veckit
Local scorer for the Virtual Embryo Challenge: score a submitted .h5ad
against the T1/T2/T3 metric panels, entirely offline, against reference files you supply. This package
never reads the official (held-out) validation/test data, and never imports anything that names where that
data lives -- it only knows how to compute a metric panel from arrays you hand it.
pip install veckit
Getting the real data
Register and download the released T1/T2/T3 stages from the official challenge site:
virtualembryo.ai/challenge/data (see also
aristoteleo/virtualembryo). veckit_tutorial.ipynb in this
repo also ships a few tiny (150-cell) samples if you just want to try the tool first.
CLI
Full, meaningful panel (a real prediction, scored against the real preceding stage):
veckit --task T1 --input pred.h5ad --target T1/9.5.h5ad --reference T1/8.5.h5ad
veckit --task T2 --setting heart --input pred.h5ad --target b.h5ad --reference a.h5ad
veckit --task T3 --input pred.h5ad --target mab21l2_ko.h5ad --wt wt.h5ad
Quick copy_last / wt_identity check (--input doubles as --reference/--wt when omitted):
veckit --task T1 --input T1/8.5.h5ad --target T1/9.5.h5ad
Python
from veckit import score
result = score(task="T1", input="pred.h5ad", target="T1/9.5.h5ad", reference="T1/8.5.h5ad")
print(result["metrics"])
--target (all tasks) is the pseudo target your --input prediction is scored against.
--reference (T1/T2) / --wt (T3) is the reference expression de_score/de_direction/severity_slope
are computed relative to -- these are PRIMARY metrics ("did you predict the right change", not just
"does this look plausible"), so pass a real one whenever you're scoring a real model; omitting it defaults
to --input itself, which is only correct when you're deliberately testing a no-change baseline
(copy_last/wt_identity) -- those metrics then correctly read as exactly 0, matching the official
baseline tables.
This is not a preview of your real competition score. Whatever you pass as --target is, by
definition, data you already had — so this only tells you the scoring pipeline runs and your submission
format is valid, not how well you'll do on the real held-out target.
More
Full task definitions and metric rationale: virtualembryo.ai and aristoteleo/virtualembryo.
Metadata
Release files for veckit 0.1.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| veckit-0.1.2.tar.gz | 36.7 kB | Without provenance | Details for veckit-0.1.2.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| veckit-0.1.2-py3-none-any.whl | Python 3 | none | any | Details for veckit-0.1.2-py3-none-any.whl |
Total release size: 80.4 kB
Release files / veckit-0.1.2.tar.gz
| Download URL | veckit-0.1.2.tar.gz |
|---|---|
| Size | 36.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
4fde73b6730e5c24a2b73ae49a77cef6401eca81f1bbe7e71acd458db8f84530
|
|
BLAKE2b-256 checksum How to use checksums |
a6664c9a37efeebfc2f830432ca849dbcdb0e0f643d127699a27435ff69796be
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.9
|
Release files / veckit-0.1.2-py3-none-any.whl
| Download URL | veckit-0.1.2-py3-none-any.whl |
|---|---|
| Size | 43.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
2eb3c19caca571f6c5c4deb936bfd697fb7b979e6470b9eb666b244bfd29d3bd
|
|
BLAKE2b-256 checksum How to use checksums |
6fd7a46399dce07f2837915595cbab9fb792dbde85b818911727811d9bb4803e
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.9
|
Security for veckit 0.1.2
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.