Module download

Module: download

Downloads the standardized databases required by CheckAMG and sets them up in a destination directory. Two databases are retrieved by default: the smaller (21-40 GB) annotate database (curated profile HMMs and cutoff files) and the larger (76 GB) de-novo database (the pre-trained CheckAMG-PST model checkpoint, reference embeddings, FAISS index, and labels).

The databases are pre-built and packaged to ensure reproducibility across CheckAMG and database versions, and to avoid failures caused by upstream download links changing or disappearing.

About 116 GB of free disk space is required during download. This can be reduced to about 97 GB after downloading finishes if the human-readable HMM files are removed with --rm-hmm (CheckAMG only needs the binary HMM files).

checkamg download -d /path/to/db/destination --rm-hmm

The space required for the database can be further reduced to about 21 GB if you do not plan on running checkamg de-novo or checkamg end-to-end and skip downloading the de novo database with --no-download-denovo-db.

Parameters

Argument	Default	Description
-d, --db-dir	required	Directory where the CheckAMG databases will be placed.
-f, --force	off	Force re-download even if the databases already exist.
-r, --rm-hmm	off	Remove the human-readable HMM files after downloading to save space. CheckAMG only needs the binary files.
-v, --db-version	latest	Exact CheckAMG database version identifier to download (overrides the latest compatible version).
--download-annotate-db / --no-download-annotate-db	enabled	Download the database required by annotate. Use --no-download-annotate-db to skip it if it already exists.
--download-denovo-db / --no-download-denovo-db	enabled	Download the larger database required by de-novo. Use --no-download-denovo-db if you only need annotate.

Notes

• Pass the same destination directory to -d/--db-dir for annotate, de-novo, and end-to-end.
• If a download is interrupted, re-running checkamg download resumes where possible; use --force to start over.