Spectre.Operation.View (Spectre v0.3.0)

Privacy-safe, read-only projection of a committed operational loop.

wait_ref exposes only the current wait id, kind, and trigger generation so a caller can fence its reply. Progress, blocker, results, and artifacts are independently controlled by the Definition's publication policy.

Summary

Types

t()

Functions

from_loop(loop, control)

Projects a committed loop and its control plane into a flat, redacted view.

Types

t()

@type t() :: %Spectre.Operation.View{
  artifacts: term(),
  attempts: term(),
  blocker: term(),
  budget: term(),
  checkpoint: term(),
  cognitive: term(),
  context_revision: term(),
  control_state: term(),
  cycles: term(),
  definition: term(),
  definition_version: term(),
  id: term(),
  invalidations: term(),
  kind: term(),
  last_crash: term(),
  last_update: term(),
  metadata: term(),
  next_operation: term(),
  next_trigger: term(),
  observations: term(),
  operation: term(),
  partial_results: term(),
  pause_requested: term(),
  pending_command: term(),
  phase: term(),
  progress: term(),
  reconciliation: term(),
  retries: term(),
  revision: term(),
  status: term(),
  terminal_category: term(),
  trigger_generation: term(),
  updated_at: term(),
  wait_ref: term()
}

Functions

from_loop(loop, control)

@spec from_loop(Spectre.Operation.Loop.t(), Spectre.Operation.Control.t()) :: t()

Projects a committed loop and its control plane into a flat, redacted view.

Progress, partial results, artifacts, and blocker honor the loop's publication policy and become :redacted when publication is disabled; requests, updates, commands and cognitive data are reduced to safe fields.