Potential patch needed for updated assembly_summary.txt format

[Weifanwu66]

Describe the bug
ncbi-genome-download currently fails with:
ERROR: Invalid line length in summary file line 2277358. Expected 38, got 33. Attempting to fix.
'submitter' is not in list
This suggests that the current assembly_summary_bacteria.txt from NCBI no longer matches the fixed 38-column schema assumed by the tool.
Affected version
0.3.3
To Reproduce
Steps to reproduce the behavior:
Run:
ncbi-genome-download bacteria --genera "Escherichia" --assembly-level complete --formats fasta --section genbank

The error occurs during metadata parsing.

Expected behavior
The tool should download the genomes of Salmonella from genbank section

Screenshots

Additional context
Add any other context about the problem here.

[kblin]

That same command ran fine on my machine just now. I seem to recall that the NCBI had some sort of maintenance scheduled starting yesterday, maybe something broke with the assembly summary file?
Alternatively, can you get me the first 5 lines and 10 lines around the broken line of your assembly summary file?

You should be able to get these using this code snippet:

(CACHE_FILE=${HOME}/.cache/ncbi-genome-download/genbank_bacteria_assembly_summary.txt; head -n 5 ${CACHE_FILE}; tail -n +2277353 ${CACHE_FILE} | head -n 10) > summary.txt

And then attach the summary.txt file here.

[Weifanwu66]

That same command ran fine on my machine just now. I seem to recall that the NCBI had some sort of maintenance scheduled starting yesterday, maybe something broke with the assembly summary file? Alternatively, can you get me the first 5 lines and 10 lines around the broken line of your assembly summary file?

You should be able to get these using this code snippet:

(CACHE_FILE=${HOME}/.cache/ncbi-genome-download/genbank_bacteria_assembly_summary.txt; head -n 5 ${CACHE_FILE}; tail -n +2277353 ${CACHE_FILE} | head -n 10) > summary.txt
And then attach the summary.txt file here.
Please see attached summary.txt
summary.txt

[kblin]

The file that you attached has the correct line lengths for all the lines, so would not trigger this issue. That still makes me think there was some error fetching the summary file that originally broke your download. If you re-run your download in the same directory today, does it still fail?

Having said that, I see that the NCBI did change the column names, and that is what caused the fix attempt to break. I'll see if I can fix that.