Skip to content

deps: update ghcr.io/hassansaei/vntyper docker tag to v2.0.50 - #459

Open
renovate[bot] wants to merge 1 commit into
mainfrom
renovate/ghcr.io-hassansaei-vntyper-2.x
Open

renovate[bot] wants to merge 1 commit into
mainfrom
renovate/ghcr.io-hassansaei-vntyper-2.x

Conversation

@renovate

@renovate renovate Bot commented Nov 16, 2026

Copy link
Copy Markdown
Contributor

This PR contains the following updates:

Package Update Change
ghcr.io/hassansaei/vntyper patch 2.0.47 → 2.0.50

Release Notes

hassansaei/VNtyper (ghcr.io/hassansaei/vntyper)

v2.0.50

Compare Source

Deployment to vntyper.org is deliberately deferred. VNtyper-Online stays on 2.0.47; vntyper-online-backend keeps its 2.0.47 digest pin.

Image: ghcr.io/hassansaei/vntyper:2.0.50@sha256:d9cd90cd1b717d6c5fcebf3ce09d37ec4e86b4a6dc23074bb0b2bdce5af0e21c (revision d260faf). PyPI: vntyper==2.0.50.

Web service: anonymous research donations
  • Donation schema vntyper-donation/2: POST /donations/ takes a JSON record of enums, bins and booleans (no file, no run ID, no sample name, no hashes), refuses extra fields (422), non-JSON bodies (415), bodies over 2 KB (413) and answers 403 while ENABLE_DONATIONS is off. The response is {"status": "recorded"}; no identifier is returned or stored.
  • Storage: rows go to donations_v2 with the whitelisted fields, a server-side received_quarter and a random key. The v1 donations table, the archive validation and the integrity check on donated ZIPs are removed. With donations enabled the API refuses to start without a reachable POSTGRES_URL instead of falling back to an in-memory store.
  • Aggregates: /donations/aggregates/ returns counts only, null below k=5 (k=10 for confirmed positives and tool calls), with complementary suppression.
  • Logging: the API drops uvicorn access-log lines for /donations/, and a rate-limited donation request is logged without the client address. In production a missing PSEUDONYM_SECRET is now logged as an ERROR at startup of the API and every worker.
  • Research pipeline decisions: No caller thresholds, result rows, tiers or screening messages changed. docker/app only.

v2.0.49: VNtyper v2.0.49

Compare Source

Deployment to vntyper.org is deliberately deferred. The manuscript revision and VNtyper-Online stay on 2.0.47; vntyper-online-backend keeps its 2.0.47 digest pin.

Image: ghcr.io/hassansaei/vntyper:2.0.49@sha256:9144f02d1302f8132b13a7a42b026307759594bd08020c1ab850cc7dc9aa3e76 (revision 7135a19). PyPI: vntyper==2.0.49.

Runtime
  • k-mer counting threads (#​365): KAnalyze's split stage, the bottleneck on larger inputs, now gets half of --threads (at most 4); the total never exceeds --threads and the k-mer count file is byte-identical for every split. On an exclusive Xeon E5-2667 v4 node with a 1.25 GB BAM, normal mode runs 13% faster at 8 threads and 31% faster at 4; KAnalyze peak memory rises from 3.3 to 5.6 GB at 8 threads, inside the existing 12 GB Java heap.
  • Input fingerprint in the background (#​365): the SHA-256 of the input (and of the reference for FASTQ/CRAM) is computed while samtools prepares the input, instead of before it. A CRAM reference is no longer hashed twice. Fast mode on the same BAM runs 25% faster, and with code-adVNTR 18% faster.
  • Research pipeline decisions: No caller thresholds, result rows, tiers or screening messages changed; 427 output files on the test BAMs and all 232 paired cluster runs match 2.0.48.

v2.0.48: VNtyper v2.0.48

Compare Source

Deployment to vntyper.org is deliberately deferred. The manuscript revision and VNtyper-Online stay on 2.0.47; vntyper-online-backend keeps its 2.0.47 digest pin. Before any later deployment, set one stable PSEUDONYM_SECRET on the API and every worker (#​355).

Image: ghcr.io/hassansaei/vntyper:2.0.48@sha256:451797a6523b8ef12b48e7d57d596971336b29953f05811a6163f5e36fcd3689 (revision 156c6f5). PyPI: vntyper==2.0.48.

Vrbacká 2025 citation and published variant names (decision profile revision 3)
  • Corrected DOI (#​352): every Vrbacká et al. 2025 citation in known_variants cited doi:10.1101/2024.11.14.623419, an unrelated preprint. It now cites doi:10.1101/2025.09.06.673538.
  • Published names (#​353): three entries were shifted one base 5' by a MucOneUp convention before 0.50.0. 53_54insG and 53_54insA become 54_55insG and 54_55insA; 57_58insG, the same allele as 58_59insG, is removed and 58_59insG cites Olinger 2020 and Vrbacká 2025; 54dupG, not in the paper's Table 1, is removed.
  • Decision profile revision 3: known_variants decides Tier A eligibility, so the packaged profile moves to revision 3. Summaries recorded with revision 2 (2.0.1 to 2.0.47) are still accepted by vntyper cohort.
  • Research pipeline decisions: A call named 57_58insG, 53_54insG, 53_54insA or 54dupG is now representation-of-caller-call instead of known-variant and cannot reach Tier A. On the development simulation 37 rows change only their nomenclature flags and note; no name, tier or call changes. Caller thresholds and screening messages are unchanged.
Web service: job timing, job summary and keyed pseudonyms
  • Job timing (#​355): /job-status/{id}/ also returns submitted_at, started_at and completed_at (ISO 8601 UTC, null when not recorded).
  • GET /job-summary/{id}/ (#​355): returns the run's own pipeline_summary.json from the result archive, with the same identifier check, rate limit, file-safety rules and retention as /download/.
  • Keyed pseudonyms (#​355): the usage-statistics caller hash and the donation sample pseudonym are HMAC-SHA256 under PSEUDONYM_SECRET. Set one stable value on the API and every worker; unset, each process uses a random key and logs a warning, so unique-user counts overcount. Existing hashes and pseudonyms do not carry over.
  • Research pipeline decisions: No caller thresholds, result rows, tiers or screening messages changed. docker/app only.
Cohort discovery and input errors
  • Symlinked and unreadable sample directories (#​357): vntyper cohort follows directory symlinks below an input root (each real directory once) and logs every directory it cannot read. Before Python 3.13 the recursive search skipped both without a word, so completed samples were missing from the exports.
  • Empty input layout (#​358): an input with no reads in the target region and no unmapped reads now fails with a message saying so and what to check (region coverage, --reference-assembly, a blank or failed library), instead of claiming that reads would be dropped.
  • Research pipeline decisions: No caller thresholds, result rows, tiers or screening messages changed.
Test suite
  • Host- and dependency-independent web tests (#​359): the API tests no longer depend on the host's free disk, and stay green on pytest 9.2 and Starlette 1.10 (an fstat patch leaked into pytest's capture, and an ASGI test client never yielded to Starlette's disconnect listener).
Subthreshold note covers dupC candidates
  • dupC below the floor (#​362, #​363): a dupC row (ALT == alt_filtering.gg_alt_value) below the reporting floor also fails alt_filter_pass, because the shipped gg_depth_score_threshold equals the floor, so it never qualified for the subthreshold note. Such a row now qualifies when its ALT failure is the GG depth rule alone. Those no-call samples now carry the note and the negative_subthreshold state. Excluded ALTs and artifact-flagged rows stay ineligible. A dupC row at or above the floor that fails only a stricter custom GG threshold is not subthreshold.
  • Research pipeline decisions: No caller thresholds, calls, result rows or tiers changed. A subthreshold candidate is not a call.
Grade and subthreshold state in machine-readable outputs
  • pipeline_summary.json (#​356): records confidence_grade, the grade the report prints, and subthreshold_signal with state (negative_subthreshold or none), best_depth_score, event_count and support_count. The report computes both and re-anchors the summary before rendering, so the printed integrity digest covers them. vntyper report never rewrites a supplied summary.
  • Support count for the subthreshold note: the signal now carries the alternate-allele read support of the strongest below-floor candidate. A no-call run with a note writes the numbers to kestrel/subthreshold_signal.json; the ## line in kestrel_result.tsv is unchanged.
  • Cohort statistics: cohort_stats.csv, .tsv and .json append confidence_grade, subthreshold_state, subthreshold_best_depth_score, subthreshold_event_count and subthreshold_support_count after the existing columns. Summaries written before this release give empty cells.
  • Research pipeline decisions: No caller thresholds, result rows, tiers, grades or screening messages changed. A subthreshold candidate is not a call.

🤖 Generated with Claude Code


Configuration

📅 Schedule: (UTC)

  • Branch creation
    • Between 12:00 AM and 05:59 AM, only on Monday (* 0-5 * * 1)
  • Automerge
    • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.


  • If you want to rebase/retry this PR, check this box

This PR was generated by Mend Renovate. View the repository job log.

@renovate
renovate Bot requested a review from GeiserX as a code owner November 16, 2026 00:20
@renovate renovate Bot added automated Opened automatically by a bot dependencies Dependency updates labels Nov 16, 2026
@coderabbitai

coderabbitai Bot commented Nov 16, 2026

Copy link
Copy Markdown

Important

Review skipped

Bot user detected.

To trigger a single review, invoke the @coderabbitai review command.

Run configuration

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • Trigger review
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

automated Opened automatically by a bot dependencies Dependency updates

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0 participants