v0.2.3

What's Changed

• Adding Allen2022MassiveRaw by @hubertjb in #105
• Revert and re-attribute neuralbench studies by @hubertjb in #106
• Fix the building of the Bel2026PetitListen study by @BelCorentin in #110
• Fix splitting issue on non-affected words by @shouhamdi in #111
• fix: align study filenames with class names by @kbegany in #119
• Add emg/qwerty CTC keystroke decoding task by @bruAristimunha in #50
• Enable NeuralBench local-only execution through config.json by @hubertjb in #118
• Set exclude-newer to 3 days for CI build by @hubertjb in #120
• Truncate HuggingFaceText context to model positional capacity by @shouhamdi in #121
• Fix Brennan _download method to accept overwrite; Add nworker parameter to OpenNeuro by @teonbrooks in #124
• Add support for new datasets in NeuralBench by @hubertjb in #123
• Fix physionet rename by @edgarduc in #67
• neuralfetch: fix mne sample link by @kingjr in #127
• Fix TimedArray offset issue in FmriExtractor by @jrapin in #129
• Refactor HuggingFace visual extractors by @sdascoli in #102
• Clarify ImportError message for Step not found by @julien-gadonneix in #137
• Fix _check_requirements to work with git+https:// URL requirements by @shouhamdi in #133
• Unify HuggingFace model loading and more by @sdascoli in #135
• [doc] Fix Step applied to a dataframe by @jrapin in #146
• Prefetch HuggingFace extractor models by @sdascoli in #145
• Misc fixes and cleanup by @jrapin in #147
• feat(segments): add on_trigger_overlap guard to list_segments by @kingjr in #115
• Fix duplicate Sentence rows in AddSentenceToWords with multiple Text events by @julien-gadonneix in #122
• Combine fMRI/MEG into hebart2023things by @kbegany in #131
• Move HuggingFace extractor helpers to module by @sdascoli in #151
• Remove duplicate _get_max_length method in HuggingFaceText by @bkowshik in #160
• fix(neuraltrain): declare exca as a runtime dependency by @bkowshik in #156
• chore: hoist core stdlib imports to module level by @bkowshik in #159
• Remove incorrect "32-channel EEG" comments on MEG/fMRI data_shape by @bkowshik in #154
• Fix MNE sample-data re-downloads in Mne2013Sample/Fake2025Meg by @bkowshik in #157
• Adjust duration calculation to account for precision by @julien-gadonneix in #138
• Honor FmriCleaner.ensure_finite when other cleaning options are disabled by @bkowshik in #164
• CONTRIBUTING.md: add neuralbench, complete neuraltrain extras by @bkowshik in #55
• NeuralBench prediction logging by @hubertjb in #126
• Add probe_layer field to DownstreamWrapper for layer-wise linear probing by @bruAristimunha in #83
• Fix Brennan2019Hierarchical loading for the v2 (bn999738r) release by @bkowshik in #175
• Add Glasser fsaverage ROI projection by @yohann-benchetrit in #185
• Align study/file names by @kbegany in #186
• Fix CI by @jrapin in #187
• Support Algonauts CIFTI fMRI derivatives by @yohann-benchetrit in #191
• Use a Step infra for timelines in Study by @jrapin in #194
• docs: fix 02_studies walkthrough for removed infra_timelines API by @hubertjb in #202
• Add TimedArray.copy(**changes) by @jrapin in #206
• Accept mne raw input for spikes extractor when not NWB by @julien-gadonneix in #100
• Miscellaneous fixes by @jrapin in #208
• Onboard Levy2026Noninvasive (SpanishBCBL/DECOMEG) typing dataset by @hubertjb in #201
• Support for new brain model build API (build_from_context) by @hubertjb in #207
• extractors: add CiftiRoiProjector for cortical Glasser + subcortical … by @yohann-benchetrit in #210
• Add fMRI ROI query/selection options and small fixes. by @jrapin in #211
• Revert last commit and Levy to provide correct affections by @jrapin in #212
• Add Levy2026Noninvasive typing study by @jrapin in #214
• Add fMRI ROI query and selection options by @jrapin in #215
• Bump versions to 0.2.3 by @jrapin in #216

New Contributors

• @kbegany made their first contribution in #119
• @bruAristimunha made their first contribution in #50
• @teonbrooks made their first contribution in #124
• @edgarduc made their first contribution in #67
• @bkowshik made their first contribution in #160
• @yohann-benchetrit made their first contribution in #185

Full Changelog: v0.2.2...v0.2.3