Writing new scan point generators

Let’s walk through the simplest generator, LineGenerator, and see how it is written.

We import the baseclass Generator and the Point class that we will be generating instances of.

Our new subclass includes a docstring giving a short explanation of what it does

The initialiser stores the arguments given to it, then generates the three properties that are required by the baseclass:

• position_units: Dict of str position_name -> str position_unit
• index_dims: List of int dimension sizes for the dataset
• index_names: List of str dimension names for the dataset

It is important to note that the position_units property will have the same number of elements as index_dims for grid based scans (like LineGenerator). For non grid based scans (like SpiralGenerator), index_dims will typically have less elements, because the last two or more dimensions will be unrolled into one long array. This avoids sparse datasets.

We have a repeated bit of code here, so have pulled it out into a function. It calculates the position of a point given an index

This is the entry point for external code. It is expecting us to produce a number of Point instances, one for each point in the scan. We are required to fill in the following dictionaries of str position_name -> float position:

• positions: The capture position corresponding to the centre of the scan frame
• lower: The lower bound of the scan frame if the scan is to be used for continuous scanning
• upper: The upper bound of the scan frame if the scan is to be used for continuous scanning

We also fill in the list of datapoint indexes:

• indexes: The index into the dataset that the data frame should be stored in

The yield keyword turns the python function into a generator, which can then be used by the external program to iterate through points without evaluating them all at the start.