mljar/mljar-supervised

📋 Changes

• publish_app() now reuses the last published URL by default (#830)
• Fixed missing mljar_app.json upload in publish_app() (#829)
• Improved numeric widget steps in generated apps (#828)
• Added AI Act transparency messaging in generated apps, reports, and training output (#810)

📋 Changes

• (#803) automatically generate web app for your automl model
• (#813) add fairness compliance certificate generation

📋 Changes

• Simplified `report_structured()` default output to compact summary (leaderboard + global feature importance).
• Added on-demand detailed report for a selected model via `report_structured(model_name=...)`.
• Improved selected-model markdown layout: clear section order, cleaner headings, filtered noisy hyperparameters, richer fairness details and fairness explanation text.
• Included fairness columns in summary leaderboard when fairness is enabled.
• Fixed global feature-importance aggregation metadata (`models_present`) and updated examples/tests for the new report workflow.

📦 mljar-supervised 1.2.0

• Added `AutoML.report_structured(format="markdown"|"dict"|"json", model_details=...)` for LLM-friendly reporting.
• New on-demand artifact: `<results_path>/report_structured.json` (full structured payload).
• Improved markdown report readability (clean run summary, clearer best/model sections, better feature-importance naming).
• Added examples for classification, regression, and fairness structured reports.
• Updated structured-report tests and fixed macOS smoke workflow by installing `libomp`.

🐛 Fixes

• removed unused dependencies, cleaned code -> the installation should be a little faster :) Changes in the #800

📋 Changes

• fixed issues with new sklearn API #789, #788, #787
• setup matplotlib backend for AutoML and switch it back to original #785

📋 Changes

• a lot of warning and bug fixes done in #777, #771, #762, #761, #760, #759, #758, #757, #756, #755, #754, #753, #752, #751, #750, #749, #743, #742, #733

📋 Changes

• (#380) disable boost on errors step for custom strategy
• (#728) fix accuracy metric for Lightgbm

📋 Changes

• (#725) fix styling of AutoML report, apply styles in the `mljar-automl-report` class

📋 Changes

• fixed problems with `report()` (#714)

📋 Changes

• fix xgboost warning (#667)

🐛 Fixes

• fix sklearn/scipy warnings (#709)
• fix report display in JupyterLab (#710)

📋 Changes

• 🐼 pandas > 2.0.0
• 🚀 xgboost > 2.0.0 (#649)
• 🌳 dtreeviz > 2.2.2 (#631)
• 🌈 shap > 0.42.1

🐛 Fixes 🛠️

• Alrighty, with great power (read: updates) comes great responsibility (read: fixes)! We've rolled up our sleeves to zap those pesky warnings caused by our major package glow-up:
• 🎓 Added classes_ for those classy classifiers (#654)
• 📊 Patched up a boo-boo in the calibration plot (#655)
• 🔧 Tweaked a model type warning that was acting all sassy (#638)
• Keep rocking and happy coding! 🎸🤖🚀

📋 Changes

• #637 fix problem with font loading for report

📋 Changes

• #634 fix problem with categorical values in target and nan values for fairness metric
• #635 add tests for fairness feature
• #636 switch off shap exceptions printouts

📋 Changes

• For implementation details please check issue #612
• For example usage please check article https://mljar.com/blog/fairness-machine-learning/

📋 Changes

• #595 replace boston example dataset with California housing dataset, replace `mse` metric with `squared_error` for tree based algorithms from sklearn
• #596 change the import method for `dtreeviz` package

📋 Changes

• #590 dynamically set font in a report, thanks @yairVanti!

📦 Enhancements

• #523 Add type hints to AutoML class, thank you @DanielR59
• #519 save train&validation index to file in train/test split, thanks @filipsPL @MaciekEO

🐛 Bug fixes

• #496 fix exception in baseline mode, thanks @DanielR59 @moshe-rl
• #522 fixed requirements issue, thanks @DanielR59 @MaciekEO
• #514 remove warning, thanks @MaciekEO
• #511 disable EDA, thanks @MaciekEO

📋 Changes

• #463 change multiprocessing to Parallel with loky
• #462 handle large data for tree visualization in regression
• #419 remove/hide warnings
• #411 loose dependencies for numpy and scipy

📋 Changes

• #81 add scatter plot predicted vs target in regression
• #158 add ROC curve for binary classification
• #336 add visualization for Optuna results
• #352 add support for Colab
• #374 update seaborn
• #378 set golden features number
• #379 switch off boost_on_errors step in Optuna mode
• #380 add custom cross validation strategy
• + 8 more

📋 Changes

• #343 set seed in Optuna
• #344 set eval_metric directly in all algorithms
• #350 add estimated train time in Optuna mode
• #342 add `optuna_verbose` param in `AutoML()`
• #354 add KNN in Optuna
• #356 and Neural Network in Optuna
• #357, #348 use mljar wrapper for Random Forest and Extra Trees
• #358 add `extra_tree` param in LightGBM
• + 6 more

📋 Changes

• #332 We added Optuna framework for hyperparameters tuning. It can be used by setting `mode="Optuna"` in AutoML. You can read more details at blog post: https://mljar.com/blog/automl-optuna/

📋 Changes

• #179 add `need_retrain()` method to detect performance decrease
• #226 extract rules from decision tree
• #310 add support for MAPE
• #312 optimize prediction time
• #313 set stacking time threshold depending on best model train time
• #320 search for model with prediction time constraint
• #322 `n_jobs` as a parameter
• #328 disable stacking for small (nrows < 500) datasets
• + 8 more