PDDLSharp 1.1.9

<p align="center"> <img src="https://github.com/kris701/PDDLSharp/assets/22596587/6c7c3516-bb1e-4713-ad17-e2eaff67107b" width="200" height="200" /> </p>

PDDLSharp

This is a package to make a PDDL parser, contextualiser, analyser, code generator and much more for C#. The parser is fully PDDL 2.2 compatible. The package can be found through Nuget or the Git package manager.

Parsers

There is a few parsers included in PDDLSharp, each of them will be explained in the subsections.

PDDL Parser

The PDDL Parser is the main part of PDDLSharp. Its a fully fledged parser that can parser up to PDDL 2.2 files.

A usage example of how to use the PDDL parser:

IErrorListener listener = new ErrorListener();
IParser<INode> parser = new PDDLParser(listener);
PDDLDecl decl = new PDDLDecl(
    parser.ParseAs<DomainDecl>(new FileInfo("domain.pddl")),
    parser.ParseAs<ProblemDecl>(new FileInfo("problem.pddl"))
)

To parse a file as a specific PDDL object:

IErrorListener listener = new ErrorListener();
IParser<INode> parser = new PDDLParser(listener);
ActionDecl decl = parser.ParseAs<ActionDecl>(new FileInfo("action-file.pddl"));

To contextualise a domain/problem:

IErrorListener listener = new ErrorListener();
IParser<INode> parser = new PDDLParser(listener);
PDDLDecl decl = new PDDLDecl(
    parser.ParseAs<DomainDecl>(new FileInfo("domain.pddl")),
    parser.ParseAs<ProblemDecl>(new FileInfo("problem.pddl"))
)
contextualiser.Contexturalise(decl);

To analyse a domain/problem (Note, the analyser will also contextualise the PDDL declaration, if it have not been contextualised):

IErrorListener listener = new ErrorListener();
IParser<INode> parser = new PDDLParser(listener);
IAnalyser analyser = new PDDLAnalyser(listener);
PDDLDecl decl = new PDDLDecl(
    parser.ParseAs<DomainDecl>(new FileInfo("domain.pddl")),
    parser.ParseAs<ProblemDecl>(new FileInfo("problem.pddl"))
)
analyser.Analyse(decl);

Plan Parser

There is also a plan parser that can parse Fast Downward output plans.

IErrorListener listener = new ErrorListener();
IParser<ActionPlan> parser = new FastDownwardPlanParser(listener);
ActionPlan plan = parser.Parse(new FileInfo("planFile.plan"));

SAS Parser

There is also a SAS parser that can parse Fast Downward intermediate output SAS format.

IErrorListener listener = new ErrorListener();
IParser<ISASNode> parser = new SASParser(listener);
SASDecl sas = parser.ParseAs<SASDecl>(new FileInfo("file.sas"));

Code Generators

PDDL Code Generator

To generate PDDL code from a PDDL declaration:

IErrorListener listener = new ErrorListener();
ICodeGenerator<INode> generator = new PDDLCodeGenerator(listener);
PDDLDecl decl = new PDDLDecl(...);
// If you want a "pretty" output, use:
// generator.Readable = true;
generator.Generate(decl.Domain, "domain.pddl");
generator.Generate(decl.Problem, "problem.pddl");

Plan Code Generator

To generate a Fast Downward plan from a ActionPlan declaration:

IErrorListener listener = new ErrorListener();
ICodeGenerator<ActionPlan> generator = new FastDownwardPlanGenerator(listener);
ActionPlan plan = new ActionPlan(...);
generator.Generate(plan, "planFile");

SAS Code Generator

To generate a Fast Downward intermediate output from a SASDecl declaration:

IErrorListener listener = new ErrorListener();
ICodeGenerator<ISASNode> generator = new SASCodeGenerator(listener);
SASDecl sas = new SASDecl(...);
generator.Generate(sas, "output.sas");

Toolkit

PDDLSharp includes a few tools, that can be found in the namespace PDDLSharp.Toolkit.

State Space Simulator

There is a State Space Simulator included with PDDLSharp. This is a simulator that is capable of simulating the state changes for each action execution. If there are invalid arguments or type issues, the simulator will throw an exception.

PDDLDecl declaration = new PDDLDecl(...);
IStateSpaceSimulator simulator = new StateSpaceSimulator(declaration);
simulator.Step("actionName", "obj1", "obj2");

Plan Validator

There is a simple plan validator included in PDDLSharp. It is capable of taking in a ActionPlan and a PDDLDecl and verify if the given plan is even possible or not.

IErrorListener listener = new ErrorListener();
IParser<ActionPlan> parser = new FastDownwardPlanParser(listener);
ActionPlan plan = parser.Parse(new FileInfo("planFile"));

PDDLDecl declaration = new PDDLDecl(...);
IPlanValidator validator = new PlanValidator();
validator.Validate(plan, declaration);

The Validate(...) method returns true if the plan is valid, false otherwise.

Mutex Detector

There is a simple predicate mutex detector included in PDDLSharp. It is able to find simple action predicate mutexes. You just give it a PDDLDecl and it will try and find mutex predicates in it:

IErrorListener listener = new ErrorListener();
IParser<INode> parser = new PDDLParser(listener);
PDDLDecl decl = new PDDLDecl(...)

IMutexDetectors detector = new EffectBalanceMutexes();
var mutexes = detector.FindMutexes(decl);

Static Predicate Detector

There is a simple static predicate detector included in PDDLSharp. It is able to find predicates that are not ever set in action effects. You just give it a PDDLDecl and it will try and find static predicates in it:

IErrorListener listener = new ErrorListener();
IParser<INode> parser = new PDDLParser(listener);
PDDLDecl decl = new PDDLDecl(...)

IStaticPredicateDetectors detector = new SimpleStaticPredicateDetector();
var predicates = detector.FindStaticPredicates(decl);

Predicate Grounder

There is a predicate grounder included in PDDLSharp. It is able to take in a predicate, and instanciate it with all possible valid combinations of objects (and constants).

PDDLDecl decl = new PDDLDecl(...)
IGrounder<PredicateExp> grounder = new PredicateGrounder(decl);
PredicateExp predicate = new PredicateExp(...);
List<PredicateExp> groundedPredicates = grounder.Ground(predicate);

IParametized Grounder

Another grounder that is included is one that can ground IParametized nodes. This is nodes such as ActionDecl, ForAllExp, ExistsExp, etc. It takes in an IParametized node and makes grounded copies of it.

PDDLDecl decl = new PDDLDecl(...)
IGrounder<IParametized> grounder = new ParametizedGrounder(decl);
ActionDecl action = new ActionDecl(...);
List<IParametized> groundedActions = grounder.Ground(action);

Sequential Macro Generator

PDDLSharp also have a simple sequential macro generator. It can generate lifted macros based on reoccuring sequences in ActionPlans

PDDLDecl decl = new PDDLDecl(...)
IMacroGenerator<List<ActionPlan>> generator = new SequentialMacroGenerator(decl);
List<ActionPlan> plans = new List<ActionPlan>(...);
List<ActionDecl> macros = generator.FindMacros(plans);

Supported Requirements

PDDLSharp supports a large set of requirements, all the way up to PDDL 2.2:

• STRIPS (:strips)
• Typing (:typing)
• Disjunctive Preconditions (:disjunctive-preconditions)
• Equality (:equality)
• Quantified Preconditions (:quantified-preconditions)
  • Existential Preconditions (:existential-preconditions)
  • Universal Preconditions (:universal-preconditions)
• Conditional Effects (:conditional-effects)
• Domain Axioms (:domain-axioms)
  • Subgoals Through Axioms (:subgoals-through-axioms)
  • Expression Evaluation (:expression-evaluation)
• ADL (:adl)
• Fluents (:fluents)
• Durative Actions (:durative-actions)
  • Durative Inequalities (:durative-inequalities)
  • Continuous Effects (:continuous-effects)
• Negative Preconditions (:negative-preconditions)
• Derived Predicates (:derived-predicates)
• Timed Initial Literals (:timed-initial-literals)
• Action Expansions (:action-expansions)
• Foreach Expansions (:forach-expansions)
• DAG Expansions (:dag-expansions)
• Safety Constraints (:safety-constraints)
• Open World (:open-world)
• True Negation (:true-negation)
• UCPOP (:ucpop)
• Constraints (:constraints)
• Preferences (:preferences)

Notes

The system tests uses the Downward Benchmark Set to test on. A repository with solved instances of the benchmarks are also used for the plan parsing and verification.