Skip to content

Releases

v0.4.1

Choose a tag to compare

@github-actions github-actions released this 09 Oct 14:44
5e23597

0.4.1

Fixed

  • JitterSimulator.transform before fit now raises instead of returning the
    data unchanged, which would have scored as perfect jitter stability. This
    affects the Rust JitterSimulator and pypasteur.JitterSimulator.
  • The pasteur-model README example passes Some(1) for the positive class
    index, matching the 0.4.0 Option<usize> signature.
  • Positive-class-index errors from OnnxModel::from_file name the parameter
    as well as the CLI flag.

Added

  • pasteur-cli --version.
  • pypasteur.__version__.

What's Changed

Full Changelog: v0.4.0...v0.4.1

v0.4.0

Choose a tag to compare

@github-actions github-actions released this 09 Oct 13:33
c5e59ba

0.4.0

Added

  • Multiclass and multilabel models. metadata.json gains task_type
    (binary | multiclass | multilabel) and output.{name, classes, thresholds}. simulate, evaluate, and compare take --task, and
    --positive-group-id is repeatable (one cohort per model output, in
    output.classes order). All sklearn-onnx probability formats are read:
    [n, K] tensors, ZipMaps with int or string keys (matched to classes by
    name), and MultiOutputClassifier's per-label tensor sequence.
  • Every metric is computed per class/label; headline fields are macro
    averages. A new multi section in the result JSON reports micro ROC AUC,
    worst-label resiliency, a jitter decision-flip rate, the share of flipper
    pairs that never flip, the multiclass detour rate, labels excluded for
    having one class, and per-label scores. Binary output has no multi key.
  • Multiclass jitter stability sums the per-class prediction variances and
    halves them, so it equals the binary score for 2 classes and does not
    depend on how many classes the model has. Multilabel averages per label.
  • Multiclass flipper finds where the pairwise margin p[b] − p[a] crosses 0;
    multilabel flipper samples pairs per label (new pair_label grid column)
    and crosses each label's own threshold.
  • Single-target regression models. task_type: "regression" in
    metadata.json (default output variable, then predictions), and
    --task regression with true values from --targets <parquet> --target-col <name>, joined by ID as text. baselines are rmse, mae
    and r2; resiliency is blackout R² ÷ clean R² (the share of skill kept)
    under metric_based.r2; jitter variance is scaled by the target's variance.
    Flipper pairs patients across a clinical cutoff (--flip-threshold, in
    target units, passed to both simulate and evaluate) and finds where the
    prediction crosses it. A new regression section reports these in target
    units; classifier JSON is unchanged.

Changed

  • Breaking (library): Model::predict_proba returns a DataFrame with
    one column per output, SimulationVariant.y is a DataFrame label matrix,
    EvaluationConfig gains task, and OnnxModel::from_file takes
    positive_class_index: Option<usize>. TaskType gains Regression and
    EvaluationResult gains an optional regression field. CLI usage for
    binary models is unchanged.
  • evaluate/compare --labels is no longer required at parse time; it is
    required for classifiers and refused for regression. --flip-threshold
    still defaults to 0.5 for classifiers.
  • ROC AUC handles tied scores. Ties count half (Mann–Whitney) instead of
    depending on row order. Results without ties agree with the previous
    implementation to floating-point rounding.
  • The probability output is resolved when the model loads. A model without
    output_probability/probabilities (or the contract's output.name) now
    fails at load with the list of its real outputs, instead of at first
    prediction.
  • compare refuses two --model files with the same file name instead of
    panicking on duplicate prediction columns.

Docs

  • New pages for hospital IT and model reviewers: security and data handling
    (data flow, network behaviour, offline install, PHI in outputs), metric
    definitions and caveats, and an end-to-end model-selection walkthrough.
  • One-page usage overview PDF for IT review.
  • Fixed broken schema links and removed internal references from the READMEs
    and --help text.

What's Changed

New Contributors

Full Changelog: v0.3.0...v0.4.0

v0.3.0

Choose a tag to compare

@github-actions github-actions released this 29 Sep 14:14

0.3.0

First open-source release. Versions 0.1 and 0.2 were internal and are not in
this repository's history.

Added

  • pasteur-core, pasteur-cli, pasteur-hf, and pasteur-model.
  • pypasteur (import pypasteur). Wheels are abi3 for CPython ≥ 3.12:
    Linux x86_64/aarch64, macOS x86_64/arm64, and Windows x64, plus an sdist.
  • A release publishes those artifacts to PyPI, the four crates to crates.io,
    and the GitHub Release.

Behavior

  • Blackout masks companion columns. BlackoutConfig.companions is
    required. apply_mask nulls the feature and its companions together, and
    errors if a companion is not in the frame. A blacked-out assay that kept
    {assay}_measured == 1 is not a row an EHR emits, so flip rates would
    measure extrapolation rather than robustness.
    pasteur-cli simulate --blackout-companion <COL> is repeatable.
  • Null fill defaults to NaN. OnnxModel::from_file takes null_fill,
    and evaluate / compare take --null-fill. Resolution order: flag →
    the model's metadata.json input.absent_sentinel → NaN. 0.0 TSH is
    not absence, it is profoundly suppressed. Models that are not NaN-native
    error instead of returning NaN probabilities that poison ROC AUC.
    A model trained with 0.0 as the missing-value token should declare
    "absent_sentinel": 0.0 under input in metadata.json, or pass
    --null-fill 0.
  • Feature order is validated against the loaded model. from_file
    resolves the input name against the session's actual inputs, checks the
    input width, and — when a metadata.json sits beside the .onnx —
    compares input.feature_order element-wise. Wrong width already errored
    inside ONNX Runtime; wrong order at the right width did not.

Changed

  • Licensed under BSD-3-Clause. Crates carry crates.io metadata and inherit
    version, license, and MSRV from the workspace.
  • Builds on stable Rust ≥ 1.95. 1.95 is the floor polars 0.54 actually
    needs (polars-ooc uses APIs stabilized in 1.95), even though no
    dependency declares it.

Full Changelog: https://github.com/Krv-Labs/pasteur/commits/v0.3.0