Js refactor perception (#874)

* wip refactor

* made a ros split for fieldline detection, objectdetection, camera. first refactor of detector_fieldline, object+detect_node. deleated goal post detection and will use yolov5 as that is much more accurate. split the tests and got most of them working. got ros unit tests to work.

* moved some functions and removed ready from camera. put all testing data into data folder. fixed all the pathways for downloading datasets and models. getting rid of detectir objects since some of it should be in strategy and others in object detection node. Made a utils file for common functions. finished all documentation other then functionaly analysis since these modules are too small. Good enough for now until we rework certain things

* fixed the ros versions so they can be used in sim

.dockerignore

@@ -19,3 +19,4 @@ soccerbot/bags/
 LICENSE
 venv/
 *.webm
+/soccer_perception/data/images/*

.gitignore

@@ -27,8 +27,9 @@ venv/
 /docs/_build/
 /docs/generated/
 *.bag
-/soccer_perception/soccer_object_localization/images/
-/soccer_perception/soccer_object_detection/images/*
+
+/soccer_perception/data/images/*
+/soccer_perception/data/videos/*
 /soccer_strategy/src/soccer_strategy/log.txt
 .mypy_cache/
 cuda-ubuntu2004.pin.1

soccer_common/src/soccer_common/__init__.py

@@ -2,7 +2,7 @@
 from soccer_common.transformation2d import Transformation2D
 
 try:
-    from soccer_common.camera import Camera
+    from soccer_common.perception.camera_calculations import CameraCalculations
     from soccer_common.pid import PID
 except:
     print("not using camera")