yeatmanlab/pyAFQ 0.6

Version 0.6

on GitHub

This version includes many fixes, documentation enhancements and
performance optimizations. It also drops Python 3.6 support.
This version matches our first paper describing/using the software.

• Add more timing information (#622)
• Allow CLI to specify what api method is called (#623)
• Plot tweaks to make paper quality (#576)
• Reduce apm test workload (#621)
• Split up nightly 5 (#620)
• Fix test_AFQ_init, may allow nightly 3 to pass (#619)
• Dilate the ROIs. (#618)
• Enh Add Callosum ROIs support (#538)
• BF: Need to read these parameters from file, before getting the data. (#615)
• Drop Python 3.6 support. (#612)
• BF: use get instead of get_nearest (#610)
• [ENH] [DOC] Add matlab to python file conversion functions, add docs for custom tractography integration (#599)
• [FIX] calculate sl counts on the spot (#605)
• DOC: Example that explores BIDS and includes tractography from another pipeline. (#577)
• Allow more flexible definition of endpoint filtering atlas. (#589)
• DOC: Explain that trk files are saved in RASMM. (#604)
• Removes several unused CLIs. (#588)
• BF: Use the provided x and y inputs. (#606)
• [ENH] added reco80 example (#567)
• [DOC]Add mask.rst file to give context and explanation to masks (#598)
• Reuse the CSD sh coefficients if you already have them. (#591)
• [ENH] Allow the user to specify what range the color by volume should shade over (#594)
• Adding dummy end points for custom bundles (#543)
• [ENH] Allow user to optionally normalize density map maximum values to 1 (#595)
• [ENH] Add cloudknot example (#533)
• [ENH] Robust tensor fitting (#580)
• FIX: Traverse BIDS hierarchy to find masks, bvals, and bvecs (#587)
• NF: Adds DKI AWF to scalars. (#592)
• Read and resample ROI data. (#545)
• DOC: Adds some documentation for developers of the software. (#546)
• initialize for subject and session pairs where dwi files exist (#583)
• [FIX] Put tg in rasmm first for SLR registration (#566)
• [FIX] Unflip Plotly x axis; multiple flexibility improvements in plotly plotting for paper (#581)
• WIP DEP: Bump numpy version (#579)
• adding nb_streamlines to segmentation parameters (#570)
• [ENH] add weighted option for dice (#568)
• Adds prealign stage to the examples. (#555)
• Eliminate force_recompute option. (#552)
• Warn when templateflow creates directory (#557)
• Save non anat positioned bundles in their own plots (#539)
• Allow bundle dict as input to afq object (#540)
• Put msmt in nightly 3 (#542)
• Actually use MSMT in API call (#530)
• Update test_init to additional df columns; add nightly 5 test yml (#531)
• NIGHTLY: move reco80 to 4; break up nightly 3 to isolate bug (#525)
• Multi-shell, multi-tissue model (#474)
• BF: Reset the input tractogram space to what you originally got. (#520)
• FIX: Remove xvfb from being needed in tests (#522)
• FIX: Reco80 fixes (#503)
• FIX: specify that the b val range is inclusive (#523)
• ENH #443 Callosal Group Example (#476)
• adding separate example output directories to avoid name collisions (#490)
• FIX: Plotly viz bug fixes, and update to custom bundles (#513)
• FIX: Update all b0_thresholds to 50 (#507)
• FIX: update reg_subject arg checking to include dicts (#515)
• FIX: updates docstring in clean_bundle which returns sft, not nib.streamlines (#514)
• Try plotly as default, re-organize usage documentation (#439)
• FIX: Update tractography max_length docstring to be consistent (#508)
• Split off nightly 4 from nightly 3, nightly rests run python 3.8 (#501)
• ENH: Sort the bundles list from csv alphabetically in group csv comparison (#499)
• ENH Disabling progress bars for sphinx-gallery (#492)
• ENH: Dice coeff (#484)
• enh adding distclean and realclean targets for sphinx build (#489)
• FIX: Downsample number of streamlines vizzed down to 200 when vizzing all bundles (#482)
• ENH: Return contast index dataframe from contrast_index (#483)
• Require dipy versions higher than 0.12, so that we can use current ma… (#488)
• Reg algo automatically chosen based on whether mapping is provided, syn mapping for sls fixed, recobundle defaults updated (#472)
• Apply brain mask to subject img before registration (#478)
• FIX: export registered b0 should use inverse pre align to read mapping file (#479)
• Fix typos in api.afq type checking (#477)
• Type check AFQ object inputs (#456)
• Allow reg subject to be bids filters; refactor getting stanford data into temp folder (#458)
• Removes extraneous underscore. (#475)
• Adds total number of streamlines, to complete the sl counts table. (#469)
• Follow up on #462, to fix failing nightly test. (#470)
• FIX: remove whole brain from bundle list for sl count when using recobundles (#471)
• Fixes the timing table (#467)
• Updating pyAFQ documentation (#455)
• Streamline counts table (#468)
• Adjusting group comparison figures (#466)
• Adds 80-bundle atlas for RB (#431)
• Catch no subject error early (#463)
• Colorful legend when using cbv (#465)
• Give the APM map a more BIDS-ish name, and generate metadata file. (#462)
• FIX: Validate bids_path exists (#459)
• Give users the option to disable individual subject lines in plot (#446)
• Add PFT Mask. (#444)
• Rename profile columns back to standard column names (#445)
• Update plots to paper Quality (#429)
• Remove whole brain from bundle dict for reco viz (#438)
• Allow option to upload combined profile (#437)
• Add packaging requirement (#436)
• Allow user to input custom tractography (#422)
• Automatically choose bundle names for recobundles (#420)
• Skip nightly test that may be causing OOM (#428)