Refactor object detection

assets/__init__.py

@@ -4,7 +4,7 @@
 
 import darkdetect
 
-from src.utils import get_file_path
+from src.path_utils import get_file_path
 
 ASSET_DIRPATH = path.dirname(path.abspath(__file__))
 
@@ -20,7 +20,7 @@ def get_icon(
     sys_platform = platform.system()
     if sys_platform == "Linux":
         # Do this since linux doesn't support icons like windows and mac
-        return None, None 
+        return None, None
     is_dark_mode = (
         override_is_dark_mode or darkdetect.isDark()
         if darkdetect.isDark() is not None

pyproject.toml

@@ -25,6 +25,7 @@ dependencies = [
     "pyobjc-framework-Cocoa; sys_platform == 'darwin'",
     "pyvirtualcam",
     "watchdog>=6.0.0",
+    "lapx>=0.9.4",
 ]
 name = "commander"
 requires-python = "==3.12.*"
@@ -41,7 +42,13 @@ commander-tk = "talos:tk_interface"
 commander-pyside = "talos:pyside_interface"
 
 [dependency-groups]
-dev = ["pyinstaller>=6.16.0", "textual-dev>=1.8.0", "pytest", "pytest-cov", "pytest-mock"]
+dev = [
+    "pyinstaller>=6.16.0",
+    "textual-dev>=1.8.0",
+    "pytest",
+    "pytest-cov",
+    "pytest-mock",
+]
 
 [project.optional-dependencies]
 mediapipe = [
@@ -74,7 +81,5 @@ include = ["src*", "assets*", "config*"]
 addopts = ["--cov=src", "--cov-report=html"]
 
 [tool.coverage.run]
-omit =[
-    "tests/*"
-]
-relative_files = true
+omit = ["tests/*"]
+relative_files = true

src/arg_parser.py

@@ -1,5 +1,6 @@
 import argparse
-from src.model_options import MODEL_OPTIONS
+
+from src.tracking import MODEL_OPTIONS
 
 
 def _create_arg_parser():

src/config/path.py

@@ -1,6 +1,6 @@
 import os
 
-from src.utils import get_file_path
+from src.path_utils import get_file_path
 
 _CONFIG_DIR = os.path.join(os.path.dirname(__file__), "../../config")
 ROBOT_CONFIG_FILENAME = "robot_configs.local.yaml"

src/config/read.py

@@ -14,8 +14,7 @@
     ROBOT_CONFIGS_PATH,
 )
 from src.config.schema.app import AppSettings
-
-from ..utils import get_file_path
+from src.path_utils import get_file_path
 
 
 def read_default_robot_config() -> dict[str, Any]:

src/config/schema/app.py

@@ -5,15 +5,15 @@
 from pydantic import Field
 from pydantic_settings import (
     BaseSettings,
-    PydanticBaseSettingsSource,
-    YamlConfigSettingsSource,
     CliSettingsSource,
+    PydanticBaseSettingsSource,
     SettingsConfigDict,
+    YamlConfigSettingsSource,
 )
 
-from src.arg_parser import ARG_PARSER
 from src.config.path import APP_SETTINGS_DEFAULT_PATH, APP_SETTINGS_PATH
 
+
 def _parse_args(root_parser: argparse.ArgumentParser, args) -> argparse.Namespace:
     """
     Implement overrides for App Settings that are triggered by providing certain cli args
@@ -27,6 +27,7 @@ def _parse_args(root_parser: argparse.ArgumentParser, args) -> argparse.Namespac
         args.draw_bboxes = True
     return args
 
+
 class AppSettings(BaseSettings):
     """
     Application-wide settings that are not specific to individual connections.
@@ -38,7 +39,9 @@ class AppSettings(BaseSettings):
         env_file_encoding="utf-8",
         case_sensitive=False,
         extra="ignore",
-        yaml_file=APP_SETTINGS_PATH if os.path.exists(APP_SETTINGS_PATH) else APP_SETTINGS_DEFAULT_PATH,
+        yaml_file=APP_SETTINGS_PATH
+        if os.path.exists(APP_SETTINGS_PATH)
+        else APP_SETTINGS_DEFAULT_PATH,
         yaml_file_encoding="utf-8",
         cli_parse_args=True,
         cli_kebab_case=True,
@@ -53,15 +56,19 @@ def settings_customise_sources(
         dotenv_settings: PydanticBaseSettingsSource,
         file_secret_settings: PydanticBaseSettingsSource,
     ) -> tuple[PydanticBaseSettingsSource, ...]:
+        from src.arg_parser import ARG_PARSER
+
         return (
             init_settings,
-            CliSettingsSource(settings_cls, root_parser=ARG_PARSER, parse_args_method=_parse_args),
+            CliSettingsSource(
+                settings_cls, root_parser=ARG_PARSER, parse_args_method=_parse_args
+            ),
             YamlConfigSettingsSource(settings_cls),
             dotenv_settings,
             env_settings,
         )
 
-    log_level: Literal['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL'] = Field(
+    log_level: Literal["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"] = Field(
         default="INFO",
         description="Logging level (e.g., DEBUG, INFO, WARNING, ERROR)",
     )
@@ -73,3 +80,7 @@ def settings_customise_sources(
         default=30,
         description="Frames per second for pulling video streams (can be lower than max_fps to reduce load)",
     )
+    disable_performance_warnings: bool = Field(
+        default=False,
+        description="Whether to disable warnings about performance issues (e.g., if processing is taking too long and frames are being dropped or the opposite)",
+    )

src/config/watchers/app_settings_file_handler.py

@@ -11,7 +11,8 @@
     FileSystemEventHandler,
 )
 
-from ...utils import get_file_path
+from src.path_utils import get_file_path
+
 from ..path import APP_SETTINGS_PATH, BACKUP_DIR
 from ..schema.app import AppSettings
 
@@ -65,7 +66,7 @@ def on_deleted(self, event: DirDeletedEvent | FileDeletedEvent) -> None:
         logger.info(f"Backed up current app settings to {path}")
 
     def on_created(self, event: DirCreatedEvent | FileCreatedEvent):
-        self.on_modified(event)
+        self.on_modified(event)  # pyright: ignore[reportArgumentType]
 
     def on_moved(self, event: DirMovedEvent | FileMovedEvent):
-        self.on_modified(event)
+        self.on_modified(event)  # pyright: ignore[reportArgumentType]

src/config/watchers/robot_config_handler.py

@@ -13,9 +13,9 @@
     FileSystemEventHandler,
 )
 
-from ...utils import get_file_path
-from ..path import BACKUP_DIR, ROBOT_CONFIGS_PATH
-from ..schema.robot import RobotConfigs
+from src.config.path import BACKUP_DIR, ROBOT_CONFIGS_PATH
+from src.config.schema.robot import RobotConfigs
+from src.path_utils import get_file_path
 
 
 def take_backup() -> str:

src/connection/publisher.py

@@ -22,13 +22,13 @@ class Direction(IntEnum):
     """Directional Enum for interface controls
 
     The values can be summed to get combined directions(from -4 to 4)
-    -4: DL, -3: L, -2: UL, -1: D, 0: None, 1: U, 2: DR, 3: R, 4: UR
+    -4: DR, -3: R, -2: UR, -1: D, 0: None, 1: U, 2: DL, 3: L, 4: UL
     """
 
     UP = 1
     DOWN = -1
-    RIGHT = 3
-    LEFT = -3
+    LEFT = 3
+    RIGHT = -3
 
     @staticmethod
     def toDirectionTuple(sum_direction: int) -> tuple[int, int]:

src/directors/continuous_director.py

@@ -3,7 +3,7 @@
 from loguru import logger
 
 import src.config as config
-from src.connection.publisher import Publisher
+from src.connection.publisher import Direction, Publisher
 from src.directors.base_director import BaseDirector
 from src.utils import calculate_acceptable_box, calculate_center_bbox
 
@@ -83,27 +83,27 @@ def process_frame(
             center_bottom = center + frame_buffer
 
             if average < center_top:
-                logger.info(f"Move camera up: Average:{average} Center:{center}")
+                logger.debug(f"Move camera up: Average:{average} Center:{center}")
                 change_in_y = average - center_top
             elif average > center_bottom:
-                logger.info(f"Move camera down: Average:{average} Center:{center}")
+                logger.debug(f"Move camera down: Average:{average} Center:{center}")
                 change_in_y = average - center_bottom
 
             if change_in_x > 0:
-                publisher.polar_pan_continuous_start(-1, 0)
-                # print("start")
+                publisher.polar_pan_continuous_direction_start(Direction.RIGHT)
+                logger.debug("start right")
                 self.last_command_stop = False
             elif change_in_x < 0:
-                publisher.polar_pan_continuous_start(1, 0)
-                # print("start")
+                publisher.polar_pan_continuous_direction_start(Direction.LEFT)
+                logger.debug("start left")
                 self.last_command_stop = False
             elif change_in_y < 0:
-                publisher.polar_pan_continuous_start(0, 1)
-                # print("start")
+                publisher.polar_pan_continuous_direction_start(Direction.UP)
+                logger.debug("start up")
                 self.last_command_stop = False
             elif change_in_y > 0:
-                publisher.polar_pan_continuous_start(0, -1)
-                # print("start")
+                publisher.polar_pan_continuous_direction_start(Direction.DOWN)
+                logger.debug("start down")
                 self.last_command_stop = False
             elif not self.last_command_stop:
                 publisher.polar_pan_continuous_stop()

src/interface/pyside_gui/main_interface.py

@@ -25,7 +25,7 @@
 from src.interface.pyside_gui.qtwidgets import Toggle
 from src.interface.pyside_gui.styles import get_main_stylesheet
 from src.talos_app import App, ControlMode
-from src.tracking import MODEL_OPTIONS
+from src.tracking.options import MODEL_OPTIONS
 from src.utils import (
     add_termination_handler,
     remove_termination_handler,

src/path_utils.py

@@ -0,0 +1,18 @@
+import os.path as path
+import sys
+
+
+def get_file_path(relative_path: str) -> str:
+    """
+    Get the absolute file path for a given relative path. This method is necessary
+    to ensure the program works when compiled with PyInstaller.
+
+    Args:
+        relative_path (str): The relative path to the file.
+
+    Returns:
+        str: The absolute file path if the program is compiled with PyInstaller, otherwise the relative path.
+    """
+    if getattr(sys, "frozen", False) and hasattr(sys, "_MEIPASS"):
+        return path.join(sys._MEIPASS, relative_path)  # pyright: ignore[reportAttributeAccessIssue]
+    return relative_path

src/streaming/streamer.py

@@ -41,10 +41,10 @@ def draw_visuals(bboxes, frame):  # -> Any:
 
     for box in bboxes:
         # Draw each bbox
-        x1, y1, x2, y2 = box
+        x1, y1, x2, y2 = map(int, box)
         cv2.rectangle(frame, (x1, y1), (x2, y2), BBOX_COLOR.GREEN.value, 2)
 
-        bbox_center_x, bbox_center_y = calculate_center_bbox(box)
+        bbox_center_x, bbox_center_y = map(int, calculate_center_bbox(box))
         # Draw center of bounding box dot, this the value the commander is using
         cv2.circle(
             frame,

src/talos_app.py

@@ -233,9 +233,6 @@ def change_model(self, option: str | None = None) -> bool:
         if option is not None and len(self.connections) == 0:
             logger.warning("No connections available, skipping model initialization")
             return False
-        if self.director is not None and self.director.is_active():
-            self.director.stop_auto_control()
-            logger.info("Director stopped")
         if option is None:
             self.tracker.swap_model(None)
             self.model_selection = None
@@ -267,6 +264,7 @@ def set_manual_control(self, manual: bool) -> None:
         """Sets the active connection's manual/automatic control mode"""
         if self.director is None:
             return logger.error("No active director")
+        logger.debug("Setting manual control to {}", manual)
         self.director.set_manual_control(manual=manual)
 
     def set_control_mode(self, ctrl_mode: ControlMode) -> ControlMode:

src/tracking/__init__.py

@@ -1,49 +1,3 @@
-from loguru import logger
-from src.model_options import MODEL_OPTIONS, ModelOption
-
-from .detector import ObjectModel
-from .haar_cascade.basic_model import BasicModel
+from .options import MODEL_OPTIONS, USABLE_MODELS, ModelOption
 
 __all__ = ["ModelOption", "MODEL_OPTIONS", "USABLE_MODELS"]
-
-
-USABLE_MODELS: dict[str, ObjectModel.__class__] = {}
-
-# haar_cascade is imported via opencv-python by default
-USABLE_MODELS["basic"] = BasicModel
-
-try:
-    from .media_pipe import MediaPipeModel, MediaPipePoseModel
-
-    # USABLE_MODELS[ModelOption.KEEPAWAY] = (KeepAwayModel, KeepAwayDirector)
-    USABLE_MODELS[ModelOption.MEDIAPIPEPOSE] = MediaPipeModel
-    USABLE_MODELS[ModelOption.MEDIAPIPE] = MediaPipePoseModel
-except ImportError as e:
-    logger.warning(
-        f"""{e}
-        Failed to import mediapipe. This is an optional import, but may limit the ability to run this model.
-        This can be installed using `uv sync --extra mediapipe` or `uv sync --all-extras`"""
-    )
-
-try:
-    from .yolo.model import (
-        YOLOLargeModel,
-        YOLOMediumModel,
-        YOLONanoModel,
-        YOLOSmallModel,
-        YOLOXLargeModel,
-    )
-
-    USABLE_MODELS[ModelOption.YOLO_NANO] = YOLONanoModel
-    USABLE_MODELS[ModelOption.YOLO_SMALL] = YOLOSmallModel
-    USABLE_MODELS[ModelOption.YOLO_MEDIUM] = YOLOMediumModel
-    USABLE_MODELS[ModelOption.YOLO_LARGE] = YOLOLargeModel
-    USABLE_MODELS[ModelOption.YOLO_XLARGE] = YOLOXLargeModel
-except ImportError as e:
-    logger.warning(
-        f"""{e}
-        Failed to import Yolo model. This is an optional import, but may limit the ability to run this model.
-        This can be installed using `uv sync --extra yolo` or `uv sync --all-extras`"""
-    )
-
-