rank_genes_groups_dotplot does not work when using reference and using rankby_abs, or setting; values_to_plot='logfoldchanges'

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of scanpy.
• (optional) I have confirmed this bug exists on the master branch of scanpy.

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

As the title says. A specific set of combinations of keywords to rank gene groups and plotting throws an error unexpectedly.

Minimal code sample (that we can copy&paste without having any data)

adata = sc.datasets.paul15()
sc.tl.rank_genes_groups(adata, groupby='paul15_clusters', key_added='GG', use_raw=False, reference='1Ery')
rax = sc.pl.rank_genes_groups_dotplot(adata, key='GG', # , rankby_abs= None,
                                      n_genes=3, cmap='PiYG_r', swap_axes=True,
                                      show=False, values_to_plot='logfoldchanges',
                                      vmin=None, vmax=None)

WARNING: In Scanpy 0.*, this returned logarithmized data. Now it returns non-logarithmized data.
... storing 'paul15_clusters' as categorical
Trying to set attribute `.uns` of view, copying.
WARNING: Default of the method has been changed to 't-test' from 't-test_overestim_var'
WARNING: It seems you use rank_genes_groups on the raw count data. Please logarithmize your data before calling rank_genes_groups.
ERROR: the given dot_color_df data frame has a different shape thanthe data frame used for the dot size. Both data frames needto have the same index and columns
---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-110-708ec3ea001f> in <module>
      1 adata = sc.datasets.paul15()
      2 sc.tl.rank_genes_groups(adata, groupby='paul15_clusters', key_added='GG', use_raw=False, reference='1Ery')
----> 3 rax = sc.pl.rank_genes_groups_dotplot(adata, key='GG', # , rankby_abs= None,
      4                                       n_genes=3, cmap='PiYG_r', swap_axes=True,
      5                                       show=False, values_to_plot='logfoldchanges',

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/scanpy/plotting/_tools/__init__.py in rank_genes_groups_dotplot(adata, groups, n_genes, groupby, values_to_plot, var_names, gene_symbols, min_logfoldchange, key, show, save, return_fig, **kwds)
    861     tl.rank_genes_groups
    862     """
--> 863     return _rank_genes_groups_plot(
    864         adata,
    865         plot_type='dotplot',

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/scanpy/plotting/_tools/__init__.py in _rank_genes_groups_plot(adata, plot_type, groups, n_genes, groupby, values_to_plot, var_names, min_logfoldchange, key, show, save, return_fig, gene_symbols, **kwds)
    534             from .._dotplot import dotplot
    535 
--> 536             _pl = dotplot(
    537                 adata,
    538                 var_names,

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/scanpy/plotting/_dotplot.py in dotplot(adata, var_names, groupby, use_raw, log, num_categories, expression_cutoff, mean_only_expressed, cmap, dot_max, dot_min, standard_scale, smallest_dot, title, colorbar_title, size_title, figsize, dendrogram, gene_symbols, var_group_positions, var_group_labels, var_group_rotation, layer, swap_axes, dot_color_df, show, save, ax, return_fig, vmin, vmax, vcenter, norm, **kwds)
    940         del kwds['color_map']
    941 
--> 942     dp = DotPlot(
    943         adata,
    944         var_names,

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/scanpy/plotting/_dotplot.py in __init__(self, adata, var_names, groupby, use_raw, log, num_categories, categories_order, title, figsize, gene_symbols, var_group_positions, var_group_labels, var_group_rotation, layer, expression_cutoff, mean_only_expressed, standard_scale, dot_color_df, dot_size_df, ax, vmin, vmax, vcenter, norm, **kwds)
    215             # get the same order for rows and columns in the dot_color_df
    216             # using the order from the doc_size_df
--> 217             dot_color_df = dot_color_df.loc[dot_size_df.index][dot_size_df.columns]
    218 
    219         self.dot_color_df = dot_color_df

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/pandas/core/indexing.py in __getitem__(self, key)
    893 
    894             maybe_callable = com.apply_if_callable(key, self.obj)
--> 895             return self._getitem_axis(maybe_callable, axis=axis)
    896 
    897     def _is_scalar_access(self, key: Tuple):

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/pandas/core/indexing.py in _getitem_axis(self, key, axis)
   1111                     raise ValueError("Cannot index with multidimensional key")
   1112 
-> 1113                 return self._getitem_iterable(key, axis=axis)
   1114 
   1115             # nested tuple slicing

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/pandas/core/indexing.py in _getitem_iterable(self, key, axis)
   1051 
   1052         # A collection of keys
-> 1053         keyarr, indexer = self._get_listlike_indexer(key, axis, raise_missing=False)
   1054         return self.obj._reindex_with_indexers(
   1055             {axis: [keyarr, indexer]}, copy=True, allow_dups=True

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/pandas/core/indexing.py in _get_listlike_indexer(self, key, axis, raise_missing)
   1264             keyarr, indexer, new_indexer = ax._reindex_non_unique(keyarr)
   1265 
-> 1266         self._validate_read_indexer(keyarr, indexer, axis, raise_missing=raise_missing)
   1267         return keyarr, indexer
   1268 

~/.virtualenvs/pytorch_latest/lib/python3.8/site-packages/pandas/core/indexing.py in _validate_read_indexer(self, key, indexer, axis, raise_missing)
   1319 
   1320             with option_context("display.max_seq_items", 10, "display.width", 80):
-> 1321                 raise KeyError(
   1322                     "Passing list-likes to .loc or [] with any missing labels "
   1323                     "is no longer supported. "

KeyError: "Passing list-likes to .loc or [] with any missing labels is no longer supported. The following labels were missing: CategoricalIndex(['1Ery'], categories=['1Ery', '2Ery', '3Ery', '4Ery', '5Ery', '6Ery', '7MEP', '8Mk', ...], ordered=False, name='paul15_clusters', dtype='category'). See https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#deprecate-loc-reindex-listlike"

Versions

Start with the reproducer for rank_genes_groups_dotplot and trace the calls through _rank_genes_groups_plot and dotplot, focusing on the dot color and size data frames. Check the combinations involving reference, rankby_abs, and values_to_plot='logfoldchanges'. Done means the supplied example and affected combinations no longer raise the reported shape or KeyError failures.