Configuration Management

The DTPatternRecognition framework uses a centralized configuration system to manage all analysis settings, ensuring consistency and flexibility throughout your workflow. This is handled by a global configuration class, Config, which loads YAML files and exposes their contents as Python attributes.

The configuration is managed by a single global instance, RUN_CONFIG, imported from dtpr.base.config. This instance is initialized at startup with a specific YAML configuration file (typically run_config.yaml), and all top-level keys in this file become attributes of RUN_CONFIG. This design allows you to access and modify configuration parameters from anywhere in your codebase without hardcoding values.

The configuration file controls: - Particle and event initialization - Preprocessing and selection logic - Histogram definitions - Plotting styles and options - Any other analysis tool settings

Example: Accessing Configuration Settings

A snippet from run_config.yaml might look like:

particle_types:
  digis:
    amount: 'digi_nDigis'
    attributes:
      wh:
        branch: 'digi_wheel'
      BX:
        expr: 'time // 25 if time is not None else None'
    sorter:
      by: 'p.BX'

ntuple_preprocessors:
  genmuon_matcher:
    src: "dtpr.utils.genmuon_functions.analyze_genmuon_matches"

histo_names:
  - seg_eff_MB1
  - AM_rate_allBX_MB1
  - LeadingMuon_pt

plot_configs:
  mplhep-style: 'CMS'
  figure-configs:
    figure.dpi: 100

You can access these settings directly in your Python code:

from dtpr.base.config import RUN_CONFIG

# Accessing particle type definitions
digi_config = RUN_CONFIG.particle_types["digis"]
print(f"Digi particle amount source: {digi_config['amount']}")
print(f"Digi BX attribute expression: {digi_config['attributes']['BX']['expr']}")

# Accessing preprocessor definition
genmuon_matcher_src = RUN_CONFIG.ntuple_preprocessors["genmuon_matcher"]["src"]
print(f"Genmuon matcher function source: {genmuon_matcher_src}")

# Accessing list of histogram names to fill
print(f"First histogram to fill: {RUN_CONFIG.histo_names[0]}")

# Accessing plotting style
print(f"Matplotlib HEP style: {RUN_CONFIG.plot_configs['mplhep-style']}")

Output

Digi particle amount source: digi_nDigis
Digi BX attribute expression: time // 25 if time is not None else None
Genmuon matcher function source: dtpr.utils.genmuon_functions.analyze_genmuon_matches
First histogram to fill: seg_eff_MB1
Matplotlib HEP style: CMS

This demonstrates how RUN_CONFIG centralizes all settings, making them easily accessible throughout the application without hardcoding values or logic into Python files.

Changing the Configuration File

A key feature of RUN_CONFIG is its ability to switch between different configuration files at runtime. This is useful for running multiple analysis scenarios, testing different event setups, or experimenting with parameter variations.

You can change the configuration file using the change_config_file method:

from dtpr.base.config import RUN_CONFIG
import os

# Print initial config file and style
print(f"Initial config file: {RUN_CONFIG.path}")
print(f"Initial Matplotlib HEP style: {RUN_CONFIG.plot_configs['mplhep-style']}")

# Load a custom configuration file
custom_config_path = os.path.abspath("my_custom_config.yaml")
RUN_CONFIG.change_config_file(config_path=custom_config_path)
print(f"\nNew config file: {RUN_CONFIG.path}")
print(f"New Matplotlib HEP style: {RUN_CONFIG.plot_configs['mplhep-style']}")

# Switch back to the default config
RUN_CONFIG.change_config_file(config_path=os.path.abspath("dtpr/utils/yamls/run_config.yaml"))
print(f"\nSwitched back to default. Matplotlib HEP style: {RUN_CONFIG.plot_configs['mplhep-style']}")

Output:

Initial config file: .../dtpr/utils/yamls/run_config.yaml
Initial Matplotlib HEP style: CMS

New config file: .../my_custom_config.yaml
New Matplotlib HEP style: ATLAS

Switched back to default. Matplotlib HEP style: CMS

CLI Configuration

There is also a separate configuration instance, CLI_CONFIG, intended for internal use by the command-line interface tools. Regular users should interact only with RUN_CONFIG for analysis configuration. CLI_CONFIG is primarily for developers and advanced CLI customization.