Workflow control

Targeting specific rules

Aviary uses Snakemake under the hood, which means you can target specific rules rather than running a full module. Use -w/--workflow to specify the rule to run up to:

aviary recover -w rosella --assembly scaffolds.fasta \
    -1 reads_1.fq.gz -2 reads_2.fq.gz --output output_dir/ --max-threads 12

All steps leading up to the targeted rule will still run if they haven't been completed yet.

The available rules for each module are defined in the module's Snakefile under aviary/modules/.

Resuming interrupted runs

If an Aviary workflow is interrupted, re-running the same command will resume from the last completed step — Snakemake tracks completed outputs automatically.

If Snakemake reports that the directory is locked, first confirm that no Aviary or Snakemake process is still using it. Then pass Snakemake's unlock operation:

aviary complete \
  -1 sample_R1.fastq.gz \
  -2 sample_R2.fastq.gz \
  --output output_dir \
  --snakemake-cmds '--unlock'

Do not unlock an active run.

Dry run

Test the workflow order and verify conda environments without executing any rules:

aviary recover -1 reads_1.fq.gz -2 reads_2.fq.gz --dry-run

Skipping steps

Several steps can be skipped to speed up runs when not needed:

Flag	Effect
--skip-qc	Skip read quality control
--skip-taxonomy	Skip GTDB-tk taxonomy assignment
--skip-abundances	Skip CoverM abundance calculations
--skip-singlem	Skip SingleM recovery assessment
--binning-only	Stop after binning, skip all downstream steps

Temporary file cleanup

By default Aviary removes temporary files (BAM files, intermediate FASTQs) on completion. To keep them for partial reruns:

aviary recover -1 reads_1.fq.gz -2 reads_2.fq.gz --clean false

Note: Keeping temporary files avoids re-generating them on reruns but will use significantly more disk space.