Package index • gp3ml

Package

gp3ml gp3ml-package: gp3ml: Governance-First Predictive Modelling for 'Gazepoint' Research

Leakage auditing

Audit partition leakage and export structured results.

audit_gazepoint_ml_leakage(): Audit leakage between predictive-analysis partitions
print(<gazepoint_ml_leakage_audit>): Print a Gazepoint ML leakage audit
write_gazepoint_ml_leakage_audit_csv(): Write a Gazepoint ML leakage-audit table to CSV

Feature provenance

Record and validate intended predictor provenance.

create_gazepoint_feature_manifest(): Create a Gazepoint feature-provenance manifest
validate_gazepoint_feature_manifest(): Validate a Gazepoint feature-provenance manifest
print(<gazepoint_feature_manifest_validation>): Print feature-manifest validation
write_gazepoint_feature_manifest_csv(): Write a Gazepoint feature manifest or validation table to CSV

Group-aware splitting

Create and validate deterministic grouped holdout partitions.

split_gazepoint_ml_data(): Create a deterministic group-aware Gazepoint holdout split
validate_gazepoint_ml_split(): Validate a group-aware Gazepoint holdout split
print(<gazepoint_ml_split>): Print a group-aware Gazepoint split
print(<gazepoint_ml_split_validation>): Print group-aware split validation
write_gazepoint_ml_split_csv(): Write group-aware split tables to CSV

Group-aware resampling

Create, validate, audit, and export deterministic grouped folds.

create_gazepoint_group_folds(): Create deterministic group-aware Gazepoint resampling folds
validate_gazepoint_group_folds(): Validate group-aware Gazepoint resampling folds
audit_gazepoint_group_folds(): Aggregate leakage audits across group-aware folds
print(<gazepoint_group_folds>): Print group-aware Gazepoint resampling folds
print(<gazepoint_group_folds_validation>): Print group-aware fold validation
print(<gazepoint_group_folds_audit>): Print aggregated group-fold leakage auditing
write_gazepoint_group_folds_csv(): Write group-aware resampling tables to CSV

Resampling diagnostics

Diagnose fold balance, coverage, exclusions, and outcomes.

diagnose_gazepoint_group_folds(): Diagnose group-aware Gazepoint resampling folds
validate_gazepoint_fold_diagnostics(): Validate Gazepoint fold diagnostics
print(<gazepoint_fold_diagnostics>): Print Gazepoint fold diagnostics
print(<gazepoint_fold_diagnostics_validation>): Print Gazepoint fold-diagnostics validation
write_gazepoint_fold_diagnostics_csv(): Write Gazepoint fold diagnostics to CSV files

Repository-aware fold evaluation

Fit fold-local preprocessing and models across materialized grouped folds.

evaluate_gazepoint_group_folds() print(<gp3ml_resample_evaluation>): Evaluate a governed model specification across materialized grouped folds
collect_gazepoint_fold_predictions(): Collect predictions from a grouped-fold evaluation
summarize_gazepoint_resample_performance() print(<gp3ml_resample_performance_summary>): Summarize repeated grouped-resampling performance
validate_gazepoint_resample_evaluation() print(<gp3ml_resample_evaluation_validation>): Validate a grouped-fold evaluation result
write_gazepoint_resample_evaluation(): Write grouped-fold evaluation tables

Governed comparison and tuning

Materialize, evaluate, compare, and review explicit model candidates.

create_gazepoint_tuning_grid() print(<gp3ml_tuning_grid>): Create an explicit governed tuning grid
tune_gazepoint_model() print(<gp3ml_model_tuning>): Evaluate every governed candidate on the same grouped folds
compare_gazepoint_models(): Compare governed model candidates without selecting a winner
select_gazepoint_model() print(<gp3ml_model_selection>): Select a governed candidate using an explicit metric and direction
validate_gazepoint_model_tuning() print(<gp3ml_model_tuning_validation>): Validate governed tuning results
write_gazepoint_model_tuning(): Write governed tuning and selection tables

Nested grouped resampling

Isolate inner tuning inside every outer analysis partition.

create_gazepoint_nested_folds() print(<gp3ml_nested_folds>): Create nested grouped resampling from mature outer folds
audit_gazepoint_nested_resampling() print(<gp3ml_nested_resampling_audit>): Audit nested grouped resampling for outer-assessment leakage
validate_gazepoint_nested_folds() print(<gp3ml_nested_folds_validation>): Validate nested grouped folds
evaluate_gazepoint_nested_resampling() print(<gp3ml_nested_evaluation>): Evaluate nested grouped resampling with inner governed tuning
validate_gazepoint_nested_evaluation() print(<gp3ml_nested_evaluation_validation>): Validate a nested evaluation
write_gazepoint_nested_evaluation(): Write nested-resampling evaluation tables

Target-aligned uncertainty

Record observation, cluster, fold, and repeat uncertainty without relabelling units.

bootstrap_gazepoint_metrics_by_unit() print(<gp3ml_target_uncertainty>): Generalization-target-aligned bootstrap uncertainty
summarize_gazepoint_resample_uncertainty() print(<gp3ml_resample_uncertainty>): Summarize uncertainty across folds or repeats
validate_gazepoint_target_uncertainty() print(<gp3ml_uncertainty_validation>): Validate target-aligned uncertainty metadata
write_gazepoint_target_uncertainty(): Write target-aligned uncertainty tables

External validation and transportability

Declare independent data and report performance, drift, schema, and coverage.

declare_gazepoint_external_dataset() print(<gp3ml_external_dataset_declaration>): Declare an external dataset and its independence status
evaluate_gazepoint_external_transportability() print(<gp3ml_transportability_report>): Evaluate external transportability and validation status
validate_gazepoint_transportability() print(<gp3ml_transportability_validation>): Validate an external transportability report
write_gazepoint_transportability_report(): Write an expanded transportability report

Synthetic workflows and release reporting

Deterministic demonstrations and release-ready governance records.

simulate_gazepoint_governed_data(): Simulate governed synthetic Gazepoint-derived data
create_gazepoint_synthetic_manifest(): Create a synthetic governed feature manifest
create_gazepoint_synthetic_task(): Create one of the governed synthetic demonstration tasks
create_gazepoint_release_model_card() print(<gp3ml_release_model_card>): Create a release-ready governed model card
write_gazepoint_release_model_card(): Write a release-ready governed model card
create_gazepoint_release_evidence() print(<gp3ml_release_evidence>): Create a release evidence manifest

Governed modelling core

Task governance, preprocessing, model fitting, performance, calibration, uncertainty, external validation, model cards, and reproducibility reporting.

declare_gazepoint_task(): Declare a governed Gazepoint prediction task
assert_gp3ml_use_case(): Assert that a task is within the permitted gp3ml scope
validate_gazepoint_ml_roles(): Validate outcome, predictor, identifier, and grouping roles
gp3ml_prohibited_uses(): Prohibited gp3ml uses
fit_gazepoint_preprocessor(): Fit a fold-local preprocessing engine
bake_gazepoint_preprocessor(): Apply a fitted preprocessing engine
gp3ml_available_engines(): List available model engines
integrate_black_box_model(): Integrate a controlled black-box model engine
fit_gazepoint_model(): Fit a governed Gazepoint model
train_gazepoint_classifier(): Generic governed binary-classifier training wrapper
predict(<gp3ml_model>): Predict from a gp3ml model
fit_gazepoint_deep_model(): Fit an optional governed deep-learning model through keras3
gazepoint_classification_metrics(): Binary classification metrics
gazepoint_regression_metrics(): Regression metrics
gazepoint_performance_metrics(): Task-aware performance metrics
bootstrap_gazepoint_metrics(): Bootstrap uncertainty intervals for performance metrics
fit_gazepoint_calibrator(): Fit a probability calibrator
apply_gazepoint_calibrator(): Apply a fitted probability calibrator
assess_gazepoint_calibration(): Calibration assessment with bootstrap uncertainty
evaluate_external_validation(): Evaluate an independent external-validation dataset
create_external_validation_report(): Create an external-validation report object
write_external_validation_report(): Write an external-validation report
create_gazepoint_model_card(): Create a governance-focused model card
write_gazepoint_model_card(): Write a model card
create_gazepoint_reproducibility_report(): Create a reproducibility report
write_gazepoint_reproducibility_report(): Write a reproducibility report

API stability and contracts

Audit the established public API, S3 classes, and return schemas.

gp3ml_api_contracts(): gp3ml public API contracts
gp3ml_object_schema(): Describe the schema of a gp3ml object
validate_gp3ml_object_contract(): Validate an object against the gp3ml public-object contract
audit_gp3ml_api_stability(): Audit gp3ml API stability
write_gp3ml_api_contracts(): Write gp3ml API contracts
plot(<gp3ml_api_stability_audit>): Plot a gp3ml API stability audit

Cross-package interoperability

Validate and combine prepared handoffs without duplicating upstream preprocessing.

gp3ml_interop_contracts(): Cross-package interoperability contracts
create_gazepoint_handoff(): Create a lightweight cross-package Gazepoint handoff
validate_gazepoint_handoff(): Validate a Gazepoint handoff
combine_gazepoint_handoffs(): Combine validated cross-package handoffs
as_gp3ml_data(): Extract model-ready data from a gp3ml handoff object
plot(<gp3ml_handoff_validation>): Plot handoff validation checks

Reproducibility hardening

Normalize and audit runtime-specific generated output.

normalize_gazepoint_artifact_text(): Normalize volatile text in generated research artifacts
audit_gazepoint_reproducibility(): Audit generated artifacts for volatile output
write_gazepoint_reproducibility_audit(): Write a reproducibility-hardening audit
with_gazepoint_reproducible_output(): Evaluate code with deterministic documentation-output settings
plot(<gp3ml_reproducibility_audit>): Plot a reproducibility-hardening audit

Engine portability

Review optional-engine availability and controlled degradation.

gp3ml_engine_capabilities(): Audit gp3ml model-engine capabilities
assert_gp3ml_engine_available(): Assert that a gp3ml engine is available
plot(<gp3ml_engine_capabilities>): Plot gp3ml engine availability

Integrated research workflow

Simulate and validate realistic cross-package Gazepoint research handoffs.

simulate_gazepoint_research_handoffs(): Simulate realistic cross-package Gazepoint research handoffs
validate_gazepoint_research_bundle(): Validate a synthetic cross-package research bundle
plot(<gp3ml_research_bundle_validation>): Plot research-bundle validation

Decision governance

Predeclare, evaluate, select, apply, and audit classification decisions and abstention.

create_gazepoint_decision_rule(): Create a governed classification decision rule
validate_gazepoint_decision_rule(): Validate a governed decision rule
evaluate_gazepoint_thresholds(): Evaluate explicit classification thresholds
select_gazepoint_threshold(): Select a threshold from a governed threshold evaluation
apply_gazepoint_decision_rule(): Apply a governed classification decision rule
audit_gazepoint_abstention(): Audit abstention decisions
plot(<gp3ml_threshold_evaluation>): Plot threshold evaluation
plot(<gp3ml_abstention_audit>): Plot abstention audit

Target-aware conformal prediction

Prediction-level uncertainty with explicit calibration-unit semantics.

fit_gazepoint_conformal(): Fit target-aware split-conformal calibration
predict_gazepoint_interval(): Predict conformal regression intervals
predict_gazepoint_set(): Predict binary conformal prediction sets
assess_gazepoint_conformal_coverage(): Assess conformal coverage
validate_gazepoint_conformal(): Validate a conformal fit
plot(<gp3ml_conformal_coverage>): Plot conformal coverage

Dataset shift and robustness

Audit predictor distributions, missingness, and analytical stability without one synthetic drift score.

audit_gazepoint_dataset_shift(): Audit predictor distribution shift
audit_gazepoint_missingness_shift(): Audit missingness shift
summarize_gazepoint_shift(): Summarize dataset shift without collapsing it to one drift score
evaluate_gazepoint_seed_stability(): Evaluate seed stability
evaluate_gazepoint_feature_stability(): Evaluate leave-one-feature-out stability
evaluate_gazepoint_threshold_stability(): Evaluate threshold stability around the optimum
evaluate_gazepoint_missingness_sensitivity(): Evaluate named missingness-sensitivity scenarios
audit_gazepoint_model_robustness(): Audit multiple robustness dimensions
plot(<gp3ml_dataset_shift_audit>): Plot a dataset shift audit
plot(<gp3ml_model_robustness_audit>): Plot a robustness audit

Analysis-plan governance

Freeze pre-model analytical commitments and audit later deviations.

declare_gazepoint_analysis_plan(): Declare a frozen-analysis-plan contract
validate_gazepoint_analysis_plan(): Validate an analysis plan
lock_gazepoint_analysis_plan(): Lock an analysis plan using SHA-256
audit_gazepoint_plan_deviations(): Audit deviations from a locked analysis plan
write_gazepoint_analysis_plan(): Write an analysis plan
plot(<gp3ml_plan_deviation_audit>): Plot analysis-plan deviations

Portable models and research artifacts

Serialize governed models, capture environments, package research evidence, and validate cryptographic provenance.

create_gazepoint_model_artifact(): Create a portable governed model artifact
restore_gazepoint_model_artifact(): Restore a model artifact
validate_gazepoint_model_artifact(): Validate a model artifact
test_gazepoint_model_portability(): Test model-artifact serialization and optional fresh-process prediction
capture_gazepoint_environment(): Capture a reproducibility environment record
compare_gazepoint_environments(): Compare two environment records
validate_gazepoint_environment(): Validate the current environment against a reference record
write_gazepoint_ro_crate(): Write a minimal RO-Crate-oriented research object
validate_gazepoint_ro_crate(): Validate a gp3ml RO-Crate-oriented export
write_gazepoint_release_checksums(): Write SHA-256 checksums for release artifacts
validate_gazepoint_release_checksums(): Validate SHA-256 release checksums
plot(<gp3ml_model_artifact_validation>): Plot model-artifact validation
plot(<gp3ml_environment_comparison>): Plot an environment comparison
plot(<gp3ml_ro_crate_validation>): Plot RO-Crate validation
plot(<gp3ml_release_checksum_validation>): Plot release checksum validation

Governance evidence profiles

Organize gp3ml evidence and produce explicitly non-certifying standards-oriented crosswalks.

create_gp3ml_governance_profile(): Create a governance-evidence profile
audit_gp3ml_governance_profile(): Audit a governance-evidence profile
write_gp3ml_governance_profile(): Write a governance profile audit
plot(<gp3ml_governance_profile_audit>): Plot a governance profile audit