Closed WAEliasmith closed 3 years ago
We have to change how the btree_root is used and passed down. For example, in ScenarioSetup.py:132
, it only contains the default btree name. I believe it should already contain an absolute path to the resolved btree. That is,
ROOT_DIR/btrees/sdv/<btree name>
But then the problem remains when we're looking for subtrees. Somehow, we have to pass down a new variable btrees_path
consisting of ROOT_DIR/btrees/sdv/
. Or, we can actually extract the path for searching for subtrees from btree_root
using os.path.split
. That way we don't have to introduce a new variable. But that won't work in the future when we have multiple search paths.
Alex and Michal decided on the following design:
btree_root
to root_btree_name
to better indicate what the variable is used for,btrees_path
, which will be passed down from ScenarioSetup
and it'll be used for searching for subtrees.
This PR copies the btree files and organises them by pedestrian or vehicle btrees. This will help for -b searching in the future.
Should the pedestrian drive be renamed to walk?
Main Testing: Test gs_ringoad_modified_with_ego:
(similar error for paperversion) Test base_eval_scenario769:
Experimental_maps testing: Test gn_gs.osm: have to remove maps/ before the lanelet in the osm file
Test gs_4_crossing_paths_running_red_light: have to remove maps/ before the lanelet in the osm file
Testing gs_18_lane_change_same_direction: have to remove maps/ before the lanelet in the osm file
Early Testing: Pulled wavelab/anm_unreal_test_suite#90 onto anm_unreal_test_suite
Further testing of this code is done on https://github.com/rodrigoqueiroz/geoscenarioserver/pull/51 Better handling of not finding btrees is also done on 51