release raiwidgets and responsibleai 0.17.0

[imatiach-msft]

Description

release raiwidgets and responsibleai 0.17.0

v0.17.0

• educational materials
  • add e2e test steps to CONTRIBUTING.md file for developers/contributors
  • add steps for validating UX and SDK changes to documentation
  • replace boston housing dataset with california housing dataset in tests and example notebook due to ethical concerns and biases
  • update dnn pytorch notebook example to give more interesting results by increasing model epochs
• new features

  • Responsible AI Dashboard

    • move performance metric chart, disaggregated metric table, and StatisticsUtils into core-ui
    • add highchart, all components will be moved to highchart instead of plotly gradually
    • add metrics to StatisticsUtils - F1 score, selection rate, mean absolute error
    • add cohort and filter definitions in raiwidgets SDK
    • enable highchart exporting functionality

  • Responsible AI Dashboard for Text

    • add new interpret-text library skeleton

  • Causal

    • add treatment bar chart
• bug fixes and tests

  • Responsible AI Dashboard

    • add e2e UI tests for Causal Analysis - classification tests
    • add e2e UI tests for Feature Importance - aggregate importances
    • add e2e UI tests for Feature Importance - individual importances
    • add e2e UI tests for What-If Counterfactuals
    • add e2e UI tests for Data Explorer - diabetes-regression-model-debugging notebook
    • add e2e UI tests for Causal Analysis - diabetes-regression-model-debugging notebook
    • add e2e UI tests for Model Overview - diabetes-regression-model-debugging notebook
    • add e2e UI tests for responsibleaidashboard-housing-classification-model-debugging notebook
    • fix e2e UI What-if sort test
    • add e2e UI tests for Feature Importance - responsibleaidashboard-diabetes-decision-making notebook
    • add e2e UI tests for What-if, Model Overview and Data explorer - responsibleaidashboard-diabetes-decision-making notebook
    • add e2e UI tests for Aggregate Causal Analysis - diabetes-decision-making notebook
    • add multiclass classification dataset to UI test app and set up basic model-assessment tests for Model Overview
    • add notebook test for pytorch example notebook and add pytorch to CI
    • move to new fluentUI dependency
    • correctly annotate python warnings
    • separate out raiwidgets and UI tests into its own workflow
    • add js and raiwidget code coverage
    • add upload step to e2e test for CI RAIWidgets Python Typescript pipeline
    • add upload step to e2e test to CI notebooks pipeline
    • return exception details to UI
    • add docstrings for locale parameter

  • RAIInsights

    • catch dirty data conditions in RAI Insights validation
    • update responsibleai to interpret-community 0.24.0
    • update responsibleai to interpret-community 0.24.1
    • update responsibleai to interpret-community 0.24.2
    • make responsibleai managers public and add back compat connectors
    • make explanations save behavior similar to other managers
    • change logic to walk categorical columns in validate_train_test_categories
    • cover missing test scenario in state_directory_management.py
    • default categorical feature to array
    • add error analysis test for regression task in RAIInsights
    • make causal manager add and compute behavior similar to other managers

  • Counterfactual

    • add feature importance description for counterfactual local importance
    • update Counterfactual explanation text description

  • Error Analysis

    • fix importances taking too long to calculate for large data in error analysis
    • add a perf test for validating training the surrogate error tree on large data and a minor optimization for traversing the tree
    • add filtering on index to error analysis backend for RAI dashboard cohort switch
    • remove unnecessary multiple calls to tree api
    • refactor tree view state to be private static in tree view renderer and not in parent
    • create github action to release erroranalysis
    • add excludes cohort filter test in erroranalysis
    • refactor matrix filter and area state to be private static
    • make search case insensitive in feature list
    • add more docs for error analysis surrogate tree

  • Interpret

    • allow ExplanationDashboard to show limited view for more than 1k features instead of erroring

  • Causal

    • improve documentation for nuisance_model and heterogeneity_model
    • implement list method in causal manager

• other

  • fix pypi releases by reverting back to pypi token per package
  • fix dummy counterfactual feature importance values for UI test app
  • add pseudo language for localization test
  • integrate codecov into responsible-ai-toolbox repository
  • remove test warning due to use of deprecated np.object
  • add python 3.9 tests
  • fix ipython dependency in rai-core-flask and remove from raiwidgets
  • update setup.py files to carry python 3.9 classifiers
  • fix broken builds due to pip upgrade which broke pip-tools
  • remove mac temporarily from rai widgets step to fix failing builds
  • remove node-version from CI-python,yml
  • fix flaky flask test
  • improve description for responsibleai and raiwidgets pypi packages

Areas changed

npm packages changed:

• responsibleai/causality
• responsibleai/core-ui
• responsibleai/counterfactuals
• responsibleai/dataset-explorer
• responsibleai/fairness
• responsibleai/interpret
• responsibleai/localization
• responsibleai/mlchartlib
• responsibleai/model-assessment

Python packages changed:

• raiwidgets
• responsibleai
• erroranalysis
• rai_core_flask

Tests

• No new tests required.
• New tests for the added feature are part of this PR.
• I validated the changes manually.

Screenshots (if appropriate):

Documentation:

• My change requires a change to the documentation.
• I have updated the documentation accordingly.

[codecov-commenter]

Codecov Report

Merging #1211 (6476436) into main (c3b7a7e) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #1211   +/-   ##
=======================================
  Coverage   67.09%   67.09%           
=======================================
  Files          91       91           
  Lines        4379     4379           
=======================================
  Hits         2938     2938           
  Misses       1441     1441           

Flag	Coverage Δ
unittests	67.09% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c3b7a7e...6476436. Read the comment docs.

[github-actions]

https://responsibleai.blob.core.windows.net/pullrequest/microsoft/responsible-ai-toolbox/ilmat/release-raiwidgets-0.17.0/dashboard/index.html

[github-actions]

https://responsibleai.blob.core.windows.net/pullrequest/microsoft/responsible-ai-toolbox/ilmat/release-raiwidgets-0.17.0/dashboard/index.html

[github-actions]

https://responsibleai.blob.core.windows.net/pullrequest/microsoft/responsible-ai-toolbox/ilmat/release-raiwidgets-0.17.0/dashboard/index.html

[github-actions]

https://responsibleai.blob.core.windows.net/pullrequest/microsoft/responsible-ai-toolbox/ilmat/release-raiwidgets-0.17.0/dashboard/index.html

[github-actions]

https://responsibleai.blob.core.windows.net/pullrequest/microsoft/responsible-ai-toolbox/ilmat/release-raiwidgets-0.17.0/dashboard/index.html