ttl")

# Equipment count validation
equip_result = validator.validate_equipment_count("building_105.ttl")

3. Run Analytics Application

from hhw_brick import apps

# Discover available applications
available_apps = apps.list_apps()

# Load temperature difference analysis
app = apps.load_app("secondary_loop_temp_diff")

# Run on qualified building
results = app.run(
    brick_model="building_105.ttl",
    timeseries_data="building_105_data.csv"
)

Batch Processing

Process multiple buildings in parallel:

from hhw_brick import BatchConverter, BrickModelValidator

# Batch conversion
batch = BatchConverter()
results = batch.convert_all_buildings(
    metadata_csv="metadata.csv",
    vars_csv="vars_available_by_building.csv",
    output_dir="output/",
    show_progress=True  # Show progress bar
)

# Batch validation
validator = BrickModelValidator()
validation_results = validator.batch_validate_ontology(
    test_data_dir="output/",
    max_workers=4
)

Examples

The examples/ folder contains 8 complete examples:

1. 01_convert_csv_to_brick.py - Single and batch CSV conversion
2. 02_ontology_validation.py - SHACL-based ontology validation
3. 03_point_count_validation.py - Sensor/point count verification
4. 04_equipment_count_validation.py - Equipment count verification
5. 05_subgraph_pattern_matching.py - Structural pattern validation
6. 06_application_management.py - Discover and manage analytics apps
7. 07_run_application.py - Run analytics on single building
8. 08_batch_run_application.py - Batch run analytics on multiple buildings

Run any example:

python examples/01_convert_csv_to_brick.py

---

Input Data Format

Two CSV files are required:

1. metadata.csv - Building configuration (system type, equipment counts, building ID)
2. vars_available_by_building.csv - Sensor availability matrix (building ID × sensor names)

See example files in examples/fixtures/ for format details.

---

Output Format

The converter generates Brick Schema 1.4 models in Turtle (.ttl) format with:

• Building and system hierarchy (rec:Building → brick:Hot_Water_System → brick:Hot_Water_Loop)
• Equipment instances (brick:Boiler, brick:Pump)
• Sensor points (brick:Temperature_Sensor, brick:Flow_Sensor, etc.)
• Semantic relationships (brick:hasPart, brick:feeds, brick:isPointOf)

Example:

@prefix brick: <https://brickschema.org/schema/Brick#> .
@prefix hhws: <https://hhws.example.org#> .

hhws:building105 a rec:Building ;
    rec:isLocationOf hhws:building105.hws .

hhws:building105.hws a brick:Hot_Water_System ;
    brick:hasPart hhws:building105.hws.primary_loop .

---

Key Features

1. CSV-to-Brick Conversion

• Automatic system type detection
• Batch processing with parallel execution
• Support for 5 system types (condensing, non-condensing, generic boiler, district hot water, district steam)

2. Multi-Level Validation

• Ontology: SHACL validation against Brick Schema 1.4
• Point Count: Verify sensor counts (with owl:sameAs deduplication)
• Equipment Count: Validate boilers and pumps (with subclass support)
• Structural Pattern: SPARQL-based topology validation

3. Portable Analytics

Built-in applications that work across buildings:

• secondary_loop_temp_diff: Secondary loop temperature analysis
• primary_loop_temp_diff: Primary loop temperature analysis (boiler systems)

Applications use SPARQL queries to auto-discover sensors from Brick models, eliminating hardcoded point names.

---

Methods

The package implements a three-stage validation approach:

1. Syntactic Validation: RDF syntax and Brick Schema conformance (SHACL)
2. Semantic Validation: Point and equipment counts against ground truth (calculated from input CSV)
3. Structural Validation: Subgraph pattern matching for system topology

Ground truth values (expected counts) are computed independently from the input CSV data, not from the generated Brick model, ensuring unbiased validation.

---

Contact

Author: Mingchen Li
Email: liwei74123@gmail.com
Repository: https://github.com/CenterForTheBuiltEnvironment/HHW_brick
Issues: https://github.com/CenterForTheBuiltEnvironment/HHW_brick/issues

---

Version

Version: 0.1.0
Last Updated: 2025-01-01