refactor(pi052): introduce generic language runtime

tests/policies/language_conditioned/test_runtime.py

@@ -0,0 +1,88 @@
+from lerobot.policies.language_conditioned import (
+    LanguageConditionedRuntime,
+    RuntimeState,
+    ToolCall,
+    VQAResult,
+)
+
+
+class FakeAdapter:
+    def __init__(self):
+        self.updated = False
+        self.text_calls = []
+
+    def select_action(self, observation, state):
+        assert observation == {"observation.state": 1}
+        assert state.task == "clean"
+        return ["a0", "a1"]
+
+    def select_text(self, kind, observation, state, user_text=None):
+        self.text_calls.append((kind, user_text))
+        return "new plan <say>ok</say>"
+
+    def parse_tool_calls(self, text):
+        assert text == "new plan <say>ok</say>"
+        return [ToolCall("say", {"text": "ok"})]
+
+    def answer_vqa(self, question, camera, observation, state):
+        return VQAResult(answer=f"answer: {question}")
+
+    def update_language_state(self, observation, state):
+        self.updated = True
+        state.set_context("subtask", "pick cup", label="subtask")
+
+
+class FakeTool:
+    def __init__(self):
+        self.calls = []
+
+    def call(self, args):
+        self.calls.append(args)
+
+
+def test_runtime_tick_updates_language_enqueues_and_dispatches_action():
+    adapter = FakeAdapter()
+    executed = []
+    runtime = LanguageConditionedRuntime(
+        policy_adapter=adapter,
+        observation_provider=lambda: {"observation.state": 1},
+        action_executor=executed.append,
+    )
+    runtime.set_task("clean")
+
+    logs = runtime.step_once()
+
+    assert adapter.updated
+    assert runtime.state.language_context["subtask"] == "pick cup"
+    assert executed == ["a0"]
+    assert list(runtime.state.action_queue) == ["a1"]
+    assert "  subtask: pick cup" in logs
+
+
+def test_runtime_handles_user_interjection_and_dispatches_tools():
+    adapter = FakeAdapter()
+    tool = FakeTool()
+    runtime = LanguageConditionedRuntime(
+        policy_adapter=adapter,
+        observation_provider=lambda: {"observation.state": 1},
+        tools={"say": tool},
+    )
+    runtime.set_task("clean")
+    runtime.state.extra["recent_interjection"] = "please say ok"
+    runtime.state.emit("user_interjection")
+
+    logs = runtime.step_once()
+
+    assert ("interjection", "please say ok") in adapter.text_calls
+    assert runtime.state.language_context["plan"] == "new plan <say>ok</say>"
+    assert tool.calls == [{"text": "ok"}]
+    assert "  speech: ok" in logs
+
+
+def test_runtime_state_aliases_legacy_keys_to_language_context():
+    state = RuntimeState()
+    state["current_subtask"] = "open drawer"
+    state["current_memory"] = "drawer open"
+
+    assert state.get("current_subtask") == "open drawer"
+    assert state.language_context == {"subtask": "open drawer", "memory": "drawer open"}

tests/policies/pi052/test_pi052_runtime_adapter.py

@@ -0,0 +1,54 @@
+from types import SimpleNamespace
+
+from lerobot.policies.language_conditioned import RuntimeState
+from lerobot.policies.pi052.inference.pi052_adapter import PI052PolicyAdapter, split_plan_and_say
+
+
+def test_pi052_adapter_builds_recipe_prompts_from_runtime_state():
+    adapter = PI052PolicyAdapter(policy=object())
+    state = RuntimeState(
+        task="clean the kitchen",
+        language_context={"memory": "cup moved", "plan": "pick then place"},
+        extra={"prior_subtask": "pick the cup"},
+    )
+
+    assert adapter.messages_for("subtask", state) == [{"role": "user", "content": "clean the kitchen"}]
+    assert adapter.messages_for("memory", state) == [
+        {"role": "user", "content": "clean the kitchen"},
+        {"role": "assistant", "content": "Previous memory: cup moved"},
+        {"role": "user", "content": "Completed subtask: pick the cup"},
+    ]
+    assert adapter.messages_for("interjection", state, user_text="wait") == [
+        {"role": "user", "content": "clean the kitchen"},
+        {"role": "assistant", "content": "Previous plan:\npick then place"},
+        {"role": "user", "content": "wait"},
+    ]
+    assert adapter.messages_for("vqa", state, user_text="where is the cup?") == [
+        {"role": "user", "content": "where is the cup?"}
+    ]
+
+
+def test_pi052_adapter_parses_say_tool_calls_and_plan_text():
+    adapter = PI052PolicyAdapter(policy=object())
+    text = "Move to the sink. <say>heading to the sink</say>"
+
+    assert split_plan_and_say(text) == ("Move to the sink.", "heading to the sink")
+    assert adapter.parse_tool_calls(text)[0].name == "say"
+    assert adapter.parse_tool_calls(text)[0].arguments == {"text": "heading to the sink"}
+    assert adapter.plan_from_text(text) == "Move to the sink."
+
+
+def test_pi052_runtime_cli_smoke_does_not_load_model(monkeypatch):
+    from lerobot.policies.pi052.inference import runtime_cli
+
+    fake_policy = SimpleNamespace(config=SimpleNamespace(device="cpu"))
+
+    monkeypatch.setattr(
+        runtime_cli,
+        "_load_policy_and_preprocessor",
+        lambda policy_path, dataset_repo_id: (fake_policy, None, None, None),
+    )
+    monkeypatch.setattr(runtime_cli, "_build_tools", lambda no_tts, tts_voice: {})
+    monkeypatch.setattr(runtime_cli, "_run_repl", lambda runtime, initial_task, max_ticks: 0)
+
+    assert runtime_cli.main(["--policy.path=fake", "--no_robot", "--task=clean", "--max_ticks=0"]) == 0