CellEngine Python API Toolkit

A Python interface for the CellEngine API that enables you to securely, reproducibly and easily interact with your data from a Python or Jupyter environment.

Please refer to Python Toolkit Documentation and CellEngine API Documentation for more information.

Upcoming v1.0.0 release (master branch)

v1.0.0 will be the first stable release of the CellEngine Python Toolkit.

Breaking changes

• path property removed. Previously, all classes had a path property that was the URL of the resource. This conflicted with the path property of Folders and Experiments.

• FcsFile.header is now a dict. Previously this was a JSON string. For convenience, it's now pre-parsed into a dict.

• Events DataFrames now consistently use a multi-level index with the first level being the channel name and the second level being the reagent name. This applies to the return values of FcsFile.get_events(), Compensation.apply() and ScaleSet.apply().

• Experiment.get_statistics() return type defaults to DataFrame. Previously it was a dict.

• Cls.from_dict() and Cls.to_dict() have been removed. These were meant for internal use and are no longer necessary.

• RangeGate, RectangleGate, PolygonGate and EllipseGate now inherit from SimpleGate, and QuadrantGate and SplitGate now inherit from CompoundGate. This means that simple gates no longer have a names property and compound gates no longer have a name property.

• Experiment.scalesets is now Experiment.scaleset and Experiment.get_scaleset() takes no args. There is only one ScaleSet in an experiment, so passing an _id or name is unnecessary.

Additional improvements

• Most dependencies have been removed.
• MyPy types have been significantly improved.
• Support for Experiment.path property was added.
• Support for compensation in the Experiment.plot() method was added.
• Cls.update() now only sends changed values to CellEngine. This reduces the permissions required for some operations.
• All tests now run against a real CellEngine instance instead of using mocks, avoiding bugs due to stale mocks.
• Support for experiment.save_revision().