Skip to content

Latest commit

 

History

History
179 lines (114 loc) · 5.15 KB

NEWS.md

File metadata and controls

179 lines (114 loc) · 5.15 KB
Raw

v0.16.0

2019-07-04

  • New Bash script fdt_network_matrix.sh
  • New Slurm script fdt_network_matrix.launcher.slurm

v0.15.0

2019-04-11

  • New script make_unusable.sh moves specified data to the unusable directory

v0.14.0

2019-04-05

  • probtrackx2.launcher.slurm replaces fsl_probtrackx2_run.launcher.slurm
    • Add CLI arguments
  • dti_create_network.sh replaces fsl_probtrackx2_mean.sh
    • Add CLI arguments
  • dti_create_network.launcher.slurm replaces fsl_probtrackx2_mean.launcher.slurm
    • Add CLI arguments

v0.13.0

2019-03-31 -- 2019-04-04

  • dti_reg_FS_to_diff.sh replaces dti_probtrackx2_setup.sh
    • Remove the dependence on Tracula; calculate registrations directly in the script
  • Add dti_qc_reg.sh to create screenshots for QCing registrations
  • Add dependency for Xvfb (required by dti_qc_reg.sh)
  • Add dti_reg.launcher.slurm for the above script

v0.12.0

2019-03-28

  • setup_vars.sh replaces dti_vars.sh

v0.11.0

2019-03-25

  • Change QC directories, so that each QC step is under a single directory
  • New R scripts, dti_qc_eddy.Rscript and realignment_params.R, for more DTI QC measures
    • Called by the new dti_qc_other.sh script
  • New argument --fd-cutoff for dti_eddy.sh

v0.10.0

2019-03-24

  • Add basic QC check of image dimensions

v0.9.0

2019-03-08

  • Initial (very basic) support for multiple modalities

v0.8.0

2019-02-12 -- 2019-02-15

  • Update eddy-related code so it works on Lonestar5 and with a singularity container
  • Update bedpostx-related code to work on Lonestar5

v0.7.2

2019-02-09

  • Fix some launcher-related bugs
    • There cannot be any "heredoc"s in the script being sent to the compute nodes (from a launcher script).

v0.7.1

2019-02-02

  • Add a fix specific to FSL v6.0.0 in which fslroi changed the image type

v0.7.0

2018-12-08

  • Add a few CLI options for eddy_quad
  • Add an error check for the ${target}_dicom.tar.gz file
  • Fix bug in dti_dicom2nifti in which the preproc.json file was incorrectly written to the qc_bet directory

v0.6.0

2018-11-12

New files

  • Moved the color variables to globals.sh; will be sourced by scripts for better usage messages
  • dicom2nifti.launcher.slurm runs the first processing step for multiple subjects
  • eddy_cuda.slurm runs dti_eddy.sh on Lonestar5 for a single subject

Removed files

  • fsl_qa_preproc.launcher.slurm is no longer needed
  • preproc.slurm is superseded by dicom2nifti.slurm and eddy_cuda.slurm
  • The code from dti_qc_eddy.sh is now in dti_eddy.sh

Minor changes

  • dti_dicom2nifti_bet.sh now directly calls dti_qc_bet.sh

v0.5.0

2018-11-06

  • Added functions to log software, system, and parameter information in json files, using jq and jo
  • Use cut to get correct jq version info
  • Remove backticks from usage functions

v0.4.0

2018-11-04

General updates

  • Added a LICENSE file; using Apache version 2.0
  • Added a script to hold "utility" functions, called utilities.sh
  • Put error checking into its own script, check_dependencies.sh
  • Changed the exit codes for non-trivial errors
  • Added ${FSLDIR}/bin in front of some FSL-based programs
  • Changed the code to get scriptdir to work across more systems (e.g., realpath isn't available on CentOS 6)

Other

  • Added an R script, slice_times.R, to calculate the slice timing for interleaved acquisitions
    • Assume sequential acqusition for both Philips and GE

v0.3.0

2018-10-23

  • dti_dicom2nifti_bet now allows the specific input of the DICOM .tar.gz file, via the option --tgz. This means it is not required to have the sourcedata directory tree already set-up
  • dti_vars will create ${srcdir} if it doesn't exist, to facilitate the above change

v0.2.0

2018-10-22

File renames and additions

  • Removed the leading fsl_ from the scripts, as it was redundant
  • Initial preprocessing script no longer runs eddy
    • Created a separate dti_eddy.sh and dti_qc_eddy.sh since eddy is time-consuming
  • Renamed the first QC script to dti_qc_bet.sh

v0.1.0

2018-10-21

Overhaul of the preprocessing script

General updates

  • Remove the --bids option from all scripts; the project directory is now required to conform to the BIDS standard
  • Initial commit of README.md
  • Move atlas text files to atlases directory

Preprocessing updates

  • Now creates output directories if they don't already exist
  • Gets Manufacturer info from the DICOM header
    • Extraction via tar now differs based on this; for Philips data, the directory structure is unchanged
    • For GE and Siemens, any leading directories are removed, and the files are extracted directly into ${srcdir}/tmp
  • Automatically calculate the number of b0 images and average them together for nodif.nii.gz
  • If slspec.txt doesn't exist, try to guess the slice timing information for slice-to-volume correction in eddy_cuda
    • For Philips data, hard-coded for sequential acquisition
  • Change from eddy_openmp to eddy_cuda
    • Add --mporder, --slspec, --residuals, and --cnr_maps as arguments

v0.0.1

2018-10-11

Move all files in bin/fsl up, to bin/.