[AVR-761] Lberg/refactor map api #196
Conversation
There was a problem hiding this comment.
I am guessing that the goal of exposing this is that we can select the elements of interest by their bounds, right?
Though for the geometry itself we will want to use get_lane_coords and get_crosswalk_coords
Is this right @lucabergamini
Exactly :) |
|
PTAL @anasrferreira , I've added a function to sample from the lane coords |
| @@ -15,7 +15,7 @@ | |||
| CV2_SHIFT_VALUE = 2 ** CV2_SHIFT | |||
|
|
|||
|
|
|||
| def elements_within_bounds(center: np.ndarray, bounds: np.ndarray, half_extent: float) -> np.ndarray: | |||
| def indices_in_bounds(center: np.ndarray, bounds: np.ndarray, half_extent: float) -> np.ndarray: | |||
There was a problem hiding this comment.
Should this one also get moved to map_api?
lucabergamini
force-pushed
the
lberg/refactor-map-api
branch
from
7acdd5e
to
3b7aad5
Compare
* improve transform through interpolation * add comment * rename variable
| INTER_METER = 0 # fixed interpolation at a given step in meters | ||
| INTER_ENSURE_LEN = 1 # ensure we always get the same number of elements |
There was a problem hiding this comment.
I think you'll need to add #: to autodoc be able to render this doc.
| xyz_inter[:, 0] = np.interp(steps, xp=cum_dist, fp=xyz[:, 0]) | ||
| xyz_inter[:, 1] = np.interp(steps, xp=cum_dist, fp=xyz[:, 1]) | ||
| xyz_inter[:, 2] = np.interp(steps, xp=cum_dist, fp=xyz[:, 2]) |
There was a problem hiding this comment.
I'm not sure about it but I think that the np.interp would allow you to do this on a single call, but have to double check.
There was a problem hiding this comment.
according to the docstring it's only 1D and it doesn't take axis as an argument :(
l5kit/l5kit/data/map_api.py
Outdated
| return False | ||
| traffic_el = element.element.traffic_control_element | ||
| return traffic_el.HasField("traffic_light") is True | ||
|
|
||
| def is_traffic_face_colour(self, element_id: str, colour: str) -> bool: |
There was a problem hiding this comment.
| def is_traffic_face_colour(self, element_id: str, colour: str) -> bool: | |
| def check_traffic_light_state(self, element_id: str, tl_state: str) -> bool: |
There was a problem hiding this comment.
state is something like `ACTIVE, INACTIVE, here we're just checking if the element is a traffic light face and which colour it has. It's a static info, not a dynamic one
| lanes_lines[lane_type].extend([xy_left, xy_right]) | ||
| for lane_area in lanes_area.reshape((-1, INTERPOLATION_POINTS * 2, 2)): | ||
| # need to for-loop otherwise some of them are empty | ||
| cv2.fillPoly(img, [lane_area], (17, 17, 31), **CV2_SUB_VALUES) |
There was a problem hiding this comment.
where does (17, 17, 31) come from?
There was a problem hiding this comment.
no idea, it was there before I started working on this
Refactor of MapAPI: