edfEazyDataFix

API Reference

EazyDataFix 2.0 adds reusable recipe and pipeline objects while preserving the stable v1 transformation, validation and readiness APIs.

Reusable pipeline API (v2.0)

These additive v2 objects turn explicit preparation rules into a reviewable, repeatable pipeline. See the pipeline recipe guide for complete examples and boundaries.

edf.PipelineRecipe(...)→ PipelineRecipe2.0
Define and save versioned JSON/YAML preparation rules.
edf.ColumnRule(...)→ ColumnRule2.0
Declare an explicit type, mapping, missing-value rule and conversion failure policy.
edf.Pipeline(...)→ PipelineResult2.0
Preview, apply or run a recipe with evidence and guarded exports.
edf.RecipeMLArtifact(...)→ DataFrame2.0
Replay recipe-bound, training-fitted preprocessing on prediction inputs.

Stable convenience workflows (v1 compatible)

These APIs form the stable v1 workflow layer. See the v1.0.0 release notes for complete examples.

edf.run(...)→ RunResult1.x
Compose profile, assessment, controlled cleaning and deterministic EDA.
edf.prepare_with_report(...)→ PreparationReport1.x
Prepare a dataset and return deterministic changes and warnings.
edf.infer_schema(...)→ DataContract1.x
Infer expected columns and data types from a trusted dataset.
edf.validate_contract(...)→ ContractValidationReport1.x
Validate schema and reusable quality rules with pass/fail output.

Core data APIs

edf.assess()→ AssessmentReport
Compute a structured data-quality assessment for a dataset.
edf.fix()→ FixResult
Apply configurable, deterministic cleaning to a dataset.
edf.profile()→ DatasetProfile
Inspect the structural metadata of a dataset.

Deterministic Agentic EDA

The Agentic EDA foundation remains fully supported in v1.0. See the historical v0.3.0 release notes for the full workflow.

edf.eda(...)
Generate a structured deterministic EDA result.
edf.plan_eda(...)
Create a reproducible follow-up analysis plan.
edf.execute_eda(...)
Execute selected EDA steps using deterministic handlers.
edf.run_agentic_eda(...)
Run the complete deterministic Agentic EDA workflow.
edf.export_agentic_eda_report(...)
Export HTML, JSON, Markdown and supported PNG visualisations.

Additional stable utilities

edf.assess_ai_readiness(...)
Assess whether a dataset is ready for downstream AI or ML use.
edf.prepare(...)
Return a deterministically prepared DataFrame.
edf.analysis_ready(...)
Clean and prepare a dataset for analytics and machine learning.
edf.analysis_ready_with_report(...)
Return analysis-ready data with scores, changes, warnings and validations.
edf.ml_ready(...)
Create leakage-safe train/test inputs and a reusable preprocessing artifact.
edf.powerbi_ready(...)
Prepare validated single-table or multi-table Power BI model inputs.
edf.generate_agentic_eda_narrative(...)
Generate an optional evidence-cited narrative from deterministic EDA.
edf.export_agentic_eda_notebook(...)
Export a ready-to-run deterministic Jupyter Notebook.

Conventions

Public workflows accept a supported dataset path or pandas.DataFrame and do not mutate caller-owned DataFrames. Stable result objects keep structured data available for notebooks, applications, tests and CI pipelines. Catch EazyDataFixErrorfor package-level failures.