Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* access to data structure adapted * use tapas_hgf_binary_plotTraj instead of tapas_h2gf_binary_plotTraj; data structure has been adapted to run with the HGF plotting function * plot boxplot as well for zeta * plot boxplots for RW * plot trajectories as well for RW * jobs for rw added * run hgfv5.2 for srl1 and srl2 * correction * figure title: iterations replaced by samples * plotting part updated * final noise ROIs are written in a volume, and displayed through spm_orthviews noise ROIs erosion is now 3D and using spm_erode * Update Contributor-License-Agreement.md reformatted table for better display * plot mean and ste for h2gf and hgf results * updates * added name and affiliation in Contributor-License-Agreement.md * noiseROIs after processing are now displayed only when verbose>=2 * some more help in the GUI about the expected geometry of the maks/tissue probability * noiseROI GUI : added a Yes/No switch to force Coregister:Estimage&Reslice yes : always do it no : compare geometry and decide to skip coreg or still perform it * Updated demo to increase length of sampling chains * included physio-public/merge-tapas-development branch * removed examples/ from .gitignore * Implemented automatic ploting of posteriors. * Added an example for flat model. * Updated changelog. * Solved a bug with invgamma. * Added a figure with an example of the results. * Small bug in plotting. * PhysIO: bugfix BIDS .gz, file still open when trying to delete temp folder for unzipping * hotfix github issue #38; restricting valid searchPositions for findpeaks_template_correlation for early peaks * PhysIO bugfix github issue #37, first scan event and first cardiac pulse coinciding * PhysIO: bugfix #43, prepending c/r time courses, if recording started after scan, irrespective of specified logfile names, as long as c/r time courses exist * PhysIO: fixes BIDS bugs issue #43, #45 * WIP PhysIO R2018.2.0 update, removed suptitle dependence and added BIDS examples and default template example * PhysIO bugfix suptitle; new findpeaks wrapper * PhysIO bugfix issue #46, skipped tics samples in logfiles * Start implementing summaries for seria. * Almost complete summaries of the parameters. * Implemented the summaries. * Implemented summaries for prosa and nice displaying. * Typos from the readme file removed. * Cleaned up demo * Small typo. * Typo in readme file. * Deal with invalid trials. * Added a youtube video. * Improved handling of the errors. * HGF Toolbox Version 5.2.2 * New Release of rDCM * Check for empty inputs * Added comment on rDCM to the TAPAS CHANGELOG.md * - HUGE: minor bugfix * Improved the documentation. * Added change log. * Corrected small error in function header * Added get revision info. * Demo with simulations on a grid and varying noise * HGF Toolbox Version 5.3.0 * bugfix in HGF softmax_binary_sim according to issue #48; to establish consistency with softmax_binary by also taking reward values from the input into account * Improved the changelog. * Updated CHANGELOG to reflect hgf 5.3. * Improved documentation and removed some warnings. * Improved documentation and new video. * Added table for summary stats to docs. * Added code for the summaries of prosa. * Updated requirements. * Updated the python code. * Added wrappers for the generation of summaries. * Updated the summaries to use c code. * Summaries are done in c. * Updated data and data file. * Small typo. * Improved docs of download_example_data. * Update the name of the examples and scripts. * Improved documentation. * Updated the examples download. * PhysIO R2019a-v7.1.0 release: updated documentation and examples; introduced new semantic versioning; modularization of functions reflected in figure names; see CHANGELOG.md for full details * PhysIO R2019-v7.1.0 release: added units test BIDS reader and findpeaks * Included info on PhysIO-R2019a-v7.1.0. Note that the CHANGELOG infos from TAPAS 3.0.1 and 3.0.0 seem to be mixed up with the unreleased (3.1) here * Updated the CHANGELOG.md * Implemented WBIC, WAIC, parameters of second level and solved bug. * README with external contributors and mentioning serial correlation cleanup by PhysIO, removing misleading warnings for cardiac phase and gradient gain correction * Bugfix tapas_physio_init discovered by Dario; was not working, because utils functions used in init were not on path * Removed the mlx file and updated the h2gf demo. * Added utilities for delta plots. * Small changes. * commits from physio-public tag R2019b-v7.2.0 * Added PhysIO R2019b change summary to CHANGELOG.md * - huge: rewrote HUGE as Matlab class * - references for user manual * - tapas naming scheme * Update the version number and added references. * Removed .DS_Store and updated example hash. * Updated change log. * noise ROI : PCA is now performed with MATLB built-in SVD function, instead of PCA function in "stats" toolbox noiseROI : keep variables name coherent * new function tapas_physio_pca to perform PCA with new SVD method or old STATS toolbox method added also tapas_physio_pca_test to test if both methods give the same result * fix : tapas_physio_pca_test uses now randn() for noramlly distributed numbers instead of integer numbers fix : dimension of the output from tapas_physio_pca * bugfix in the tests for PCA : use randn instead of randi tests for PCA : added verification for nVoxels <= NVolumes * New run of demo (including est struct) * typo * - added info to CHANGELOG.md of TAPAS about PhysIO changes * PhysIO: updated its internal version and changelog * PhysIO: updated README version * Commented and change the implementation of the early unit. * Changes during discussion with Eduardo - alpha and beta Gaussian-gamma hyperparameter definitions adapted - Improved comments * HUGE: bugfix x-tick labels * Update CHANGELOG.md with changes to HUGE * Improved the general structure of SEM. * Further comments and changes while discussing * Updated small bug in preapre priors and improved prior. * Improved the documentation. * Renamed correctly a function. * Improved documentation and adapted demo. * Empirical delta plot analysis. * Solved an apparent bug in the code. * Improved the structure displaying posterior (SEM) * Better posterior summaries (SEM). * Added function for WAIC. * Now compute the WAIC as part of the summary. * Improved posterior summaries (SEM). * Improved the setup script for tapas. * Improved the prosa model. * Updated the setup file. * Debugging version for summary stats. * Experimental. * Trying to solve numerical bug. * Founda a really serious bug with the python parameters. * Small change. * Simplified a bit prosa. * Standarized the implementation of prosa. * Seria model compatible with python3. * Small change. * Python3 compatible. * Found weird bugs. * Version 3.3.0 with new SERIA. * included PhysIO v7.2.2 bugfix Siemens VB sync to DICOM, using PPU3T and added this new example * PhysIO v7.2.3 (bugfix manual peak selection, github issue #85) * allow figures to be showed/saved/closed upon review * tweak figure show/save/close code * tweaks * updates figure display/save/close functionality * make isfile/exists switch dependent on Matlab ver 'isfile' is clearer than 'exist', because checks for correct absolute/relative path, not existence anywhere on Matlab path, but only available in Matlab version >= R2017b * PhysIO v7.2.4, bugfix for docked figures -nodisplay now (Sam Harrison * placeholder for release date * HGF Toolbox v5.3.3 * Ensure parameter covariance in h2gf is positive semi-definite * Streamlined h2gf setup loops, removed unused files * Continued streamlining and removing unused files and functions * Removed all instances of 'initilize', replaced with 'initialize' * In h2gf: removed all 'initilize', replaced with 'initialize' * HGF Toolbox v5.3.5 * HGF Toolbox v5.3.6 * PhysIO Release R2020a-v7.3.0 * HGF Toolbox v6.0 * - preparing current release * - preparing release * release 3.3.0 (#100) * Added hierarchical inference part. * Documented and simplifid output. * Deleted some typos. * Solved a bug in SERIA introduced by new feature. * Added mcmcdiag and diagnostics for h2gf. * Fixed typos * Started editing h2gf/README.md * Now the expected states are stored in the output structure. * Now the correlation matrix is only of the valid parameters. * Now return the pseudo model evidence. * Updated the REDME file.. * Updates to h2gf_example * Plot function provided by Sandra * Further updates to h2gf_example - Vary omega2 instead of kappa2 - Various other improvements * Replaced tapas_h2gf_example.m with h2gf_demo.mlx * Updates to h2gf_demo * Improved h2gf_summary - All plotting and diagnostics function from the HGF Toolbox now work - Plotting function provided by Sandra now obsolete and removed - Adapted h2gf_demo - h2gf_demo now also in source (.m) not just binary (.mlx) * Minor changes. * Added nchain input, improved docs, simplified input. * h2gf for srl2 data from the PRSSI EEG project, different Nr of iterations, config files and eta values can be specified * script to run h2gf_demo_link_figures for all hgf configurations and h2gf estimations * x-axes adjusted * demo srl2 * Issue: conflict between pars.nchains and pars.T - When pars.nchains = 8 and pars.T left on default, this produces an error - Reproducible by running h2gf_demo.mlx * Solved bug with chains. * save as well as png * replaced by h2gf_demo_srl2_link_figures * h2gf estimations for long version of SSL task * save figure as png with higher resolution * jobs submitted to euler * loops modified * scripts to plot individual trajectories * euler jobs added to plot individual trajectories * ze prior mu set to log(48) instead of 0.5) * shuffle position of subject's data on data structure included * paths added * script with paths to data structure for short srl task * plot histogram of samples * Store the rng seed into the output. * Documented the rng_seed option. * RW implementation * run simulations with srl1 inputs * variance disabled * paths adapted * timing adapted * access to data structure adapted * use tapas_hgf_binary_plotTraj instead of tapas_h2gf_binary_plotTraj; data structure has been adapted to run with the HGF plotting function * plot boxplot as well for zeta * plot boxplots for RW * plot trajectories as well for RW * jobs for rw added * run hgfv5.2 for srl1 and srl2 * correction * figure title: iterations replaced by samples * plotting part updated * plot mean and ste for h2gf and hgf results * updates * Updated demo to increase length of sampling chains * Cleaned up demo * Deal with invalid trials. * Improved handling of the errors. * Demo with simulations on a grid and varying noise * Implemented WBIC, WAIC, parameters of second level and solved bug. * Removed the mlx file and updated the h2gf demo. * Added utilities for delta plots. * Small changes. * New run of demo (including est struct) * Commented and change the implementation of the early unit. * Changes during discussion with Eduardo - alpha and beta Gaussian-gamma hyperparameter definitions adapted - Improved comments * Improved the general structure of SEM. * Further comments and changes while discussing * Updated small bug in preapre priors and improved prior. * Improved the documentation. * Renamed correctly a function. * Improved documentation and adapted demo. * Empirical delta plot analysis. * Solved an apparent bug in the code. * Improved the structure displaying posterior (SEM) * Better posterior summaries (SEM). * Added function for WAIC. * Now compute the WAIC as part of the summary. * Improved posterior summaries (SEM). * Improved the prosa model. * Updated the setup file. * Debugging version for summary stats. * Experimental. * Trying to solve numerical bug. * Founda a really serious bug with the python parameters. * Small change. * Simplified a bit prosa. * Standarized the implementation of prosa. * Seria model compatible with python3. * Small change. * Python3 compatible. * Found weird bugs. * Version 3.3.0 with new SERIA. * placeholder for release date * HGF Toolbox v5.3.3 * Ensure parameter covariance in h2gf is positive semi-definite * Streamlined h2gf setup loops, removed unused files * Continued streamlining and removing unused files and functions * Removed all instances of 'initilize', replaced with 'initialize' * In h2gf: removed all 'initilize', replaced with 'initialize' * HGF Toolbox v5.3.5 * HGF Toolbox v5.3.6 * PhysIO Release R2020a-v7.3.0 * HGF Toolbox v6.0 * - preparing current release * - preparing release Co-authored-by: Eduardo Aponte <[email protected]> Co-authored-by: Christoph Mathys <[email protected]> Co-authored-by: siglesias <[email protected]> Co-authored-by: Christoph Mathys <[email protected]> Co-authored-by: Eduardo Aponte <[email protected]> Co-authored-by: Lars Kasper <[email protected]> * Added FDT toolbox to tasks * Added necessary folders * Updates to folder location checks * Added functs to check if new TAPAS version on github * Used try/catch in case API changed/offline * Started functions to find dependencies * Comments and minor improvements for init * Renamed tapas_new_init into tapas_init * Adapted renaming of files to file content... * - misc: skip toolboxes with failed init * - misc: bandwidth optimization, return path to example data * - misc: ignore examples folder * - misc: external code as package * - h2gf: replace dependency on external toolbox with internal function for psrf calculation * - h2gf: move binary data to example files * - h2gf: add dependencies * - huge: release 2020-09 * Update CHANGELOG.md * First release of the tapas-ceode toolbox to integrate DCM for ERPs efficiently and accurately * Update README.md * Update tapas_ceode_fx_erp.m * Update tapas_ceode_fx_cmc.m * Update tapas_ceode_gen_erp.m * Update tapas_ceode_int_euler.m * Update tapas_ceode_compare_integrators_cmc.m * Update tapas_ceode_compare_integrators_erp.m * Update tapas_ceode_synData_cmc.m * Update tapas_ceode_synData_erp.m * Update CHANGELOG.md * ceode: referral to github issues * Initial commit of version v1.2 * Initial commit of version v1.2 * Fixed small bug related to rs-fMRI data * rDCM Release v1.2 * rDCM Release v1.2 * rDCM Release v1.2 * Update CHANGELOG.md * - release 4.0.0 * - init: huge init function * Update CHANGELOG.md * Update CHANGELOG.md Co-authored-by: siglesias <[email protected]> Co-authored-by: Christoph Mathys <[email protected]> Co-authored-by: Benoît Béranger <[email protected]> Co-authored-by: Lars Kasper <[email protected]> Co-authored-by: Christoph Mathys <[email protected]> Co-authored-by: Eduardo Aponte <[email protected]> Co-authored-by: Stefan Frässle <[email protected]> Co-authored-by: Eduardo Aponte <[email protected]> Co-authored-by: nbuergi <[email protected]> Co-authored-by: Eduardo Aponte <[email protected]> Co-authored-by: jsheunis <[email protected]> Co-authored-by: Lars Kasper <[email protected]> Co-authored-by: [ofaull] <[[email protected]]> Co-authored-by: Matthias Müller-Schrader <[email protected]> Co-authored-by: dschoebi <[email protected]>
- Loading branch information
16 people
authored
Sep 9, 2020
1 parent
45beaf2
commit 416a9fe