0.0.12

This release continues ongoing work on improving memory usage. We have eliminated the memory_limit option in our Estimators in favor of using sparse arrays. We expect to see a corresponding increase in fit times, especially for Monte Carlo FWE correction- however, we plan to address this in future releases.

What's Changed

🛠 Breaking Changes

• Replace multiprocessing with joblib for parallelization and change n_cores default to 1 by @tsalo in #597
• Incorporate joblib into ALESubtraction and fix SCALE docstring by @tsalo in #641
• Stop storing MetaResults as attributes of fitted Estimators by @tsalo in #657
• Refactor Correctors and remove statsmodels requirement by @tsalo in #679

🎉 Exciting New Features

• Add FocusCounter diagnostic tool by @tsalo in #649
• Support cluster-level Monte Carlo FWE correction in the MKDAChi2 Estimator by @tsalo in #650
• Support vfwe_only in CBMAEstimator even when null_method isn't montecarlo by @tsalo in #678
• Add warning when coordinates dataset contains both positive and negative z_stats by @JulioAPeraza in #699
• Add parameter estimate standard error to IBMA results by @tsalo in #691
• Use sparse array in ALE, ALESubtraction, SCALE, KDA, and MKDADensity by @JulioAPeraza in #725

🐛 Bug Fixes

• Retain updated Estimator in Corrector-generated MetaResults by @tsalo in #633
• Do not inherit IBMAEstimator's aggressive_mask from previous Datasets by @tsalo in #652
• Use beta maps in PermutedOLS instead of z maps by @tsalo in #715

Other Changes

• Reduce SCALE memory usage by @tsalo in #632
• Improve memory management in MKDAChi2 Estimator by @tsalo in #638
• Remove Peaks2Maps-related tests by @tsalo in #643
• Disable MA map pre-generation in CorrelationDecoder by @tsalo in #637
• Switch testing from CircleCI to GitHub Actions by @tsalo in #642
• Override unusable methods and improve documentation by @tsalo in #645
• Document other meta-analysis tools outside our ecosystem by @tsalo in #654
• Reorganize and streamline examples by @tsalo in #656
• Convert CBMAEstimator method to function by @tsalo in #658
• Add explicit support for Python 3.10 by @tsalo in #648
• Use BibTeX citations in documentation by @tsalo in #670
• Replace relative imports with absolute ones by @tsalo in #674
• Simplify organization of base classes by @tsalo in #675
• Note why we don't implement TFCE in NiMARE (currently) by @tsalo in #680
• Dropping the memory-mapping option for Estimators and kernel transformers by @JulioAPeraza in #681
• Optimize locating coordinates in convert_neurosynth_to_dataset by @ryanhammonds in #682
• Reduce memory usage of KernelTransformer.transform and meta.utils.compute_kda_ma by @JulioAPeraza in #676
• Generate automatic CHANGELOG from release note and add it to docs by @JulioAPeraza in #684
• Add manual changelog to documentation by @tsalo in #635
• Automatically update CHANGELOG.md for prereleases as well by @JulioAPeraza in #688
• Fix tag-name issue in update-changelog workflow by @JulioAPeraza in #689
• Optimize numpy operations in MKDADensity Estimator and (M)KDAKernel by @adelavega in #685
• Add PAT to automatically commit release notes to CHANGELOG.md by @JulioAPeraza in #695
• Fix CHANGELOG formatting issues by @JulioAPeraza in #701
• Add citation information to documentation by @tsalo in #712
• Add a glossary page to the documentation by @tsalo in #706
• Remove extraneous copy() statements by @jdkent in #662
• Add information about maintaining NiMARE to developer's guide by @tsalo in #703
• Pin minimum version of pandas by @jdkent in #722

New Contributors

• @ryanhammonds made their first contribution in #682
• @adelavega made their first contribution in #685

Full Changelog: 0.0.11...0.0.12