Welcome to the MILUV devkit page. This Python devkit provides useful functions and examples to accompany the MILUV dataset. To begin using this devkit, clone or download and extract the repository.
03-07-2024: MILUV devkit v1.0.0 released.
The devkit requires Python 3.8 or greater. To install the devkit and its dependencies, run
$ pip3 install .
inside the devkit's root directory (~/path/to/project/MILUV).
Alternatively, run
$ pip3 install -e .
inside the devkit's root directory, which installs the package in-place, allowing you make changes to the code without having to reinstall every time.
For a list of all dependencies, refer to requirements.txt in the repository's root directory.
To ensure installation was completed without any errors, test the code by running
$ pytest
in the root directory.
To get started, download the MILUV dataset. By default, the devkit expects the data for each experiment is present in /miluv/data/EXP_NUMBER, where EXP_NUMBER is the number of the experiment.
If you wish to change the default data location, be sure to modify the data directory in the devkit code.
Using the MILUV devkit, retrieving sensor data by timestamp from experiment 1c can be implemented as:
from miluv.data import DataLoader
import numpy as np
mv_1c = DataLoader(
"3a",
barometer=False,
height=False,
)
timestamps = np.arange(0, 10, 1) # Time in s
data_at_timestamps = mv_1c.data_from_timestamps(timestamps)This example can be made elaborate by selecting specific robots and sensors to fetch from at the given timestamps.
from miluv.data import DataLoader
import numpy as np
mv_1c = DataLoader(
"3a",
barometer=False,
height=False,
)
timestamps = np.arange(0, 10, 1) # Time in s
robots = ["ifo001", "ifo002"] # We are leaving out ifo003
sensors = ["imu_px4", "imu_cam"] # Fetching just the imu data
data_at_timestamps = mv_1c.data_from_timestamps(
timestamps,
robots,
sensors,
)For more information regarding the MILUV development kit, please refer to the documentation.
This development kit is distributed under the MIT License.
