New package for decoding neural activity

[ncguilbeault]

Summary

This PR adds 2 new packages to the Bonsai.ML namespace. The first package, called Bonsai.ML.NeuralDecoder is a package designed to interface with the bayesian-neural-decoder python package for performing online decoding of neural activity. The second package, Bonsai.ML.NeuralDecoder.Design builds on this package by adding custom visualizers to observe the decoded posterior distribution over time with respect to the prediction, and adds a mashup visualizer to allow the user to overlay the true position of the animal.

Changes

• Updated package version to preview
• Added visualizer for unidimensional array heatmap time series data
• Added new neural decoding package
• Make plot model public
• Make plot public
• Added classes for getting posterior and position range
• Added posterior time series heatmap visualizer
• Remove position range in favor of a single class
• Rename posterior time series heatmap to posterior visualizer
• Added mashup visualizer
• Exposed methods to specifically update axes
• Added get properties to relay current count and length of data array
• Change position range to value range
• Added default value range mapping
• Changed to true position overlay
• Updated visualize method to correctly merge source and mashup streams
• Call base unload method
• Added null checks when showing data
• Fixed missing XML comment
• Added public getting to retrieve current count information
• Initialize heatmap plot with specific properties
• Update heatmap to display data correctly in show method
• Updated visualize function to make the correct call to the observable visualize
• Correctly adding values to the series
• Updated name to true position overlay
• Added method to update y axis title on load
• Added method to update y axis title
• Changed name to decoder
• Added include workflows
• Removed preview from package version
• Updated README
• Added getting started guide for neural decoder package