fix: achieve goal if can't fail

Allow subgoals in triggers and empty plan

See merge request ics/sp/2025/n25b/pepperplus-cb!50

pyproject.toml

@@ -24,6 +24,7 @@ dependencies = [
   "sphinx-rtd-theme>=3.0.2",
   "tf-keras>=2.20.1",
   "torch>=2.8.0",
+  "tornado ; sys_platform == 'win32'",
   "uvicorn>=0.37.0",
 ]
 

src/control_backend/agents/bdi/agentspeak_generator.py

@@ -4,6 +4,7 @@ University within the Software Project course.
 © Copyright Utrecht University (Department of Information and Computing Sciences)
 """
 
+import logging
 from functools import singledispatchmethod
 
 from slugify import slugify
@@ -30,7 +31,6 @@ from control_backend.schemas.program import (
     BasicNorm,
     ConditionalNorm,
     EmotionBelief,
-    FaceBelief,
     GestureAction,
     Goal,
     InferredBelief,
@@ -67,6 +67,7 @@ class AgentSpeakGenerator:
     """
 
     _asp: AstProgram
+    logger = logging.getLogger(__name__)
 
     def generate(self, program: Program) -> str:
         """
@@ -106,7 +107,7 @@ class AgentSpeakGenerator:
         check if a keyword is a substring of the user's message.
 
         The generated rule has the form:
-        keyword_said(Keyword) :- user_said(Message) & .substring(Keyword, Message, Pos) & Pos >= 0
+        keyword_said(Keyword) :- user_said(Message) & .substring_case_insensitive(Keyword, Message, Pos) & Pos >= 0
 
         This enables the system to trigger behaviors based on keyword detection.
         """
@@ -118,7 +119,7 @@ class AgentSpeakGenerator:
             AstRule(
                 AstLiteral("keyword_said", [keyword]),
                 AstLiteral("user_said", [message])
-                & AstLiteral(".substring", [keyword, message, position])
+                & AstLiteral(".substring_case_insensitive", [keyword, message, position])
                 & (position >= 0),
             )
         )
@@ -134,7 +135,6 @@ class AgentSpeakGenerator:
         """
         self._add_reply_with_goal_plan()
         self._add_say_plan()
-        self._add_reply_plan()
         self._add_notify_cycle_plan()
 
     def _add_reply_with_goal_plan(self):
@@ -198,40 +198,6 @@ class AgentSpeakGenerator:
             )
         )
 
-    def _add_reply_plan(self):
-        """
-        Adds a plan for general reply actions.
-
-        This plan handles general reply actions where the agent needs to respond
-        to user input without a specific conversational goal. It:
-        1. Marks that the agent has responded this turn
-        2. Gathers all active norms
-        3. Generates a reply based on the user message and norms
-
-        Trigger: +!reply
-        Context: user_said(Message)
-        """
-        self._asp.plans.append(
-            AstPlan(
-                TriggerType.ADDED_GOAL,
-                AstLiteral("reply"),
-                [AstLiteral("user_said", [AstVar("Message")])],
-                [
-                    AstStatement(StatementType.ADD_BELIEF, AstLiteral("responded_this_turn")),
-                    AstStatement(
-                        StatementType.DO_ACTION,
-                        AstLiteral(
-                            "findall",
-                            [AstVar("Norm"), AstLiteral("norm", [AstVar("Norm")]), AstVar("Norms")],
-                        ),
-                    ),
-                    AstStatement(
-                        StatementType.DO_ACTION,
-                        AstLiteral("reply", [AstVar("Message"), AstVar("Norms")]),
-                    ),
-                ],
-            )
-        )
 
     def _add_notify_cycle_plan(self):
         """
@@ -269,6 +235,39 @@ class AgentSpeakGenerator:
             )
         )
 
+    def _add_stop_plan(self, phase: Phase):
+        """
+        Adds a plan to stop the program. This just skips to the end phase,
+        where there is no behavior defined.
+        """
+        self._asp.plans.append(
+            AstPlan(
+                TriggerType.ADDED_GOAL,
+                AstLiteral("stop"),
+                [AstLiteral("phase", [AstString(phase.id)])],
+                [
+                    AstStatement(
+                        StatementType.DO_ACTION,
+                        AstLiteral(
+                            "notify_transition_phase",
+                            [
+                                AstString(phase.id),
+                                AstString("end")
+                            ]
+                        )
+                    ),
+                    AstStatement(
+                        StatementType.REMOVE_BELIEF,
+                        AstLiteral("phase", [AstVar("Phase")]),
+                    ),
+                    AstStatement(
+                        StatementType.ADD_BELIEF,
+                        AstLiteral("phase", [AstString("end")])
+                    )
+                ]
+            )
+        )
+
     def _process_phases(self, phases: list[Phase]) -> None:
         """
         Processes all phases in the program and their transitions.
@@ -285,21 +284,6 @@ class AgentSpeakGenerator:
                 self._process_phase(curr_phase)
             self._add_phase_transition(curr_phase, next_phase)
 
-        # End phase behavior
-        # When deleting this, the entire `reply` plan and action can be deleted
-        self._asp.plans.append(
-            AstPlan(
-                type=TriggerType.ADDED_BELIEF,
-                trigger_literal=AstLiteral("user_said", [AstVar("Message")]),
-                context=[AstLiteral("phase", [AstString("end")])],
-                body=[
-                    AstStatement(
-                        StatementType.DO_ACTION, AstLiteral("notify_user_said", [AstVar("Message")])
-                    ),
-                    AstStatement(StatementType.ACHIEVE_GOAL, AstLiteral("reply")),
-                ],
-            )
-        )
 
     def _process_phase(self, phase: Phase) -> None:
         """
@@ -326,6 +310,9 @@ class AgentSpeakGenerator:
         for trigger in phase.triggers:
             self._process_trigger(trigger, phase)
 
+        # Add force transition to end phase
+        self._add_stop_plan(phase)
+
     def _add_phase_transition(self, from_phase: Phase | None, to_phase: Phase | None) -> None:
         """
         Adds plans for transitioning between phases.
@@ -501,9 +488,13 @@ class AgentSpeakGenerator:
             if isinstance(step, Goal):
                 subgoals.append(step)
 
-        if not goal.can_fail and not continues_response:
+        if not goal.can_fail:
             body.append(AstStatement(StatementType.ADD_BELIEF, self._astify(goal, achieved=True)))
 
+        if len(body) == 0:
+            self.logger.warning("Goal with no plan detected: %s", goal.name)
+            body.append(AstStatement(StatementType.EMPTY, AstLiteral("true")))
+
         self._asp.plans.append(AstPlan(TriggerType.ADDED_GOAL, self._astify(goal), context, body))
 
         self._asp.plans.append(
@@ -564,10 +555,10 @@ class AgentSpeakGenerator:
             )
         )
         for step in trigger.plan.steps:
-            body.append(self._step_to_statement(step))
             if isinstance(step, Goal):
-                step.can_fail = False  # triggers are continuous sequence
-                subgoals.append(step)
+                new_step = step.model_copy(update={"can_fail": False})  # triggers are sequence
+                subgoals.append(new_step)
+            body.append(self._step_to_statement(step))
 
         # Arbitrary wait for UI to display nicely
         body.append(
@@ -611,6 +602,7 @@ class AgentSpeakGenerator:
         - check_triggers: When no triggers are applicable
         - transition_phase: When phase transition conditions aren't met
         - force_transition_phase: When forced transitions aren't possible
+        - stop: When we are already in the end phase
         """
         # Trigger fallback
         self._asp.plans.append(
@@ -642,6 +634,16 @@ class AgentSpeakGenerator:
             )
         )
 
+        # Stop fallback
+        self._asp.plans.append(
+            AstPlan(
+                TriggerType.ADDED_GOAL,
+                AstLiteral("stop"),
+                [],
+                [AstStatement(StatementType.EMPTY, AstLiteral("true"))],
+            )
+        )
+
     @singledispatchmethod
     def _astify(self, element: ProgramElement) -> AstExpression:
         """
@@ -688,10 +690,6 @@ class AgentSpeakGenerator:
     def _(self, eb: EmotionBelief) -> AstExpression:
         return AstLiteral("emotion_detected", [AstAtom(eb.emotion)])
 
-    @_astify.register
-    def _(self, fb: FaceBelief) -> AstExpression:
-        return AstLiteral("face_present")
-
     @_astify.register
     def _(self, ib: InferredBelief) -> AstExpression:
         """

src/control_backend/agents/bdi/bdi_core_agent.py

@@ -176,6 +176,8 @@ class BDICoreAgent(BaseAgent):
                         self._force_norm(msg.body)
                     case "force_next_phase":
                         self._force_next_phase()
+                    case "stop":
+                        self._stop()
                     case _:
                         self.logger.warning("Received unknown user interruption: %s", msg)
 
@@ -335,6 +337,11 @@ class BDICoreAgent(BaseAgent):
 
         self.logger.info("Manually forced phase transition.")
 
+    def _stop(self):
+        self._set_goal("stop")
+
+        self.logger.info("Stopped the program (skipped to end phase).")
+
     def _add_custom_actions(self) -> None:
         """
         Add any custom actions here. Inside `@self.actions.add()`, the first argument is
@@ -342,6 +349,28 @@ class BDICoreAgent(BaseAgent):
         the function expects (which will be located in `term.args`).
         """
 
+        @self.actions.add(".substring_case_insensitive", 3)
+        @agentspeak.optimizer.function_like
+        def _substring(agent, term, intention):
+            """
+            Find out if a string is a substring of another (case insensitive). Copied mostly from
+            the agentspeak library method .substring.
+            """
+            needle = agentspeak.asl_str(agentspeak.grounded(term.args[0], intention.scope)).lower()
+            haystack = agentspeak.asl_str(agentspeak.grounded(term.args[1], intention.scope)).lower()
+
+            choicepoint = object()
+
+            pos = haystack.find(needle)
+            while pos != -1:
+                intention.stack.append(choicepoint)
+
+                if agentspeak.unify(term.args[2], pos, intention.scope, intention.stack):
+                    yield
+
+                agentspeak.reroll(intention.scope, intention.stack, choicepoint)
+                pos = haystack.find(needle, pos + 1)
+
         @self.actions.add(".reply", 2)
         def _reply(agent, term, intention):
             """
@@ -467,7 +496,6 @@ class BDICoreAgent(BaseAgent):
                 body=str(trigger_name),
             )
 
-            # TODO: check with Pim
             self.add_behavior(self.send(msg))
 
             yield

src/control_backend/agents/perception/visual_emotion_recognition_agent/visual_emotion_recognition_agent.py

@@ -14,7 +14,7 @@ from control_backend.agents.perception.visual_emotion_recognition_agent.visual_e
 )
 from control_backend.core.agent_system import InternalMessage
 from control_backend.core.config import settings
-from control_backend.schemas.belief_message import Belief, BeliefMessage
+from control_backend.schemas.belief_message import Belief
 
 
 class VisualEmotionRecognitionAgent(BaseAgent):
@@ -44,7 +44,6 @@ class VisualEmotionRecognitionAgent(BaseAgent):
         self.timeout_ms = timeout_ms
         self.window_duration = window_duration
         self.min_frames_required = min_frames_required
-        self._face_detected = False
 
         # Pause functionality
         # NOTE: flag is set when running, cleared when paused
@@ -90,9 +89,6 @@ class VisualEmotionRecognitionAgent(BaseAgent):
         # Tracks counts of detected emotions per face index
         face_stats = defaultdict(Counter)
 
-        # How many times a face has been detected
-        face_detection_yes_no = [0, 0]
-
         prev_dominant_emotions = set()
 
         while self._running:
@@ -101,8 +97,8 @@ class VisualEmotionRecognitionAgent(BaseAgent):
 
                 width, height, image_bytes = await self.video_in_socket.recv_multipart()
 
-                width = int.from_bytes(width, "little")
-                height = int.from_bytes(height, "little")
+                width = int.from_bytes(width, 'little')
+                height = int.from_bytes(height, 'little')
 
                 # Convert bytes to a numpy buffer
                 image_array = np.frombuffer(image_bytes, np.uint8)
@@ -111,13 +107,6 @@ class VisualEmotionRecognitionAgent(BaseAgent):
 
                 # Get the dominant emotion from each face
                 current_emotions = self.emotion_recognizer.sorted_dominant_emotions(frame)
-
-                # Update face face_detection_yes_no
-                if len(current_emotions) > 0:
-                    face_detection_yes_no[0] += 1
-                else:
-                    face_detection_yes_no[1] += 1
-
                 # Update emotion counts for each detected face
                 for i, emotion in enumerate(current_emotions):
                     face_stats[i][emotion] += 1
@@ -133,31 +122,18 @@ class VisualEmotionRecognitionAgent(BaseAgent):
                             dominant_emotion = counter.most_common(1)[0][0]
                             window_dominant_emotions.add(dominant_emotion)
 
-                    if (
-                        face_detection_yes_no[0] > face_detection_yes_no[1]
-                        and not self._face_detected
-                    ):
-                        self._face_detected = True
-                        await self._inform_face_detected()
-                    elif (
-                        face_detection_yes_no[0] <= face_detection_yes_no[1] and self._face_detected
-                    ):
-                        self._face_detected = False
-                        await self._inform_face_detected()
-
-                    face_detection_yes_no = [0, 0]
-
                     await self.update_emotions(prev_dominant_emotions, window_dominant_emotions)
                     prev_dominant_emotions = window_dominant_emotions
                     face_stats.clear()
                     next_window_time = time.time() + self.window_duration
 
             except zmq.Again:
-                self.logger.warning("No video frame received within timeout.")
+                pass
 
             except Exception as e:
                 self.logger.error(f"Error in emotion recognition loop: {e}")
 
+    
     async def update_emotions(self, prev_emotions: set[str], emotions: set[str]):
         """
         Compare emotions from previous window and current emotions,
@@ -173,7 +149,9 @@ class VisualEmotionRecognitionAgent(BaseAgent):
         for emotion in emotions_to_remove:
             self.logger.info(f"Emotion '{emotion}' has disappeared.")
             try:
-                emotion_beliefs_remove.append(Belief(name="emotion_detected", arguments=[emotion]))
+                emotion_beliefs_remove.append(
+                    Belief(name="emotion_detected", arguments=[emotion], remove=True)
+                )
             except ValidationError:
                 self.logger.warning("Invalid belief for emotion removal: %s", emotion)
 
@@ -197,25 +175,11 @@ class VisualEmotionRecognitionAgent(BaseAgent):
         )
         await self.send(message)
 
-    async def _inform_face_detected(self):
-        if self._face_detected:
-            belief_message = BeliefMessage(create=[Belief(name="face_present")])
-        else:
-            belief_message = BeliefMessage(delete=[Belief(name="face_present")])
-
-        msg = InternalMessage(
-            to=settings.agent_settings.bdi_core_name,
-            thread="beliefs",
-            body=belief_message.model_dump_json(),
-        )
-
-        await self.send(msg)
-
     async def handle_message(self, msg: InternalMessage):
         """
         Handle incoming messages.
 
-        Expects messages to pause or resume the Visual Emotion Recognition
+        Expects messages to pause or resume the Visual Emotion Recognition 
         processing from User Interrupt Agent.
 
         :param msg: The received internal message.
@@ -240,3 +204,4 @@ class VisualEmotionRecognitionAgent(BaseAgent):
         """
         self.video_in_socket.close()
         await super().stop()
+        

src/control_backend/agents/user_interrupt/user_interrupt_agent.py

@@ -164,6 +164,12 @@ class UserInterruptAgent(BaseAgent):
                     else:
                         self.logger.info("Sent resume command.")
 
+                case "stop":
+                    self.logger.debug(
+                        "Received stop command."
+                    )
+                    await self._send_stop_command()
+
                 case "next_phase" | "reset_phase":
                     await self._send_experiment_control_to_bdi_core(event_type)
                 case _:
@@ -422,4 +428,16 @@ class UserInterruptAgent(BaseAgent):
             )
             await self.send(vad_message)
             # Voice Activity Detection and Visual Emotion Recognition agents
-            self.logger.info("Sent resume command to VAD and VED agents.")
+            self.logger.info("Sent resume command to VAD and VED agents.")
+
+    async def _send_stop_command(self):
+        """
+        Send a command to the BDI to stop the program (i.e., skip to end phase).
+        """
+        msg = InternalMessage(
+            to=settings.agent_settings.bdi_core_name,
+            body="",
+            thread="stop"
+        )
+
+        await self.send(msg)

src/control_backend/core/config.py

@@ -82,7 +82,7 @@ class BehaviourSettings(BaseModel):
     :ivar transcription_words_per_token: Estimated words per token for transcription timing.
     :ivar transcription_token_buffer: Buffer for transcription tokens.
     :ivar conversation_history_length_limit: The maximum amount of messages to extract beliefs from.
-    :ivar visual_emotion_recognition_window_duration_s: Duration in seconds over which to aggregate
+    :ivar visual_emotion_recognition_window_duration_s: Duration in seconds over which to aggregate 
     emotions and update emotion beliefs.
     :ivar visual_emotion_recognition_min_frames_per_face: Minimum number of frames per face required
     to consider a face valid.
@@ -112,7 +112,7 @@ class BehaviourSettings(BaseModel):
     conversation_history_length_limit: int = 10
 
     # Visual Emotion Recognition settings
-    visual_emotion_recognition_window_duration_s: int = 3
+    visual_emotion_recognition_window_duration_s: int = 5
     visual_emotion_recognition_min_frames_per_face: int = 3
     # AgentSpeak related settings
     trigger_time_to_wait: int = 2000

src/control_backend/schemas/program.py

@@ -7,7 +7,7 @@ University within the Software Project course.
 from enum import Enum
 from typing import Literal
 
-from pydantic import UUID4, BaseModel
+from pydantic import UUID4, BaseModel, field_validator
 
 
 class ProgramElement(BaseModel):
@@ -24,6 +24,13 @@ class ProgramElement(BaseModel):
     # To make program elements hashable
     model_config = {"frozen": True}
 
+    @field_validator("name")
+    @classmethod
+    def name_must_not_start_with_number(cls, v: str) -> str:
+        if v and v[0].isdigit():
+            raise ValueError('Field "name" must not start with a number.')
+        return v
+
 
 class LogicalOperator(Enum):
     """
@@ -41,8 +48,8 @@ class LogicalOperator(Enum):
     OR = "OR"
 
 
-type Belief = KeywordBelief | SemanticBelief | InferredBelief | EmotionBelief | FaceBelief
-type BasicBelief = KeywordBelief | SemanticBelief | EmotionBelief | FaceBelief
+type Belief = KeywordBelief | SemanticBelief | InferredBelief | EmotionBelief
+type BasicBelief = KeywordBelief | SemanticBelief | EmotionBelief
 
 
 class KeywordBelief(ProgramElement):
@@ -117,15 +124,6 @@ class EmotionBelief(ProgramElement):
     emotion: str
 
 
-class FaceBelief(ProgramElement):
-    """
-    Represents the belief that at least one face is currently in view.
-    """
-
-    name: str = ""
-    face_present: bool
-
-
 class Norm(ProgramElement):
     """
     Base class for behavioral norms that guide the robot's interactions.

uv.lock

@@ -1,5 +1,5 @@
 version = 1
-revision = 3
+revision = 2
 requires-python = ">=3.13"
 resolution-markers = [
     "python_full_version >= '3.14' and sys_platform == 'darwin'",
@@ -1524,6 +1524,7 @@ dependencies = [
     { name = "sphinx-rtd-theme" },
     { name = "tf-keras" },
     { name = "torch" },
+    { name = "tornado", marker = "sys_platform == 'win32'" },
     { name = "uvicorn" },
 ]
 
@@ -1579,6 +1580,7 @@ requires-dist = [
     { name = "sphinx-rtd-theme", specifier = ">=3.0.2" },
     { name = "tf-keras", specifier = ">=2.20.1" },
     { name = "torch", specifier = ">=2.8.0" },
+    { name = "tornado", marker = "sys_platform == 'win32'" },
     { name = "uvicorn", specifier = ">=0.37.0" },
 ]
 
@@ -2724,6 +2726,17 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/52/27/7fc2d7435af044ffbe0b9b8e98d99eac096d43f128a5cde23c04825d5dcf/torchaudio-2.8.0-cp313-cp313t-win_amd64.whl", hash = "sha256:d4a715d09ac28c920d031ee1e60ecbc91e8a5079ad8c61c0277e658436c821a6", size = 2549553, upload-time = "2025-08-06T14:59:00.019Z" },
 ]
 
+[[package]]
+name = "tornado"
+version = "6.5.4"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/37/1d/0a336abf618272d53f62ebe274f712e213f5a03c0b2339575430b8362ef2/tornado-6.5.4.tar.gz", hash = "sha256:a22fa9047405d03260b483980635f0b041989d8bcc9a313f8fe18b411d84b1d7", size = 513632, upload-time = "2025-12-15T19:21:03.836Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/0c/1a/d7592328d037d36f2d2462f4bc1fbb383eec9278bc786c1b111cbbd44cfa/tornado-6.5.4-cp39-abi3-win32.whl", hash = "sha256:1768110f2411d5cd281bac0a090f707223ce77fd110424361092859e089b38d1", size = 446481, upload-time = "2025-12-15T19:21:00.008Z" },
+    { url = "https://files.pythonhosted.org/packages/d6/6d/c69be695a0a64fd37a97db12355a035a6d90f79067a3cf936ec2b1dc38cd/tornado-6.5.4-cp39-abi3-win_amd64.whl", hash = "sha256:fa07d31e0cd85c60713f2b995da613588aa03e1303d75705dca6af8babc18ddc", size = 446886, upload-time = "2025-12-15T19:21:01.287Z" },
+    { url = "https://files.pythonhosted.org/packages/50/49/8dc3fd90902f70084bd2cd059d576ddb4f8bb44c2c7c0e33a11422acb17e/tornado-6.5.4-cp39-abi3-win_arm64.whl", hash = "sha256:053e6e16701eb6cbe641f308f4c1a9541f91b6261991160391bfc342e8a551a1", size = 445910, upload-time = "2025-12-15T19:21:02.571Z" },
+]
+
 [[package]]
 name = "tqdm"
 version = "4.67.1"