.. _release0.101.2:

SpikeInterface 0.101.2 release notes
------------------------------------

4th October 2024

Minor release with bug fixes

core:

* Fix `random_spikes_selection()` (#3456)
* Expose `backend_options` at the analyzer level to set `storage_options` and `saving_options` (#3446)
* Avoid warnings in `SortingAnalyzer` (#3455)
* Fix `reset_global_job_kwargs` (#3452)
* Allow to save recordingless analyzer as (#3443)
* Fix compute analyzer pipeline with tmp recording (#3433)
* Fix bug in saving zarr recordings (#3432)
* Set `run_info` to `None` for `load_waveforms` (#3430)
* Fix integer overflow in parallel computing (#3426)
* Refactor `pandas` save load and `convert_dtypes` (#3412)
* Add spike-train based lazy `SortingGenerator` (#2227)


extractors:

* Improve IBL recording extractors by PID (#3449)

sorters:

* Get default encoding for `Popen` (#3439)

postprocessing:

* Add `max_threads_per_process` and `mp_context` to pca by channel computation and PCA metrics (#3434)

widgets:

* Fix metrics widgets for convert_dtypes (#3417)
* Fix plot motion for multi-segment (#3414)

motion correction:

* Auto-cast recording to float prior to interpolation (#3415)

documentation:

* Add docstring for `generate_unit_locations`  (#3418)
* Add `get_channel_locations` to the base recording API (#3403)

continuous integration:

* Enable testing arm64 Mac architecture in the CI (#3422)
* Add kachery_zone secret (#3416)

testing:

* Relax causal filter tests (#3445)

Contributors:

* @alejoe91
* @h-mayorquin
* @jiumao2
* @samuelgarcia
* @zm711