tbnpy package

Submodules

Defining a model:

• tbnpy.variable
  • Overview
  • Quick start
  • Public API
    • Variable
      • Variable
      • Properties
      • Composite-state helpers
  • Hidden helpers (intentionally omitted)
• tbnpy.probability
  • Overview
  • The protocol
    • Discrete values
  • Reading an object
  • Checking a model
  • API
    • Probability
      • Probability.form()
      • Probability.form_details()
    • describe()
    • check_model()
    • ModelCheck
      • ModelCheck.ok
    • ModelIssue
    • ModelError
    • is_probability()
    • statement()
    • domain()
• Custom probability models
  • When to use CPT vs custom models
  • Minimal interface
  • Making the object readable
  • Checking the object
  • Device and performance notes
  • Examples in the ABCDE tutorial
• tbnpy.cpt
  • Overview
  • Quick start
  • Public API
    • Cpt
      • Cpt
      • Core data structures
      • Evidence
      • Sampling
      • Log-probability evaluation
  • Advanced / internal helpers
    • Cpt._get_C_binary()
    • Cpt.expand_and_check_compatibility()
    • Cpt.expand_and_check_compatibility_all()
  • Utility functions
    • get_names()
• tbnpy.spec
  • API
    • Node types
    • SpecError
    • VariableSpec
      • VariableSpec.kind
      • VariableSpec.states
      • VariableSpec.support
      • VariableSpec.bounds
      • VariableSpec.has_numeric_states
    • NodeSpec
      • NodeSpec.metadata
      • NodeSpec.child_names
      • NodeSpec.validate_against()
    • DistributionNode
    • EquationNode
    • DiscreteCptNode
      • DiscreteCptNode.validate_against()
    • DiscretiseNode
      • DiscretiseNode.validate_against()
    • CustomNode
      • CustomNode.child_names
    • AnalysisSpec
      • AnalysisSpec.options
    • ModelSpec
      • ModelSpec.validate()
      • ModelSpec.topological_order()
    • load_spec()
    • save_spec()
• tbnpy.compile
  • API
    • compile_model()
    • compile_file()
    • CompileError
    • FAMILIES
• tbnpy.expr
  • API
    • Expression
    • ExpressionError
    • NameSpec
      • NameSpec.kind
      • NameSpec.n_states
    • FUNCTION_NAMES

Running and checking it:

• tbnpy.inference
  • Overview
  • Glossary
  • Quick start
  • Public API
    • Topological utilities
      • get_ancestor_order()
    • Forward sampling without evidence
      • sample()
    • Forward sampling with evidence
      • sample_evidence_v0()
      • sample_evidence()
• tbnpy.adaptiveMH
  • Overview
  • Glossary
  • Utility functions
    • Variable type helpers
      • is_discrete()
      • is_continuous()
      • num_categories()
    • Factor construction utilities
      • build_Cs_3d()
      • factor_logp_2d()
      • build_factors_by_var()
  • Proposal kernels
    • Discrete proposals
      • propose_discrete_adaptive()
    • Continuous proposals
      • propose_continuous_rw_gaussian()
  • Adaptation configuration
    • AdaptConfig
  • HybridAdaptiveMH
    • HybridAdaptiveMH
    • Initialisation
      • HybridAdaptiveMH.init_state_from_forward_samples()
      • HybridAdaptiveMH.init_state_random()
    • Core MCMC update
      • HybridAdaptiveMH.mh_update_block()
    • Running the sampler
      • HybridAdaptiveMH.run()
• tbnpy.diagnostics
  • API
    • SenseCheckReport
    • run_sense_check()
    • forward_sample_all()
    • plot_report()
    • EvidenceIssue
    • check_evidence()
    • PosteriorResult
      • PosteriorResult.is_degenerate()
      • PosteriorResult.summary()
      • PosteriorResult.pmf()
    • posterior()
    • effective_sample_size()
    • weighted_quantile()

Module contents

import tbnpy gives tbnpy.__version__ and three names from tbnpy.probability: tbnpy.Probability, tbnpy.describe and tbnpy.check_model.