Too many publications associated with a single author breaks augur export

[huddlej]

Current Behavior

Running augur export v2 with recent ncov data can produce a case when an author has more than associated 52 publications. In this case, the index of the author tuple (see traceback below) exceeds the letters provided in the alphabet string.

Traceback (most recent call last):
  File "/home/jhuddles/miniconda3/envs/nextstrain/bin/augur", line 8, in <module>
    sys.exit(main())
  File "/home/jhuddles/miniconda3/envs/nextstrain/lib/python3.6/site-packages/augur/__main__.py", line 10, in main
    return augur.run( argv[1:] )
  File "/home/jhuddles/miniconda3/envs/nextstrain/lib/python3.6/site-packages/augur/__init__.py", line 74, in run
    return args.__command__.run(args)
  File "/home/jhuddles/miniconda3/envs/nextstrain/lib/python3.6/site-packages/augur/export.py", line 22, in run
    return run_v2(args)
  File "/home/jhuddles/miniconda3/envs/nextstrain/lib/python3.6/site-packages/augur/export_v2.py", line 867, in run_v2
    set_node_attrs_on_tree(data_json, node_attrs)
  File "/home/jhuddles/miniconda3/envs/nextstrain/lib/python3.6/site-packages/augur/export_v2.py", line 509, in set_node_attrs_on_tree
    author_data = create_author_data(node_attrs)
  File "/home/jhuddles/miniconda3/envs/nextstrain/lib/python3.6/site-packages/augur/export_v2.py", line 491, in create_author_data
    node_author_info[node_name]["value"] = author + " {}".format("ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"[index])
IndexError: string index out of range

Expected behavior

The above error should not happen.

How to reproduce

Download the attached data in ncov_issue.zip and run the following command with augur version 8.0.0 (the latest).

augur export v2 \
  --tree tree.nwk \
  --metadata metadata_adjusted.tsv \
  --node-data branch_lengths.json \
  --auspice-config my_auspice_config.json \
  --output ncov_with_accessions.json

Possible solution

• Append a numeric value to the authors list instead of an alphabetical value. The index of the author tuple from the current code could work for this.
• Alternately, use the URL associated with the author in the metadata table or the GISAID accession Either of these two options would also provide more information to the user than an alphabetical character or a number.

The solution should ideally be future-proof enough to not break in a data-dependent manner.

Context

The current author disambiguation approach was introduced in f0e4b1c. The implementation previous to this appears to have used the numeric approach I suggest above, so maybe the GISAID id or URL approach are better?

[dpark01]

@cmloreth and I have seen this recently too

[trvrb]

How about a simple appending of A, B, C, ..., Z, AA, AB, AC, ..., AZ, BA, BB, etc...? I'd prefer Gemma Clark et al B to Gemma Clark et al 2 for aesthetic reasons.

But just having 1, 2, 3, etc... would be okay as well.

[huddlej]

Thanks, @trvrb! Your suggestion sounds good. I'll try this out. For my own context, the reason why the numeric disambiguation is not preferred is partially because the authors are listed in auspice with the number of strains displayed in the tree like so:

Dave J. Baker et al A (2) Dave J. Baker et al B (1)

If we used numeric disambiguation, these authors would be listed like:

Dave J. Baker et al 1 (2) Dave J. Baker et al 2 (1)

which is much more confusing.