labware

Microplate and dish definitions for software: outer footprint, well geometry, skirt, material, and the vendor drawing each dimension came from.

Browse them in 3D: https://docs.ramonaoptics.com/tools/labware_viewer.html

Install

pip install labware

or

conda install -c conda-forge labware

Example

from labware import load_labware

plates = load_labware()
plate = plates["SBS_96_well_plate"]

print(plate.name)  # SBS 96 well
print(plate.well_dimensions.rows, plate.well_dimensions.columns)  # 8 12
print(plate.well_dimensions.pitch)  # 0.009, in meters

Layout

• labware/definitions/ holds one JSON file per part. The file name, without .json, is the key load_labware() returns it under.
• labware/schema.py is the pydantic model every definition is validated against.
• labware/labware.v1.schema.json is that model published as a JSON schema. Each definition names it in its $schema key.

Releasing

Uploads to PyPI are automatic, and the version comes from the git tag, so there is no version number to edit in the source.

A merge to main that changes the package is released on its own: the pypi workflow runs the tests, creates the tag and a GitHub release, and publishes.

• If anything in labware/ other than a JSON file changed, such as the schema model or the loader, the minor version goes up.
• Otherwise, if a JSON file in labware/ was added, changed or deleted, which is what a new or corrected plate definition is, the patch version goes up.
• A merge that only touches files outside labware/, such as this README, the tests or the CI, is not released.

A major version is always made by hand.

To release by hand, create the tag from the GitHub web interface:

1. Open https://github.com/ramonaoptics/labware/releases/new
2. Under "Choose a tag", type the new version with a leading v, for example v0.1.1, and pick "Create new tag on publish".
3. Leave the target as main, click "Generate release notes", then "Publish release".

The pypi workflow then builds the sdist and wheel and publishes them, and labware.__version__ in the installed package reports that same version.