Automate data labelling

[tomjholland]

This issue aims to reduce the dependence on Step for splitting up cycler data. Potential solutions to this include, but are not limited to:

• Simple segmentation with tools like aeon or ruptures
• README informed segmentation, using the pybamm experiment descriptions for validation of the identified segments