Skip to content

Commit

Permalink
Added unittest for lidarseg conversion
Browse files Browse the repository at this point in the history
  • Loading branch information
kokseang committed Jan 27, 2025
1 parent 8ce0543 commit 2b875f2
Show file tree
Hide file tree
Showing 2 changed files with 207 additions and 48 deletions.
97 changes: 49 additions & 48 deletions perception_dataset/t4_dataset/annotation_files_generator.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -220,58 +220,58 @@ def _convert_to_t4_format(
"""Convert the annotations to the NuScenes format.
Args:
scene_anno_dict (Dict[int, List[Dict[str, Any]]]): [description]
frame_index_to_sample_token (Dict[int, str]): [description]
frame_index_to_sample_data_token (Dict[int, str]):
scene_anno_dict (Dict[int, List[Dict[str, Any]]]): [description]
frame_index_to_sample_token (Dict[int, str]): [description]
frame_index_to_sample_data_token (Dict[int, str]):
scene_anno_dict:
{
0: [
{
"category_name" (str): category name of object,
"instance_id" (str): instance id of object,
"attribute_names" (List[str]): list of object attributes,
"three_d_bbox": {
"translation": {
"x" (float): x of object location,
"y" (float): y of object location,
"z" (float): z of object location,
},
"velocity" (Optional[Dict[str, float]]): {
"x" (float): x of object velocity,
"y" (float): y of object velocity,
"z" (float): z of object velocity,
},
"acceleration" (Optional[Dict[str, float]]): {
"x" (float): x of object acceleration,
"y" (float): y of object acceleration,
"z" (float): z of object acceleration,
0: [
{
"category_name" (str): category name of object,
"instance_id" (str): instance id of object,
"attribute_names" (List[str]): list of object attributes,
"three_d_bbox": {
"translation": {
"x" (float): x of object location,
"y" (float): y of object location,
"z" (float): z of object location,
},
"velocity" (Optional[Dict[str, float]]): {
"x" (float): x of object velocity,
"y" (float): y of object velocity,
"z" (float): z of object velocity,
},
"acceleration" (Optional[Dict[str, float]]): {
"x" (float): x of object acceleration,
"y" (float): y of object acceleration,
"z" (float): z of object acceleration,
},
"size": {
"width" (float): width of object size,
"length" (float): length of object size,
"height" (float): height of object size,
},
"rotation": {
"w" (float): w of object quaternion,
"x" (float): x of object quaternion,
"y" (float): y of object quaternion.
"z" (float): z of object quaternion,
},
},
"two_d_box": [
"x" (float): x of left top corner,
"y" (float): y of left top corner,
"w" (float): width of bbox,
"h" (float): height of bbox,
]
"sensor_id": id of the camera
"num_lidar_pts" (int): the number of lidar points in object,
"num_radar_pts" (int): the number of radar points in object,
},
"size": {
"width" (float): width of object size,
"length" (float): length of object size,
"height" (float): height of object size,
},
"rotation": {
"w" (float): w of object quaternion,
"x" (float): x of object quaternion,
"y" (float): y of object quaternion.
"z" (float): z of object quaternion,
},
},
"two_d_box": [
"x" (float): x of left top corner,
"y" (float): y of left top corner,
"w" (float): width of bbox,
"h" (float): height of bbox,
]
"sensor_id": id of the camera
"num_lidar_pts" (int): the number of lidar points in object,
"num_radar_pts" (int): the number of radar points in object,
},
...
],
1: []. ...
...
],
1: []. ...
}
"""
Expand Down Expand Up @@ -445,6 +445,7 @@ def _convert_lidarseg_scene_annotations(

# Create lidarseg folder in
anno_path = Path(anno_dir)
anno_path.mkdir(parents=True, exist_ok=True)
version_name = anno_path.name
lidarseg_relative_path = osp.join(
T4_FORMAT_DIRECTORY_NAME.LIDARSEG_ANNO_FOLDER.value, version_name
Expand Down
158 changes: 158 additions & 0 deletions tests/t4_dataset/test_lidarseg_annotation_files_generator.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,158 @@
import os
from pathlib import Path
import tempfile
from typing import Any, Dict, List
from unittest.mock import patch

import pytest

from perception_dataset.constants import T4_FORMAT_DIRECTORY_NAME
from perception_dataset.t4_dataset.annotation_files_generator import AnnotationFilesGenerator


def _default_lidarseg(default_tmp_dir: str) -> Dict[int, List[Dict[str, Any]]]:
"""Create a default lidarseg annotation dictionary for testing."""
default_lidarseg = {
0: [
{
"paint_categories": ["category1", "category2"],
"lidarseg_anno_file": f"{default_tmp_dir}/lidarseg_anno_file_0.pcd.bin",
"total_lidar_points": 100,
},
],
1: [
{
"paint_categories": ["category1", "category2"],
"lidarseg_anno_file": f"{default_tmp_dir}/lidarseg_anno_file_1.pcd.bin",
"total_lidar_points": 200,
}
],
2: [
{
"paint_categories": ["category1", "category2"],
"lidarseg_anno_file": f"{default_tmp_dir}/lidarseg_anno_file_2.pcd.bin",
"total_lidar_points": 300,
}
],
}
# Create duymmy lidarseg annotation files
for scene_anno in default_lidarseg.values():
for anno in scene_anno:
with open(anno["lidarseg_anno_file"], "wb") as f:
f.write(b"0")

return default_lidarseg


class DummyNuImages:
"""Dummy clas for nuimages."""

def __init__(
self,
version: str = "dummy_version",
dataroot: str = "dummy_data_root",
verbose: bool = False,
):
self.version = version
self.dummy_data_root = dataroot
self.verbose = verbose


class DummyNuscenes:
"""Dummy nuscenes class for unit test."""

def __init__(
self,
version: str = "dummy_version",
dataroot: str = "dummy_data_root",
verbose: bool = False,
):
self.version = version
self.dummy_data_root = dataroot
self.verbose = verbose
self.sample_data = [
{"token": "0", "sample_token": "0", "filename": "data/LIDAR_CONCAT/00000.pcd.bin"},
{"token": "1", "sample_token": "1", "filename": "data/LIDAR_CONCAT/00001.pcd.bin"},
{"token": "2", "sample_token": "2", "filename": "data/LIDAR_CONCAT/00002.pcd.bin"},
{"token": "3", "sample_token": "3", "filename": "data/LIDAR_CONCAT/00003.pcd.bin"},
{"token": "4", "sample_token": "4", "filename": "data/LIDAR_CONCAT/00004.pcd.bin"},
{"token": "5", "sample_token": "5", "filename": "data/LIDAR_CONCAT/00005.pcd.bin"},
]
self.sample = [{"token": "0", "data": "data/LIDAR_CONCAT/00003.pcd.bin"}]


# Note: test case1, case2 use the same scene_anno_dict
class TestAnnotationFilesGenerator:
@pytest.fixture(scope="function")
def instance_for_test(self):
description = {
"visibility": {
"full": "No occlusion of the object.",
"most": "Object is occluded, but by less than 50%.",
"partial": "The object is occluded by more than 50% (but not completely).",
"none": "The object is 90-100% occluded and no points/pixels are visible in the label.",
},
"camera_index": {
"CAM_FRONT": 0,
"CAM_FRONT_RIGHT": 1,
"CAM_BACK_RIGHT": 2,
"CAM_BACK": 3,
"CAM_BACK_LEFT": 4,
"CAM_FRONT_LEFT": 5,
},
"with_lidar": True,
"with_lidarseg": True,
}
return AnnotationFilesGenerator(description=description)

def test_convert_one_scene(
self,
instance_for_test: AnnotationFilesGenerator,
) -> None:
"""Test running convert_one_scene function."""
with patch(
"perception_dataset.t4_dataset.annotation_files_generator.NuScenes", DummyNuscenes
), patch(
"perception_dataset.t4_dataset.annotation_files_generator.NuImages", DummyNuImages
):
with tempfile.TemporaryDirectory() as dir_name:
anno_dir = os.path.join(dir_name, "t4_format")
anno_path = Path(anno_dir)
anno_path_dir = anno_path / "annotation"
anno_path_dir.mkdir(parents=True, exist_ok=True)

instance_for_test.convert_one_scene(
input_dir=dir_name,
output_dir=anno_dir,
scene_anno_dict=_default_lidarseg(default_tmp_dir=dir_name),
dataset_name="dummy_dataset_name",
)
assert os.path.exists(os.path.join(anno_path_dir, "lidarseg.json"))
lidarseg_path = (
Path(anno_dir)
/ T4_FORMAT_DIRECTORY_NAME.LIDARSEG_ANNO_FOLDER.value
/ "annotation"
)
lidarseg_files = [entry for entry in lidarseg_path.iterdir() if entry.is_file()]
assert len(lidarseg_files) == 3

def test__convert_lidarseg_scene_annotations(
self,
instance_for_test: AnnotationFilesGenerator,
) -> None:
"""Test running _convert_lidarseg_scene_annotations function."""
with tempfile.TemporaryDirectory() as dir_name:
dummy_nuscenes = DummyNuscenes()
anno_dir = os.path.join(dir_name, "annotation")
instance_for_test._convert_lidarseg_scene_annotations(
nusc=dummy_nuscenes,
scene_anno_dict=_default_lidarseg(default_tmp_dir=dir_name),
lidar_sensor_channel="LIDAR_CONCAT",
anno_dir=anno_dir,
)
assert os.path.exists(os.path.join(anno_dir, "lidarseg.json"))
lidarseg_path = (
Path(dir_name) / T4_FORMAT_DIRECTORY_NAME.LIDARSEG_ANNO_FOLDER.value / "annotation"
)
lidarseg_files = [entry for entry in lidarseg_path.iterdir() if entry.is_file()]
assert len(lidarseg_files) == 3

0 comments on commit 2b875f2

Please sign in to comment.