Merge pull request #6169 from ethereum-optimism/jg/game_agent

op-challenger: Add basic agent

op-challenger/fault/agent.go

+package fault
+
+import (
+	"context"
+	"sync"
+)
+
+type Agent struct {
+	mu        sync.Mutex
+	game      Game
+	solver    *Solver
+	trace     TraceProvider
+	responder Responder
+	maxDepth  int
+}
+
+func NewAgent(game Game, maxDepth int, trace TraceProvider, responder Responder) Agent {
+	return Agent{
+		game:      game,
+		solver:    NewSolver(maxDepth, trace),
+		trace:     trace,
+		responder: responder,
+		maxDepth:  maxDepth,
+	}
+}
+
+// AddClaim stores a claim in the local state.
+// This function shares a lock with PerformActions.
+func (a *Agent) AddClaim(claim Claim) error {
+	a.mu.Lock()
+	defer a.mu.Unlock()
+	return a.game.Put(claim)
+}
+
+// PerformActions iterates the game & performs all of the next actions.
+// Note: PerformActions & AddClaim share a lock so the responder cannot
+// call AddClaim on the same thread.
+func (a *Agent) PerformActions() {
+	a.mu.Lock()
+	defer a.mu.Unlock()
+	for _, pair := range a.game.ClaimPairs() {
+		_ = a.move(pair.claim, pair.parent)
+	}
+}
+
+// move determines & executes the next move given a claim pair
+func (a *Agent) move(claim, parent Claim) error {
+	move, err := a.solver.NextMove(claim)
+	if err != nil || move == nil {
+		return err
+	}
+	// TODO(CLI-4123): Don't send duplicate responses
+	// if a.game.IsDuplicate(move) {
+	// 	return nil
+	// }
+	return a.responder.Respond(context.TODO(), *move)
+}