Skip to content
gp3mlpy 0.1.0 Python 3.11+ gp3ml 0.3.0 parity target

Governance-first predictive modelling

Build models whose validation logic matches the scientific claim.

Leakage-resistant, group-aware modelling for Gazepoint research workflows in Python — with explicit generalization targets, auditable predictors, governed decisions, external validation, and reproducible release evidence.

100% statements 100% branches 125 tests 0 parity failures
terminal
$ python -m pip install gp3mlpy
Successfully installed gp3mlpy-0.1.0
 
import gp3mlpy as gp
gp.__version__
'0.1.0'
gp.r_reference_version
'0.3.0'

Predictive modelling with explicit scientific contracts

gp3mlpy is the Python port of gp3ml 0.3.0. Instead of treating validation as a generic train/test operation, it makes the target of generalization, feature provenance, fitted operations, uncertainty, and reporting evidence visible parts of the analysis.

01

Declare what must generalize

Choose new participants, new stimuli, or both from the scientific claim. Splits that violate that target are rejected.

02

Keep fitted operations local

Preprocessing, tuning, calibration, thresholds, and related operations remain inside the correct analysis or fold structure.

03

Make predictors auditable

Feature manifests record source, role, and permission before model performance is known.

04

Leave an evidence trail

Diagnostics, model cards, external-validation reports, checksums, environment capture, handoffs, and release evidence remain inspectable.

The governed lifecycle

01Declareoutcome · unit · use · target
02Auditprovenance · roles · leakage
03Splitgroup-aware holdout or resampling
04Fitfold-local preprocessing and tuning
05Evaluateperformance · uncertainty · robustness
06Reportvalidation · provenance · release evidence

New to the package? The Quickstart takes you from installation to a validated grouped workflow, then points to the next layer only when you need it.

Choose the route that matches your study

Generalization

New participants

Keep assessment participants completely unseen during fitting.

Participant workflow →

Model selection

Nested grouped resampling

Separate inner tuning from outer assessment while retaining grouping.

Nested workflow →

Transportability

External validation

Declare an external dataset, quantify shift, and report transportability explicitly.

External validation →

Decision layer

Governed thresholds

Keep threshold origin, error costs, calibration source, and abstention visible.

Decision governance →

A validated release, not just a feature list

Python quality floor

100% statement and branch coverage

  • 4,020 / 4,020 executable statements
  • 1,700 / 1,700 measured branches
  • 0 partial branches
  • 125 passing Python tests
  • Ubuntu, Windows, macOS × Python 3.11, 3.12, 3.13
  • Ruff, mypy, strict MkDocs, build/Twine, installed-wheel API checks

CI permanently enforces --cov-branch --cov-fail-under=100.

Frozen stable API

67 PASS · 4 expected differences · 0 FAIL

All 71 stable exports were exercised against the SHA-256-verified gp3ml 0.3.0 release archive.

The four expected differences preserve safer or functioning Python behavior instead of reproducing frozen-R recycling or reference defects.

Read the parity status →

Diagnostics are part of the evidence

The documentation build regenerates synthetic plot fixtures from the current package. The plots visualize declared analysis objects; plotting does not perform hidden model selection.

Decision-threshold evaluation plot
Decision thresholds
Inspect declared candidate thresholds and decision consequences.
Dataset-shift audit plot
Dataset shift
Separate predictor shift from calibration and performance change.
Governance profile audit plot
Governance evidence
Surface controls with evidence and controls that still need review.

Where to go next

Learn

Key concepts

Understand generalization targets, provenance, leakage resistance, parity, and the scientific boundary.

Apply

Workflow articles

Follow end-to-end, generalization, governance, validation, and reproducibility workflows.

Inspect

Workflow API map

Find functions by research stage rather than alphabetically.

Reference

Complete API index

Browse all 127 compatibility exports and their dedicated reference pages.

Hard scientific boundary

gp3mlpy is for explicitly observed, non-sensitive outcomes. It must not be used for person identification, biometric authentication, health/diagnosis inference, protected-attribute inference, or direct/indirect inference of emotion, stress, personality, deception, cognition, comprehension, intent, or other mental states.