Recent Releases of bibsnet

bibsnet - v3.5.1

What's Changed

Docker build fixes by @madisoth in https://github.com/DCAN-Labs/BIBSnet/pull/160 - this ensures that the correct SimpleITK version is installed to work with version of nnUNet used and also references tar.gz model file on S3 that is named with only major and minor version numbers (3.5) so that we don't need to create new tar.gz files with every path even though models are not updated

Full Changelog: https://github.com/DCAN-Labs/BIBSnet/compare/3.5.0...3.5.1

- Python
Published by LuciMoore 8 months ago

bibsnet - 3.5.0

Main Update

Bug Fix (Missing inferior lateral ventricle in volumes tsv) & s3 tar.gz reorganization by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/153

What's Changed

Update mkdocs.yml by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/143
Hbcd standards by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/146
added explanation of significance of BIBSNet and dependencies by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/147
add circleCI and docker badges by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/148

Full Changelog: https://github.com/DCAN-Labs/BIBSnet/compare/3.4.2...3.5.0

- Python
Published by LuciMoore about 1 year ago

bibsnet - 3.4.2

What's Changed

MOST IMPORTANT FIX: extend read/execute permissions to all files used by BIBSnet by @scott-huberty in https://github.com/DCAN-Labs/BIBSnet/pull/134
Updated documentation by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/125 https://github.com/DCAN-Labs/BIBSnet/pull/126 https://github.com/DCAN-Labs/BIBSnet/pull/128 https://github.com/DCAN-Labs/BIBSnet/pull/129

New Contributors

@scott-huberty made their first contribution in https://github.com/DCAN-Labs/BIBSnet/pull/134

Full Changelog: https://github.com/DCAN-Labs/BIBSnet/compare/3.4.1...3.4.2

- Python
Published by LuciMoore over 1 year ago

bibsnet - 3.4.1

What's Changed

Image overcropping issue in preBIBSNet resolved through implementation of SynthStrip by @LuciMoore to use a brain mask to determine the optimal axial cutting plane and Dockerfile updated by @madisoth to install SynthStrip in https://github.com/DCAN-Labs/BIBSnet/pull/124
Removal of all logic, flags, dependencies etc required for the former cropping method using robustfov. Additionally, because cropping is no longer age-dependent, a participants.tsv file is no longer required https://github.com/DCAN-Labs/BIBSnet/pull/124
Created attribute which writes out volumes.json by @tjhendrickson and fix for brain mask file name issue by @lundq163 in https://github.com/DCAN-Labs/BIBSnet/pull/121

Full Changelog: https://github.com/DCAN-Labs/BIBSnet/compare/3.4.0...3.4.1

Summary of major updates from prior pre-releases (see full details in history below) include:

Replaced previous models with new production models that now don't have chirality issues (uses the no Mirroring class and sets flipping=False when generating segmentations): 540 (T1+T2), 541 (T1-only), 542 (T2-only) by @lundq163
Remove chirality correction steps from postbibsnet by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/114
Inclusion of volumes.tsv in final outputs containing total size of each region by @tjhendrickson

- Python
Published by lundq163 over 1 year ago

bibsnet - 3.4.0

What's Changed

Update stages.md by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/118
Remove chirality correction from postbibsnet by @LuciMoore in https://github.com/DCAN-Labs/BIBSnet/pull/114
Includes new production models that dont have chirality issues: 540 (T1+T2), 541 (T1-only), 542 (T2-only)
Updated Dockerfile commands for building container by @lundq163
Inclusion of volumes.tsv in final outputs containing total size of each region by @tjhendrickson

Full Changelog: https://github.com/DCAN-Labs/BIBSnet/compare/3.3.2...3.4.0

- Python
Published by lundq163 over 1 year ago

bibsnet - Updated paths in tar to be used for testing

- Python
Published by lundq163 over 1 year ago

bibsnet -

Fixing model folder structure from previous v3.3.0 release

- Python
Published by lundq163 over 1 year ago

bibsnet - Updated model to reduce chirality issues

Now includes model 540 which uses the no Mirroring class and sets flipping=False when generating synthetic images. 540 is now the production model

- Python
Published by lundq163 over 1 year ago

bibsnet - BIBSnet v3.2.0

New Model

Model 526 is now included in BIBSnet. This model is now the default for subjects with both T1 and T2 scans. Model 552 is still available with the -model flag

Manual Cropping Reduction

New flag --reduce-cropping This is useful for cases where the default age-specific brain z size specified by BIBSNet results in overcropping. See usage for more information.

Data Archive Changes

Data archive is now stored at https://s3.msi.umn.edu/bibsnet-data/ Data is versioned to improve backwards compatibility. Changes to permissions and exclusions of some files to help with building containers on docker hub

- Python
Published by BarryTik almost 2 years ago

bibsnet - BIBSnet v3.1.2

Modified chirality correction so that it runs in an iterative fashion
- First crude correction splitting left and right hemispheres based on coronal plane
- Second re-performs chirality correction with more refined LR mask
Results reveal that it corrects superficial brain regions outside of the midline

- Python
Published by tjhendrickson about 2 years ago

bibsnet - v3.1.1

BIBSnet v3.1.1

Updated templates for all ages
- Updated in https://s3.msi.umn.edu/cabinet-data-targz/data.tar.gz and the dockerhub container
postBIBSnet removes unconnected clusters from brain mask
- added function remove_extra_culsters_from_mask
Improved logging
- added debug level available with the -d flag
Bugfix for subjects with multiple sessions
- BIBSnet no longer fails on subjects with multiple sessions when no work_dir is specified

- Python
Published by tikal004 over 2 years ago

bibsnet - v3.0.2

This release is primarily to address an issue with chirality correction using templates in the data directory. As this data directory is stored in an S3 bucket on msi.umn.edu the major update does not appear on GitHub. The following changes have been made to the data directory:

replaced 1mo template with different subject and updated readme with new subject number
deleted brainmasks folder and skull-stripped template files as these are not used in BIBSNet

In addition there are several minor documentation and logging improvements.

- Python
Published by tikal004 over 2 years ago

bibsnet - BIBSnet v3.0.1

Bugfix: -participant argument is now optional.

In 3.0.0 this was mistakenly made to be required and is now back to being optional as it was previously.

- Python
Published by tikal004 over 2 years ago

bibsnet - BIBSnet v3.0.0

CABINET is now BIBSnet

The codebase formerly known as CABINET has been primarily been used as a way to run BIBSnet with preBIBSnet and postBIBSnet. This release makes BIBSnet the primary focus, moving BIBSnet into this repository and removing the code that was meant to be used by NiBabies and XCPD. We're excited about our new version of CABINET that goes about running multiple containers in a different way.

What's new

Previously BIBSnet was only the stage that ran nnUNet. We moved that code into this codebase, instead of pulling it from a different repository. We removed all of the code that was used for running NiBabies and XCPD. This includes the parameter JSON file. You can run BIBSnet the same way you ran CABINET, but without a parameter JSON argument. The parameters formerly in the parameter JSON have their default set to the parameters in the container's version of the parameter JSON. See https://bibsnet.readthedocs.io/ for updated usage.

Where we're keeping BIBSnet related things

The renaming of this code means we've had to update some URLs. We're now hosting containers on dockerhub at dcanumn/bibsnet. dcanumn/cabinet has some old images still, but will be sunset in the future. Our readthedocs is now at https://bibsnet.readthedocs.io/. The old readthedocs will link you there. github.com/DCAN-Labs/CABINET is now the home of a new codebase that was originally forked from this repository. It has diverged significantly, which is why we've made this change. We hope you give it a try when running multiple containers back to back.

- Python
Published by tikal004 over 2 years ago

bibsnet - Sidecar JSONs

Derivatives now include sidecar JSONs
Bugfix for subjects with age = 0 in their accompanying .tsv file

- Python
Published by tikal004 over 2 years ago

bibsnet - Segmentations and Masks Output in T1-native and T2-native Space

- Python
Published by hough129 almost 3 years ago

bibsnet - Zenodo Bug Fix

Adjusting .zenodo.json license in order to publish

- Python
Published by hough129 almost 3 years ago

bibsnet - Improved Segmentations with Model 551

Addition of model 551 for T1 and T2 combined segmentation generation
grabs age and brainzsize from participants.tsv when there isn't a sessions.tsv or the information isn't provided in the sessions.tsv (note: this should only be the case when there are one or no sessions per subject)

- Python
Published by hough129 almost 3 years ago

bibsnet - sub-{}_session.tsv Implementation

brainzsize and age will now be grabbed from the sub-{}_sessions.tsv inside each subject's BIDS input directory at the level of the sessions directories instead of the participants.tsv for BIDS valid input directories
readthedocs documentation now reflects this

- Python
Published by hough129 about 3 years ago

bibsnet - Working Directory Implementation

prebibsnet through postbibsnet changed to working directories
the "precomputed" derivatives directory changed to "bibsnet" derivatives directory
working directories removed at the end of postbibsnet if the user does not specify a --work-dir

- Python
Published by hough129 about 3 years ago

bibsnet - Updated Zenodo Files and Dockerfile

Removed python FSL install to allow the docker image to build on dockerhub
Removed CITATIONS.cff
Edited .zenodo.json

- Python
Published by hough129 about 3 years ago

bibsnet - Improved Segmentations with Model 550

bug fix in averaging multiple anatomical images of the same modality (T1w or T2w) in prebibsnet
addition of model 550 (new T1w and T2w bibsnet model) to model.csv, allowing the application to run model 550 in the bibsnet stage
removal of model 512 to save space in the container, replaced with model 550
changed model 512 to model 550 in the readme

- Python
Published by hough129 about 3 years ago

bibsnet - CABINET includes T1-only and T2-only models

T1-only and T2-only models are now included
Codebase automatically chooses model based on inputs
User can specify model if desired
ReadMe documentation for additional workflows and model description
Bug fixes within non-ACPC alignment workflow

- Python
Published by hough129 over 3 years ago

bibsnet - BIDS-ified CABINET Application

CABINET can now run multiple subjects/sessions or subjects with no sessions
Added issue templates
BIDSified the application by moving args from param file to argparse
Updated README.md to reflect these changes
Implemented age-to-average-head-radius inference from table

- Python
Published by hough129 over 3 years ago

bibsnet - MVP

The container can run preBIBSnet - postBIBSnet, it produces a native T1 aligned image, there is a dataset_description.json in the precomputed output directory, and the exit code is 0 for successful runs.

Full Changelog: https://github.com/DCAN-Labs/CABINET/compare/1.3.0...1.4.0

- Python
Published by hough129 over 3 years ago

bibsnet - MVP with Dockerfile

MVP with the MVP Dockerfile - image needs tested again, but pre-postBIBSnet run on this version.

- Python
Published by hough129 over 3 years ago

bibsnet - MVP without dockerfile

Pre-PostBIBSnet work - BIBSnet needs to be run stand alone, not with run.py.

- Python
Published by hough129 over 3 years ago

bibsnet - Deprecated: preBIBSnet and postBIBSnet work

preBIBSnet and postBIBSnet have an MVP.

- Python
Published by hough129 almost 4 years ago

bibsnet - Deprecated: PreBIBSnet is working

Both the non-ACPC and ACPC tracks in preBIBSnet are working for BCP data at this time.

- Python
Published by hough129 almost 4 years ago