v0.17.0 dataframe_to_tree_by_relation TypeError str and float

[Jon Palmer]

Describe the issue
A clear and concise description of what the bug is.

Environment
Describe your environment.

• Platform: centos7
• Python version: 3.10.14
• bigtree version: 0.17.0

I'm seeing changes in the behavior of dataframe_to_tree_by_relation. This code was working in previous versions (ie v0.16.4 it works without error):

# rd is a pandas dataframe
>>> rd.shape
(412, 7)
>>> rd.columns
Index(['assembly', 'illumina', 'ont', 'parent', 'seqread', 'strain', 'talias'], dtype='object')
>>> root = dataframe_to_tree_by_relation(rd, parent_col="parent", child_col="strain", attribute_cols=["talias", "seqread", "assembly"])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/envs/vtools/lib/python3.10/site-packages/bigtree/tree/construct.py", line 947, in dataframe_to_tree_by_relation
    f"Unable to determine root node\nPossible root nodes: {sorted(root_names)}"
TypeError: '<' not supported between instances of 'str' and 'float'

I'm not sure which column of the DataFrame this is referring to?

[Kay Jan W.]

Ah there was some changes and fixes with dataframe operations in the latest version. Is it possible to share what does rd contain to recreate the error?

Edit: Can you check if your child_col (strain) is of datatype string? It could be that some strain is str and some is float.

[Kay Jan W.]

Hi, thanks for raising this issue! The fix is implemented in v0.17.1, do upgrade bigtree with the command pip install --upgrade bigtree.