Core Concepts

These pages explain the fundamental abstractions and design decisions behind the framework. Understanding these concepts will make everything else – configuration, module development, postprocessing – much more straightforward.

• Architecture Overview
  • The Challenges
  • How the Architecture Addresses These
  • Key Abstractions
  • Data Flow
  • Configuration to Code Mapping
  • Caching in Detail
• Columns and Data
  • The Column Class
  • TrackedColumns
  • The Selection Pattern
  • Metadata
  • Pipeline Data
  • ColumnCollection
• Modules in Depth
  • Module Types
  • The Module Lifecycle
  • The run() Method
  • Declaring Inputs and Outputs
  • Producing Results
  • Configuration via attrs
• Results and Provenance
  • The Result Tree
  • Result Types
  • Loading Results
  • Merging and Scaling
  • Serialization Format
  • Checking Completeness
• Pattern Matching
  • Basic Patterns
  • Negation
  • Regex Patterns
  • Deep Patterns (Metadata Matching)
  • Compound Patterns
  • Capture and Grouping
  • Summary of Pattern Syntax
• Systematics
  • How Systematics Work
  • ParameterSpec
  • Run Builders
  • Driven Parameters
  • Weight vs Shape Systematics