Add Kraken2 2.17.2 and prebuilt Kraken2 databases

[kennedydane]

What does this PR do?

Adds Kraken2 2.17.2 (bio, compiled) alongside the existing 2.1.3, and adds prebuilt Kraken2 databases at /software/data/bio/kraken2.

• 2.17.2: built from the GitHub tag archive with install_kraken2.sh as before. The new k2 wrapper uses only the Python 3 standard library, so the system python3 is enough.
• Databases (-t kraken2,kraken2-db): Standard, Viral and MinusB (2026-06-26), EuPathDB48 (2023-04-07), and the four 16S sets (Greengenes 13.5, RDP 11.5, Silva 132, Silva 138). Each one is extracted to <name>_<date>/, gets an INSTALLED.ansible marker, and is skipped on later runs. The archive is deleted after extraction.
• Module: kraken2.lua now sets KRAKEN2_DB_PATH, so kraken2 --db standard_20260626 finds the database by name.

Checklist

• Linters pass locally: cd ansible && uv run yamllint . && uv run ansible-lint (Profile 'production' ... passed); ansible-playbook site.yaml --syntax-check also passes.

• CI is green on this PR.

• Double tag wired up: kraken2 on the include; kraken2-2.17.2 / kraken2-db on the leaf blocks (both with never). The hyphenated version tag matches the existing kraken2-2.1.3.

• Checksum set on every download. 2.17.2 uses sha512. Standard, Viral and MinusB use upstream's md5, taken from their published <collection>.md5 files. EuPathDB48 and the 16S sets publish no checksums, so their sha256 was computed from complete downloads, each size-checked against S3's Content-Length.

• Module template: updated kraken2.lua.

• Symbolic file modes.

• Uses the path variables (bio_data_dir, bio_dir, bio_modules).

• README software tables: Kraken2 isn't listed there today, so they're unchanged.

Notes for the maintainer

• Run as:
  • ansible-playbook site.yaml -t kraken2,kraken2-2.17.2 for the software.
  • ansible-playbook site.yaml -t kraken2,kraken2-db for the databases.
• Disk space: each step runs over every database before the next step starts, so all the archives (~98 GB) are still on disk while everything is extracted (~128 GB). Allow about 230 GB free on /software at the peak. Downloads are staged in /software/data/bio/kraken2 (tmp_dest), not ~/.ansible/tmp, so the home directory isn't used. Databases are listed smallest first, so the small ones finish before Standard starts. If a download is interrupted, a partial temp file may be left in that directory to delete by hand.
• Pre-placed archives: if an archive is already at /software/data/bio/kraken2/<basename> and its checksum matches, get_url skips the download. EuPathDB48 was pre-placed this way. Run as the user that owns these files, because the tasks set modes, extract into the directory and delete the archive.
• Extraction: tar is used instead of ansible.builtin.unarchive, which would decompress the ~80 GB archive twice. There's a # NOTE: at the task.
• Runs per version: the database block sits inside the per-version include, so it runs once for each Kraken2 version. Repeat passes only check the markers.
• Not changed: kraken2_dir's existing g=rx mode, and undated standard → standard_<date> symlinks.
• Upstream:
  • https://github.com/DerrickWood/kraken2/releases/tag/2.17.2 (the tag has no v prefix)
  • https://benlangmead.github.io/aws-indexes/k2

🤖 Generated with Claude Code

https://claude.ai/code/session_01GttZpVB7Vniruv7g6F6jjU

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.