Checking if all observation names are of equal length can entirely be
done using Pandas functionality. This looks slightly cleaner.

Rename
* `ID_length`  to `id_length` to consistently use lower case
letters for variables.
* `newIDs` and `newID` to `new_obs_names` and `new_obs_name` to have
clearer names and consistently use lower case letters and snake case for
variable names.
* `id` to `new_obs_name` to use more informative name.

Refactor code to not rely on variable `names` but use adata.obs_names
directly.

The argument `data` needs to be an `AnnData` object. The argument name
should convey this.

Makes use of regex expression instead of relying on custom function.

* Use regex expression to find start and end positions.
* Apply to each observation using map on Pandas index.

Update docs to match new argument names.

Move `adata.obs_names_make_unique()` up to group similar parts of the
code together.

Refactor definition of prefixes when observation names have same length.