Skip to main content

argumentation

The argumentation crate is the semantic core of the workspace. It provides two independent layers: abstract argumentation frameworks in the Dung 1995 tradition (with grounded, complete, preferred, stable, ideal, and semi-stable extensions, plus Caminada three-valued labellings), and structured argumentation in the ASPIC+ tradition (Modgil & Prakken 2014) with strict and defeasible rules, preference-based defeat resolution via last-link and weakest-link orderings, and automatic generation of a Dung AF. All other crates in this workspace build on top of it.

Crate: argumentation (crates.io · rustdoc)

Key types​

ArgumentationFramework<A>​

A Dung-style abstract argumentation framework generic over argument type A. Stores arguments and directed attack edges. Implements all canonical extension semantics as methods. → Full docs

aspic::StructuredSystem​

The top-level ASPIC+ entry point. Holds a knowledge base, rule set, and rule preference ordering, then constructs arguments, computes attacks, resolves defeats via the last-link principle, and emits a Dung AF. → Full docs

aspic::BuildOutput​

The combined output of StructuredSystem::build_framework: the constructed arguments, computed attacks, and the resulting abstract framework. Provides helpers like conclusions_in and argument_by_conclusion to work with the results. → Full docs

aspic::Literal​

An atomic proposition or its negation. Use Literal::atom("name") for positive literals and Literal::neg("name") for negative ones. The reserved __applicable_ prefix is used internally for undercut markers. → Full docs

Labelling​

A Caminada three-valued labelling assigning In, Out, or Undecided to each argument. Produced by ArgumentationFramework::complete_labellings. → Full docs

Key functions / methods​

FunctionReturnsWhat it does
ArgumentationFramework::grounded_extension(&self)HashSet<A>Computes the unique grounded extension (least complete extension). Always returns exactly one set.
ArgumentationFramework::preferred_extensions(&self)Result<Vec<HashSet<A>>>Enumerates all preferred extensions via subset search. Returns Err(TooLarge) above ENUMERATION_LIMIT.
ArgumentationFramework::stable_extensions(&self)Result<Vec<HashSet<A>>>Enumerates stable extensions. A stable extension attacks every argument outside it.
StructuredSystem::build_framework(&self)Result<BuildOutput>Single-pass construction: builds arguments, computes attacks, resolves defeats via preferences, emits the AF.
StructuredSystem::prefer_rule(preferred, less_preferred)Result<()>Records a rule preference (transitive closure maintained). Rejects reflexive and cyclic preferences.
BuildOutput::conclusions_in(&self, extension)HashSet<&Literal>Maps a set of ArgumentIds back to their conclusion literals.

Errors​

Error::ArgumentNotFound(String)​

An argument referenced in an operation was not in the framework. The payload is a Debug rendering of the missing argument.

Error::Parse(String)​

A parser (e.g. parsers::apx) failed to decode its input.

Error::Aspic(String)​

An ASPIC+ structural error: cyclic rule dependencies, illegal undercut markers, or cyclic/reflexive rule preferences.

Error::TooLarge { arguments, limit }​

Extension enumeration rejected because the framework exceeds ENUMERATION_LIMIT (22) arguments. The subset-search algorithm is O(2^n); use a smaller framework or wait for future SAT-based semantics.

See also​