User Guide

In-depth coverage of SeQuant’s core concepts, from the expression tree and context/configuration through operators, coupled-cluster equation generation, canonicalization, cost-based optimization, batching, code generation, and I/O. New to SeQuant? Start with Getting started instead; come back here for the fuller picture once you’re past the basics.

• The Expression Tree
  • Expr and ExprPtr
  • Leaves: Tensor, Constant, Variable
  • Composition: Product and Sum
  • ResultExpr: naming an equation
• Context and Configuration
  • The core Context
  • Canonicalizer configuration
  • The MBPT Context
  • Scoped context changes
  • Detecting changes
• Operator Interface for Symbolic Many-Body Algebra
  • Motivation
  • mbpt::Operator Anatomy
  • Operator Registry
  • Normalization conventions
  • Lie Similarity Transformation
  • Examples
• Coupled-Cluster Class
  • Overview
  • Ansatz Options
  • Key Methods
  • Examples
  • Advanced Usage
• Canonicalization
  • Why it matters
  • How it works, briefly
• Spin Tracing
  • The mbpt::Spin quantum number
  • Tracing spin out of an expression
• Optimizing Evaluation Cost
  • What it does
  • Tuning
• Batched (Memory-Bounded) Evaluation
  • Vocabulary
  • BatchPolicy
  • Deciding what to batch
  • Executing a batched plan
• Code Generation
  • The Generator interface
  • Available backends
• I/O
  • LaTeX
  • Formatting and streams
  • Serialization
• Evaluation Trace Analysis
  • Setup
  • Reading evaluation steps
  • Reading cache interactions
  • Reading batched traces
  • Reading terms