Code Sprint

From Neuroimaging Datasharing
Jump to: navigation, search
  • Standard description of both the input and output of the data from the various processing tools. Would require seeing what is in XNAT/XCEDE schema for input and for output. Basis for making sure everyone shares high-level description
  • Having an automated XNAT schema modification and restarting of the server on virtual machine. Can use the Neurodebian VM as a starting point. (Satra)
  • Get started on extracting info from DICOMs (from multiple scanners) in a validated schema. (Satra)
  • XNAT and/or other plug-ins to Scalable Brain Atlas (Rembrandt)
  • Populating XNAT server at INCF with some data (1000FC? Arno's data?) (Christian, Arno)


  • Improve Connectome Viewer Toolkit / PyXNAT bindings (Y & S)
  • Progress bar for uploading/downloading files with PyXNAT (Y & S)
  • Package PyXNAT for Debian/NeuroDebian (Y & S)
  • Use Connectome Viewer as an IDE for NiPyPE (editing, execution, viewing results) (Satra)

Outcome Code Sprint, May 31th - June 1st 2011

Lausanne, Yannick Schwartz & Stephan Gerhard

  • Work on extending XNAT with a customized schema of CFF-like multi-modal datatypes.

We first tried to extend the reconstructionImageData. We run into problems to retrieve the newly created types over the REST API, see this question. We tried then to extend the imageAssessorData as proposed. Unfortunately, updating XNAT the second time did not work for the update schema, likely due to conflicts with our first extension. The schema is here.

  • Started to prepare NeuroDebian packaging of PyXNAT
  • Implemented callback functions in PyXNAT for progress bars when sending/retrieving large data files from XNAT servers
  • Discussion

How HDF5 could be employed to support neuroimaging datasharing. One has to be clear about what type of queries one wants to support which also co-defines the data model and database infrastructures employed in the end. Alternatives databases, e.g. Document Stores, and how they would interact with XML instances, HDF5 and ontologies.

  • Test of creation of Python object model for XNAT and XCEDE using the generateDS.py library

Obviously, generateDS create classes for complex types, and not so for the actual element (names).

Somerville, Satra Ghosh and Christian Haselgrove with assistance from Andy Worth and Greg Millington

  • Continued work on the INCF XNAT instance
  • Installation of XNAT on a VirtualBox VM as a resource for local work and to explore the issues in packaging XNAT for NeuroDebian; see Install_XNAT

Download from: http://neuro.debian.net/_files/contrib/NeuroDebian_6.0.2+XNAT1.5.0-1_amd64.ova

  • Initiated discussion on code distribution and licensing issues
Retrieved from "http://datasharing.incf.org/ni/Code_Sprint"
Personal tools