tacitvenom

### Problem description The documentation of the function `collect_dataset_metadata` mentions the parameter `predicates` twice and the description of both varies: https://kartothek.readthedocs.io/en/stable/_rst/kartothek.io.dask.dataframe.html#kartothek.io.dask.dataframe.collect_dataset_metadata

For a clustering usecase, I tried different parameters and while calculating validity index, I run into the following ValueError: ``` /envs/venv/lib/python3.9/site-packages/hdbscan/validity.py:33: RuntimeWarning: invalid value encountered in divide result /= distance_matrix.shape[0]...