Skip to content

Latest commit

 

History

History
145 lines (117 loc) · 5.81 KB

TESTING.md

File metadata and controls

145 lines (117 loc) · 5.81 KB

Testing 🧪🧫 in Fiji / ImageJ2

Using pytest 🐍🔬 and Python 3 for plain Python code

Those parts of the package that do not interact / depend on ImageJ objects can be tested via pytest up to a certain level, some (most?) of them should even work in a Python 3 environment.

To perform those tests, the packges otherwise provided by ImageJ need to be mocked using the imcf-fiji-mocks package. For seting up a venv use the steps described here:

# check if we're "inside" the repo already, otherwise clone it here:
git remote -v 2>/dev/null | grep -q imcf/python-imcflibs || {
  git clone https://github.com/imcf/python-imcflibs/
  cd python-imcflibs
  git checkout -b processing-options-class origin/processing-options-class
}
# create and activate a new venv:
test -d "venv" || python3 -m venv venv
source venv/bin/activate

# install dependencies / requirements:
MOCKS_REL="0.2.0"
URL_PFX="https://github.com/imcf/imcf-fiji-mocks/releases/download/v$MOCKS_REL"
pip install --upgrade \
    $URL_PFX/imcf_fiji_mocks-${MOCKS_REL}-py2.py3-none-any.whl \
    $URL_PFX/micrometa-15.2.2-py2.py3-none-any.whl \
    $URL_PFX/sjlogging-0.5.2-py2.py3-none-any.whl \
    olefile \
    pytest \
    pytest-cov \
    pip

# now install the 'imcflibs' package in editable mode:
pip install -e .

Using this venv, tests can be triggered just the usual way. To run only specific tests, use e.g.

pytest tests/bdv/test_processingoptions.py

Using pytest 🐍🔬 and Python 2 for plain Python code

For running pytest in a C-Python 2 environment, things are slightly more complicated than the approach described for Python 3 above as pip for Python 2 cannot install a project in editable mode unless it has a setup.py file (which we don't have and don't want).

Therefore, a wheel needs to be built (e.g. using poetry) and installed (every time) into the corresponding virtualenv when performing the tests. Assuming you're having a working poetry setup on your machine, you can simply use the provided scripts/py2-pytest.sh wrapper that will create the virtualenv, build and install the imcflibs wheel and launch pytest with the parameters specified, e.g.

bash scripts/py2-pytest.sh -rv --cov --cov-report html

Common (interactive) testing with ImageJ2 / Fiji

Unfortunately there is nothing like pytest available for the parts that are running exclusively in a ImageJ2 / Fiji context. So in order to provide at least some basic, semi-interactive tests the following conventions are being used:

  • Each function in any of the imcflibs.imagej submodules should have its own directory underneath /tests/interactive-imagej/, using their fully qualified name as the path (only skipping the imcflibs. prefix). For example test scripts for imcflibs.imagej.bioformats.import_image() will be placed in the directory /tests/interactive-imagej/bioformats/import_image/.
  • The scripts inside those directories are intended to be run interactively / manually in a (freshly started) Fiji instance. Yes, really. Any other suggestions are highly welcome!
  • To facilitate this, a collection of test images (and possibly other input data) should be cloned to the local file system. Currently this sample-data repository is NOT publicly available due to legal ⚖ uncertainties. A repo containing test data 🗞 that can be published should be assembled over time though!
  • Any interactive test script should start with a header similar to the one described below. Paths to input data inside the test scripts has to be relative to the location of the sample-data repository mentioned above. This will allow for a fairly okayish testing workflow like this:
    • Make your changes in VS Code, then trigger a build by pressing Shift + Ctrl + B. If things are configured as described in the DEVELOPMENT document, the resulting .jar file will be automatically placed in Fiji's jars/ folder.
    • Next, start a fresh instance of the Fiji that received the newly built JAR.
    • After Fiji has started, simply drag and drop the desired test script onto the main window. This will open the Script Editor, then press Ctrl + R to launch the script.
    • Only on the first run on the machine being used you will have to select the base location of the sample-data repository.
    • All subsequent runs of any test script using the defined Script Parameter IMCF_TESTDATA will remember this selection, so it will be sufficient to just confirm the dialog by pressing Enter.

Quick Workflow Summary

First, make sure to have the test data 🔬🔭around (or some mocks 🪨🪵), then:

  1. Code 📝
  2. Build and deploy locally (Shift+Ctrl+B) 📦
  3. Start Fiji 🇫🇯
  4. Drag-and-Drop the respective test script 🐍🧪
  5. Hit Ctrl+R to run it 🏃‍♀️
  6. Confirm dialog with Enter ✅
  7. Inspect the output 🔎👀
  8. Repeat 🔁

Test Script Template 🏗

As described above, each test script should use the IMCF_TESTDATA parameter to facilitate the manual testing approach. Simply use this template header for creating new scripts (or look into existing ones):

# @ File (label="IMCF testdata location", style="directory") IMCF_TESTDATA

import os
from imcflibs.pathtools import join2

testfile = join2(IMCF_TESTDATA, "systems/lsm700/beads/10x_phmax.czi")
assert os.path.exists(testfile)

In case the test requires the components of the testfile's path to be used, this snippet will do the job:

# @ File (label="IMCF testdata location", style="directory") IMCF_TESTDATA

import os
from imcflibs.pathtools import parse_path

components = parse_path("systems/lsm700/beads/10x_phmax.czi", IMCF_TESTDATA)
assert os.path.exists(components["full"])