SortingAnalyzer: add read_only mode (different than lazy)

[alejoe91]

In #4623 (and #4709) we added a lazy mode for loading the SortingAnalyzer. This kept extension data and spike vector unmaterialized to avoid loading large arrays in memory. The lazy mode though did not allow any computation.
However, lazy mode could be used also to compute extensions for large datasets (see #4703) to minimize the overall memory footprint of computations.

To make these 2 use cases separate, this PR adds a read_only flag for loading SortingAnalyzer (e.g., when loading for the GUI), which prevents any saving of extensions). The lazy mode has been extended for creating analyzers, computing, and svaing/merging/splitting.

For very large datasets, one could now use this (especially in conjunction with #4703), to avoid loading large arrays in memory:

analyzer = si.create_sorting_analyzer(sorting, recording, format="zarr"/"binary_folder", folder="my-folder", lazy=True)

The PR also propagates the lazy argument to the parallel computation of PCA metrics.