You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Added header validation to processFOFN() and processAccessions() — both now check that the first line contains all required columns before processing data rows
processFOFN() requires: sample, runtype, r1, r2, se, ont, assembly, genome_size, species
Missing columns produce a clear error listing found vs expected columns and suggest using bactopia prepare or bactopia search to generate properly formatted files
Fixed SE/ONT read detection in BactopiaTools._collectInputs():
SE reads now use {sample}_SE.fastq.gz instead of {sample}.fastq.gz
ONT reads now use separate {sample}_ONT.fastq.gz file path
ONT detection changed from checking for NanoPlot HTML report to ont.txt marker file
PE read detection now correctly excludes ONT samples
Fixed JsonSchemaValidator rejecting string "true"/"false" values for boolean parameters — CLI arguments arrive as strings, so the validator now accepts these as valid booleans