hamronize summarize - local variable 'parsed_report' reference before assignment about hamronization

[fmaguire]

Sorry about the delay, didn't realise people were actually running into this issue in the wild!
This should now work as intended (release to follow imminently).

[gdemoro]

this bug is still present in the latest version (1.0.4)

[jfy133]

I would like to report the same issue, which is a problem for pipelines when we try to automate a lot of these. We are using hAMRonization heavily (because it's an awesome tool, thank you for it!) in nf-core/funcscan, but when testing this in a large pipeline run with ~126 samples, it caused the whole pipeline to fail due to this error.

The actual error arguably comes from the upstream per-tool subcommands, which produce 'nonsense' output of a single ]

Example, for abrigate output:

$ cat ABM006-megahit.txt
#FILE	SEQUENCE	START	END	STRAND	GENE	COVERAGE	COVERAGE_MAP	GAPS	%COVERAGE	%IDENTITY	DATABASE	ACCESSION	PRODUCT	RESISTANCE
$ hamronize \
    abricate \
    ABM006-megahit.txt \
     \
    --format json \
    --analysis_software_version 1.0.1 \
    --reference_database_version 2021-Mar-27 \
    > ABM006-megahit.txt.abricate.json
$ cat ABM006-megahit.txt.abricate.json 
]

So from a pipeline perspective, alternatively, it would also work at pipeline levels to have the check/failure at the tool specific subcommand. If it could produce an error message and a exit code other than 0 or 1, it would allow us to 'ignore' if the tool-level command fails.