Testing¶

The simplest way to test any additions to Jupyter-ROS is to create a fresh environment and install the package in development mode.

Create new test environment with minimal packages

$ conda create -n test_env python=3.9 jupyterlab nodejs=15 jupyter-packaging ros-noetic-desktop -c conda-forge -c robostack

$ conda activate test_env

Navigate to the root directory and install jupyter-ros
```
$ cd jupyter-ros/
$ pip install -e .
```
If there are any errors with this step, this indicates that the new additions are not configured correctly for installation. This will require some additional troubleshooting, but a common issue is to forget to include newly required dependencies in the setup.py file.
```
setup_args = {
  'install_requires': [
      'ipywidgets>=7.0.0',
      'bqplot',
      'numpy',
      'rospkg',
      'ipycanvas'
  ],
}
```
Symlink the JupyterLab extension and verify that jupyros can be imported.
```
$ jupyter labextension develop . --overwrite
$ jupyter lab
```
```
import jupyros
```
This step may also require some troubleshooting depending on the changes made to Jupyter-ROS. If you see interference from other conda environments, e.g. additional lab extensions which should not be enabled in the test environment, it is often helpful to remove the ~/.jupyter directory.
```
# List all the lab extensions
$ jupyter labextension list

# Remove jupyter directory [optional]
$ rm -r ~/.jupyter
```
Once the setup is complete, it is now time to test the new additions. This type of testing will vary greatly depending on the changes, we suggest to use your best judgement.