Smudge report crashing. Inferred coverage 0?

[jychoilab]

Hello

Thanks for making the program. I was able to run all the commands and at the smudgeplot all stage ran into this output:

Running smudgeplot v0.4.0dev
Task: all

Loading data

Masking errors using local aggregation algorithm
	            Total kmers: 12408991
	             Genomic kmers: 1196435
	             Genomic kmers in smudges: 36522
	             Sequencing errors: 11212556
	             Fraction or errors: 0.904
Inferring 1n coverage using grid algorithm
Best coverage to precision of 1/10: 18.70
Best coverage to precision of 1/100: 18.00
Best coverage to precision of 1/100 (just to be sure): 18.00

Creating centrality plot

Inferred coverage: 0.000

Creating smudge report
/home/j186c375/.local/lib/python3.11/site-packages/smudgeplot/smudgeplot.py:192: RuntimeWarning: divide by zero encountered in scalar divide
  As, Bs = round(covA / cov), round(covB / cov) # this is approximate labeling (modus of coverage / 1n coverage)
Traceback (most recent call last):
  File "/home/j186c375/.local/bin/smudgeplot", line 8, in <module>
    sys.exit(main())
             ^^^^^^
  File "/home/j186c375/.local/lib/python3.11/site-packages/smudgeplot/cli.py", line 321, in main
    smudges.local_agg_smudge_container = smudges.get_smudge_container(cov, smudge_size_cutoff, "local_aggregation")
                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/j186c375/.local/lib/python3.11/site-packages/smudgeplot/smudgeplot.py", line 192, in get_smudge_container
    As, Bs = round(covA / cov), round(covB / cov) # this is approximate labeling (modus of coverage / 1n coverage)
             ^^^^^^^^^^^^^^^^^
OverflowError: cannot convert float infinity to integer

Is the error because inferred coverage: 0? I'm not sure what is causing this crash.

[jychoilab]

I realized changing the -L parameter before this step allowed the plotting to happen so I'm just going to close this.

[KamilSJaron]

Hi @jychoilab,

I am glad you figured a workaround, but I think smudgeplot should happen data like this, so I will open the issue so we look at it at some point.

THanks for reporting.