usher_to_taxonium: ValueError: DataFrame index must be unique for orient='index'.

[AngieHinrichs]

Hi Theo! For some reason usher_to_taxonium doesn't like my metadata file today. It fails pretty quickly, with this output:

Loading metadata file..
Traceback (most recent call last):
  File "/cluster/home/angie/miniconda3/bin/usher_to_taxonium", line 8, in <module>
    sys.exit(usher_to_taxonium.main())
  File "/cluster/home/angie/miniconda3/lib/python3.9/site-packages/taxoniumtools/usher_to_taxonium.py", line 307, in main
    do_processing(args.input,
  File "/cluster/home/angie/miniconda3/lib/python3.9/site-packages/taxoniumtools/usher_to_taxonium.py", line 42, in do_processing
    metadata_dict, metadata_cols = utils.read_metadata(metadata_file, columns,
  File "/cluster/home/angie/miniconda3/lib/python3.9/site-packages/taxoniumtools/utils.py", line 31, in read_metadata
    metadata_dict = metadata.to_dict("index")
  File "/cluster/home/angie/miniconda3/lib/python3.9/site-packages/pandas/core/frame.py", line 2061, in to_dict
    raise ValueError("DataFrame index must be unique for orient='index'.")
ValueError: DataFrame index must be unique for orient='index'.

The metadata file includes GISAID data so I can't share it here. Any ideas? I did a quick check that all lines of the file have the same number of tab-separated columns.

[theosanderson]

Hi Angie, best guess for a starting point would be a duplicated node_id in the metadata - definitely not that? (I haven't yet dived into the line number yet, doing that now)

[theosanderson]

Looking in more detail, it does look like that - that the "key column" has a duplicate value. Obviously it would be great to give a more useful error message here.

[AngieHinrichs]

Ah, will check rn.

[AngieHinrichs]

That was it! My bad, thanks Theo!

[theosanderson]

Np, thanks for raising