Openclean-core

* Remove environment variable *OPENCLEAN_WORKERS*.

* Use compact serialization for HISTORE archives.
* Load and sample datasets from a data stream in `openclean.engine.base.OpencleanEngine`.
* Support stream operators on dataset snapshots in `openclean.engine.base.OpencleanEngine`.
* Add summary for data frame conflict groups.

* Make checking out a committed dataset in the `openclean.data.archive.base.ArchiveStore` optional.
* Enable cache refresh for cached datasets in `openclean.data.archive.cache.CachedDatastore`.

* Add optional version parameter when requesting metadata for a dataset version in `openclean.engine.dataset.DatasetHandle`.

* Add `openclean.function.token.base.Token` as separate class.
* Rename `openclean.function.token.base.StringTokenizer` to `Tokenizer`
* Adjust token transformer and tokenizer for new Token class.
* Change structure of datatype count in column profiler.
* Option to get set of conflicting values from `DataFrameGrouping` groups.
* Multi-threading for `ValueFunction.apply()`.
* Separate DBSCAN outlier class.
* Move us-street name functions to `openclean-geo`.

* Standardize parameter names for sample methods (\115)
* Bug fix for `openclean-notebook`