Himank Bootcamp PR #11

steadyfall · 2023-09-14T18:25:08Z

I have completed all the 4 given tasks and have created a PR, as per the bootcamp docs

modules/bootcamp/detect_landing_pad.py

Xierumeng · 2023-09-16T02:51:51Z

modules/bootcamp/detect_landing_pad.py

        # Plot the annotated image from the Result object
        # Include the confidence value
-        image_annotated = ...
+        image_annotated = prediction.plot(pil=True, conf=True)


The expected image is a numpy ndarray, not a PIL object.

According to the Ultralytics documentation, setting img = True would return numpy.ndarray type, but that was not the case.

When I tried running pytest with pil=True, conf=True, all 3 tests passed, whereas running pytest with img=True, conf=True failed all 3 tests.

The error message was the following: AttributeError: 'bool' object has no attribute 'data'.

I do not know whether this is a bug or something else.

The img parameter's type itself is expected to be np.ndarray . This is so that you can pass in an existing image that Ultralytics will draw over, but this bootcamp does not use it.

If you leave pil=True , what is image_annotated object? Do the test output images look fine?

Xierumeng · 2023-09-16T02:55:46Z

modules/bootcamp/detect_landing_pad.py

        # move the copy into CPU space,
        # and convert to a numpy array
-        boxes_cpu = ...
+        boxes_cpu = prediction.boxes.cpu().xyxy.numpy()


Use boxes_xyxy . Also detach before moving to the CPU.

What do you mean by "detach"?

Pytorch tensors have a detach() method, use that.

Xierumeng · 2023-09-16T03:04:52Z

modules/bootcamp/detect_landing_pad.py

+            if result:
+                bounding_boxes.append(box)


In statistics, it is preferred to remove the whole data point if there is an issue (here, the data point is an image). (You do not need to change anything here, just something to think about).

Does that mean removing the entire list of bounding boxes if even one bounding box has an issue?

Correct, an empty list would be returned instead (the entire image is skipped). But you don't need to change anything.

modules/bootcamp/decision_simple_waypoint.py

Xierumeng · 2023-09-16T03:16:51Z

modules/bootcamp/decision_waypoint_landing_pads.py


-        # Add your own
+        self.action_dict = dict(zip(("MOVE", "HALT", "LAND"), range(3,6)))
+        self.origin = location.Location(0,0)


What if the drone didn't start at (0,0) , but could start anywhere?

Then, in that case, wouldn't we need to keep track of the initial position of the drone?

Yes. Alternatively, don't use the initial position of the drone.

Xierumeng · 2023-09-16T03:22:17Z

modules/bootcamp/decision_waypoint_landing_pads.py

+        landing_pad_distances = list(map(lambda loc: shortest_distance(loc, given_location), landing_pad_locations))
+        landing_pad_hashmap = dict(zip(landing_pad_locations, landing_pad_distances))
+        landing_pad_hashmap = tuple(sorted(list(landing_pad_hashmap.items()), key=lambda i: i[1]))


This is very convoluted and creates many new objects that will need to be destroyed. Is there a more space efficient way?

Also, sorting is not very efficient. It's possible to get the minimum without sorting the list of landing pad locations.

Fixed this with initializing closest_location variable with first location in the list and doing a for loop for the rest of the list to replace the closer location with closest_location in steadyfall@f42441a.

modules/bootcamp/decision_waypoint_landing_pads.py

Xierumeng · 2023-09-16T03:24:09Z

modules/bootcamp/decision_waypoint_landing_pads.py

+            x_2, y_2 = target_location.location_x, target_location.location_y
+            x_square = (x_2 - x_1) ** 2
+            y_square = (y_2 - y_1) ** 2
+            return (x_square + y_square) ** 0.5


Square roots are computationally expensive. Is there another way? (You do not need to change anything here, just something to think about.

I do not square root the number and only square root it when required?

That just procrastinating the work, it doesn't eliminate it. What would happen if there was no square root?

modules/bootcamp/decision_waypoint_landing_pads.py

Xierumeng · 2023-09-16T03:48:41Z

modules/bootcamp/decision_waypoint_landing_pads.py

+            if (self.check_if_near_target(landing_pad, report_position)
+                and (shortest_distance(self.waypoint, report_position) - shortest_distance(self.waypoint, landing_pad) < 0.1)
+                and (
+                    (self.check_if_near_target(landing_pad, self.origin) and self.check_if_near_target(self.waypoint, self.origin))
+                    or (not self.check_if_near_target(report_position, self.origin))
+                )):


This is also a bit complex. What about checking whether the drone has already arrived at the waypoint (e.g. storing some state)?

This is also a bit complex. What about checking whether the drone has already arrived at the waypoint (e.g. storing some state)?

I had tried doing that, but since this method runs in an infinite loop, any data that I try to store goes away.

Declare variables (e.g. self.variable_name ) in __init__() and use those (e.g. a = self.variable_name , self.variable_name = 5 ). Declare as few or as many as you need.

Those are instance/object variables and will persist with the lifetime of the object.

modules/bootcamp/decision_simple_waypoint.py

modules/bootcamp/detect_landing_pad.py

…, UWARG#11 (comment) , UWARG#11 (comment) , UWARG#11 (comment)

…G#11 (comment)

Xierumeng · 2023-09-20T03:39:21Z

modules/bootcamp/decision_simple_waypoint.py

+        Returns the relative x and y coordinates for drone to be sent to.
+        """
+        relative_x, relative_y = self.distance_between_waypoint(given_location)
+        return (relative_x, relative_y)


Remove unneeded parantheses.

Xierumeng · 2023-09-20T03:57:00Z

modules/bootcamp/decision_simple_waypoint.py

        # ============

-        # Add your own
+        self.action_dict = dict(zip(("MOVE", "HALT", "LAND"), range(3,6)))


Can use enum instead.

Xierumeng · 2023-09-20T03:57:25Z

modules/bootcamp/decision_simple_waypoint.py

+        if action is None:
+            pass


Redundant, remove.

Xierumeng · 2023-09-20T04:04:07Z

modules/bootcamp/decision_waypoint_landing_pads.py

        # ============

+    @staticmethod
+    def shortest_distance(target_location:location.Location, given_location:location.Location) -> float:


Good use of static method!

Xierumeng · 2023-09-20T04:05:02Z

modules/bootcamp/decision_waypoint_landing_pads.py

+            y_square = (y_2 - y_1) ** 2
+            return (x_square + y_square) ** 0.5
+
+    def relative_coordinates_of_target(self, target_location:location.Location, given_location:location.Location) -> "tuple[float, float]":


This can also be static.

Xierumeng · 2023-09-20T04:05:14Z

modules/bootcamp/decision_waypoint_landing_pads.py

+        """
+        relative_location_x = target_location.location_x - given_location.location_x
+        relative_location_y = target_location.location_y - given_location.location_y
+        return (relative_location_x, relative_location_y)


Remove unneeded parantheses.

Xierumeng · 2023-09-20T04:06:29Z

modules/bootcamp/decision_waypoint_landing_pads.py

+        Returns the relative x and y coordinates for drone to be sent to.
+        """
+        relative_x, relative_y = self.relative_coordinates_of_target(target_location, given_location)
+        divider = 1


Remove this since it doesn't do anything.

Xierumeng · 2023-09-20T04:08:55Z

modules/bootcamp/decision_waypoint_landing_pads.py

+        """
+        Finds out the closest landing pad from the given location by checking out their distances.
+        """
+        closest_location = landing_pad_locations[0]


What if there were no landing pads? (You do not need to change anything, just something to think about).

Xierumeng · 2023-09-20T04:09:47Z

modules/bootcamp/decision_waypoint_landing_pads.py

+            distance_from_given_location = DecisionWaypointLandingPads.shortest_distance(landing_pad, given_location)
+            distance_from_closest_location = DecisionWaypointLandingPads.shortest_distance(closest_location, given_location)


Calculating the distance repeatedly for the closest landing pad is wasteful, is there a way to avoid this?

steadyfall added 8 commits September 12, 2023 00:33

added .python-version for pyenv-win

8ec11f3

completed first task;

a7e0a33

changed DISPLAY_SCALE; completed second task

3c67cb2

changed TIME_STEP_SIZE and DISPLAY_SCALE (system-specific)

5e64fc5

completed task 3

318294d

recorded tricky seed number

43f0dc4

completed task 5

52da233

Fixed lint issues with pylint

d8dc641

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/detect_landing_pad.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/decision_simple_waypoint.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/decision_simple_waypoint.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/decision_simple_waypoint.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/decision_simple_waypoint.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/decision_waypoint_landing_pads.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/decision_waypoint_landing_pads.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/decision_simple_waypoint.py Outdated Show resolved Hide resolved

Xierumeng reviewed Sep 16, 2023

View reviewed changes

modules/bootcamp/detect_landing_pad.py Outdated Show resolved Hide resolved

steadyfall added 6 commits September 16, 2023 16:02

Effect the changes mentioned in Bootcamp PR Comment UWARG#11 (comment) …

b521824

…, UWARG#11 (comment) , UWARG#11 (comment) , UWARG#11 (comment)

Effect the changes mentioned in Bootcamp PR UWARG#11 (comment) , UWAR…

6851633

…G#11 (comment)

removed unnecessary comment

a7d9960

Effect the changes mentioned in Bootcamp PR UWARG#11 (comment) , UWAR…

f42441a

…G#11 (comment)

Effect the changes mentioned in Bootcamp PR UWARG#11 (comment)

20f6395

Effect the changes mentioned in Bootcamp PR UWARG#11 (comment)

b6dbf23

fixed return types in methods

adcd012

steadyfall requested a review from Xierumeng September 18, 2023 15:55

Xierumeng reviewed Sep 20, 2023

View reviewed changes

modules/bootcamp/decision_simple_waypoint.py

Comment on lines +108 to +109

if action is None:

pass

Copy link

Contributor

Xierumeng Sep 20, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Redundant, remove.

Xierumeng reviewed Sep 20, 2023

View reviewed changes

		distance_from_given_location = DecisionWaypointLandingPads.shortest_distance(landing_pad, given_location)
		distance_from_closest_location = DecisionWaypointLandingPads.shortest_distance(closest_location, given_location)

Himank Bootcamp PR #11

Are you sure you want to change the base?

Himank Bootcamp PR #11

Uh oh!

Conversation

steadyfall commented Sep 14, 2023

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Xierumeng Sep 20, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Xierumeng Sep 20, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Xierumeng Sep 20, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Xierumeng Sep 20, 2023 •

edited

Loading

Xierumeng Sep 20, 2023 •

edited

Loading

Xierumeng Sep 20, 2023 •

edited

Loading