Versions and citation

latest vs dated weekly vs monthly releases, how to pin a version, and how to cite bioc-on-ice.

bioc-on-ice publishes three kinds of versions. They hold the same tables in the same layout. They differ in how long they last and how you use them (ADR-0016). Releases lists the ones available now.

Version	Where	Changes?	Kept	Use it by
`latest/`	`https://data.bioconice.cancerdatasci.org/latest/`	with each weekly release	always the newest	querying in place
Dated	`https://data.bioconice.cancerdatasci.org/v{YYYY-MM-DD}/`	never	the 4 newest, about four weeks	querying in place
Monthly	Hugging Face dataset `seandavis/bioconice`, git tag `YYYY-MM`	never	permanently	downloading, then querying locally

latest/ and the dated folders are served for querying in place: DuckDB ATTACH, read_parquet, the playground and the Explorer all read them over HTTPS. The monthly releases are the long-term archive and the version to cite. Download one, then query it locally.

Note

Releases are still made by hand, and dated folders are kept 30 days. The weekly schedule, keep-4 retention and the monthly Hugging Face archive start once a release passes the release gate (#214). Until then there are no monthly releases.

Which one to use

Exploring, or building something that should track the sources: use latest/.
An analysis you’ll rerun in the next few weeks: pin a dated folder.
A paper, a shared pipeline, a Bioconductor release, or anything you’ll cite: download a monthly release.

Pin a dated folder

Replace latest/ with a date. Each dated folder’s bioconice.duckdb reads its own Parquet files, so the views stay pinned too:

ATTACH 'https://data.bioconice.cancerdatasci.org/v2026-10-31/bioconice.duckdb' AS bioc (READ_ONLY);

To see which dates are available, open Releases or the folder listing. Older folders are deleted, so use a monthly release for anything longer than a few weeks.

Use a monthly release

Once a month, the first weekly release that passed the gate is copied to the Hugging Face dataset seandavis/bioconice and tagged with its month (e.g. 2026-11). Files sit at the repository root in the same layout as R2 (annotation/gene.parquet, …).

Download the release with the hf command-line tool (in huggingface_hub; uvx --from huggingface_hub hf runs it without installing), then query the local files:

hf download seandavis/bioconice --repo-type dataset --revision 2026-11 \
  --local-dir bioconice-2026-11
duckdb -c "SELECT count(*) FROM 'bioconice-2026-11/annotation/gene.parquet'"

--include 'annotation/*' downloads only some of the files. A Hugging Face token (hf auth login) raises the download rate limits.

Don’t query the archive in place with hf:// paths. Hugging Face limits anonymous requests, one query can make hundreds, and repeated queries soon fail with HTTP 429. For in-place queries use latest/ or a dated folder.

Cite bioc-on-ice

Cite the monthly release you used by its tag (e.g. seandavis/bioconice@2026-11), and its DOI if the dataset page for that tag shows one.

If you used a dated folder instead, record its URL and the release and git_sha from its manifest.json. The folder will be gone within weeks, so a monthly release is the better thing to cite. Each table also names the upstream source and version it came from; cite those sources as they ask (for example, BugSigDB is CC BY 4.0).

The manifest

Every version has a manifest.json:

curl -s https://data.bioconice.cancerdatasci.org/latest/manifest.json

DuckDB reads it directly. The provenance fields of the current latest/:

SELECT release, published_at, git_sha
FROM read_json('https://data.bioconice.cancerdatasci.org/latest/manifest.json');

And the row count of each published table:

SELECT t.schema, t."table", t.row_count
FROM read_json('https://data.bioconice.cancerdatasci.org/latest/manifest.json'),
     unnest(tables) AS u(t)
ORDER BY t.schema, t."table";

release is the release date and git_sha the commit of the pipeline code that built it.
provenance lists each upstream source with its version, URL, retrieval time and row count.
tables gives each table’s path, row_count, bytes, sha256, comment and column docs.