Struct DecisionRunnerAdapter 

pub struct DecisionRunnerAdapter { /* private fields */ }

Executes decision commands without moving policy into the runner.

Implementations

impl DecisionRunnerAdapter

pub fn new(executors: DecisionExecutorRegistry) -> Self

Creates an adapter backed by the supplied executor registry.

pub fn executors(&self) -> &DecisionExecutorRegistry

Returns the configured executor registry.

pub fn execution_class_for_command( &self, command: &DecisionLoopCommand, ) -> Result<DecisionExecutionClass, DecisionRunnerError>

Resolves the execution class of the executor that would run this command, using the same registry route authority as execution. Lets a host decide which resource-accounting boundary to apply before it reserves resources.

pub async fn execute_command( &self, command: &DecisionLoopCommand, knowledge: &KnowledgeBase, ) -> Result<DecisionEvidenceReceipt, DecisionRunnerError>

Resolves and executes one evidence-producing command.

The complete evidence batch is validated before it is atomically committed. Active requests capture their baseline immediately before executor invocation.

pub async fn execute_command_with_limits( &self, command: &DecisionLoopCommand, knowledge: &KnowledgeBase, limits: DecisionExecutionLimits, ) -> Result<DecisionEvidenceReceipt, DecisionRunnerError>

Resolves and executes one command under a host-owned resource allowance.

pub async fn drive_command( &self, decision_loop: &DecisionLoop, command: &DecisionLoopCommand, knowledge: &KnowledgeBase, experience: &mut ExperienceStore, session: &mut DecisionSession, ) -> Result<DecisionRunnerTurn, DecisionRunnerError>

Executes a command and resumes the matching decision-loop transition.

ExecuteAction submits passive evidence, CollectActiveEvidence submits the captured before/after snapshots, and Replan invokes the reasoner and utility planner. Terminal commands are returned unchanged. Adaptive, retry, active, and replan continuations fail before execution; use Self::drive_command_with_suppressed_actions to reauthorize them.

pub async fn drive_command_with_suppressed_actions( &self, decision_loop: &DecisionLoop, command: &DecisionLoopCommand, knowledge: &KnowledgeBase, experience: &mut ExperienceStore, session: &mut DecisionSession, host_suppressed_actions: &BTreeSet<String>, ) -> Result<DecisionRunnerTurn, DecisionRunnerError>

Executes and resumes a command under explicit current host policy.

Current suppressions are checked before executor work and remain in force through verification, adaptive authorization, and replanning.

pub async fn drive_command_with_limits( &self, decision_loop: &DecisionLoop, command: &DecisionLoopCommand, knowledge: &KnowledgeBase, experience: &mut ExperienceStore, session: &mut DecisionSession, limits: DecisionExecutionLimits, ) -> Result<DecisionRunnerTurn, DecisionRunnerError>

Drives one command under a host-owned execution allowance.

pub async fn drive_command_with_limits_and_suppressed_actions( &self, decision_loop: &DecisionLoop, command: &DecisionLoopCommand, knowledge: &KnowledgeBase, experience: &mut ExperienceStore, session: &mut DecisionSession, limits: DecisionExecutionLimits, host_suppressed_actions: &BTreeSet<String>, ) -> Result<DecisionRunnerTurn, DecisionRunnerError>

Drives one command under explicit host policy and execution allowance.

Current suppressions are checked before executor work and remain in force through verification, adaptive authorization, and replanning.