Optional: run tests

If you wish to explicitly test that tomoDRGN has been installed and is functioning correctly on your system, you can run the following tests. Each test uses a small toy dataset included in the tomoDRGN source code.

If these tests fail, please report the issue on tomoDRGN’s Github page. The most likely cause for failure is conda installing a more updated dependency version that is not compatible with tomoDRGN.

Note

Running any of the following tests requires a further installation of tomoDRGN’s testing dependencies.

cd $TOMODRGN_SOURCE_DIR
python -m pip install ".[tests]"

Note

Some useful arguments that can be supplied to the pytest commands below:

• --capture=tee-sys: log pytest and tomodrgn output to STDOUT

• --basetemp=/custom/output/directory: specify a directory to save all outputs from the test (perhaps to see sample tomoDRGN outputs)

Optional test 1

Run a quick end-to-end test of the most essential and frequently used CLI commands, tomodrgn train_vae and tomodrgn analyze. Takes about 1 minute. Produces about 20 MB of outputs.

cd $TOMODRGN_SOURCE_DIR/testing
pytest --script-launch-mode=subprocess quicktest.py

Optional test 2

Run a comprehensive end-to-end test of all CLI commands with multiple combinations of CLI options (except Jupyter notebooks). Takes about about 10 minutes (for each pytest command) on an Ubuntu workstation with a 4060Ti, or about 50 minutes on a MacBook. Produces about 1 GB of outputs.

cd $TOMODRGN_SOURCE_DIR/testing
pytest --script-launch-mode=subprocess ./commandtest.py
pytest --script-launch-mode=subprocess ./commandtest_warptools.py