Difference between revisions of "2017-12-04 - Fiji + KNIP hackathon"

(H5, not HDF5)
(add more notes to H5, not HDF5)
Line 18: Line 18:
  
 
Perhaps [https://github.com/saalfeldlab/n5/issues/15 make it versioned]?  Because ''a duck is not always a duck...''
 
Perhaps [https://github.com/saalfeldlab/n5/issues/15 make it versioned]?  Because ''a duck is not always a duck...''
 +
 +
* Philip Hanslovsky presented it
 +
* feels like HDF5, but stores chunks(blocks) in separate files in the file system.
 +
* is a Java library, but Constantin Pape already wrote a C++/python version of it: z5 (also matches "zarr" library), https://github.com/constantinpape/z5
 +
* there is a special type for label blocks
 +
* blocks can have a halo
 +
* the block grid does not need to be filled dense, some blocks could be missing
 +
* attributes are stored in an additional JSON file
 +
* Discussion: should we define standard now as to how data should be stored in there to prevent an emergence of a zoo of different flavors as there is for HDF5?
 +
    * how to do time series where each timestep / angle could have different image size
 +
    * if we want a general "N5 viewer" for images, we'd have to add calibration data
 +
    * put this information around the N5 dataset, because it behaves more like a dataset within an H5 file.
 +
    * at least add a version number!
 +
* why another file format?
 +
    * parallel writes (awesome for clusters with shared filesystem)
 +
    * provides
 +
    * couldn't this just be another flavor of HDF5?
  
 
== Hackathon Progress ==
 
== Hackathon Progress ==

Revision as of 03:41, 7 December 2017

From Monday, December 4, 2017 through Friday, December 15, 2017, the Max Planck Institute of Molecular Cell Biology and Genetics hosts ~50 developers at the Center for Systems Biology in Dresden, Germany for a hackathon to develop ImageJ2 and Fiji core infrastructure and plugins.


https://gitter.im/fiji/hackathon_dd_2017


Voluntary hackathon calendar

https://tinyurl.com/ybjcq9qw

Hackathon google doc

https://docs.google.com/document/d/1h4uCt4PAEdeGQQwwVZC73o_Anq6-AI2KZiSzxY8kWng/edit

Hot Topics (Technical Discussions)

H5, not HDF5

For more info on H5, check out the github repository here.

Perhaps make it versioned? Because a duck is not always a duck...

  • Philip Hanslovsky presented it
  • feels like HDF5, but stores chunks(blocks) in separate files in the file system.
  • is a Java library, but Constantin Pape already wrote a C++/python version of it: z5 (also matches "zarr" library), https://github.com/constantinpape/z5
  • there is a special type for label blocks
  • blocks can have a halo
  • the block grid does not need to be filled dense, some blocks could be missing
  • attributes are stored in an additional JSON file
  • Discussion: should we define standard now as to how data should be stored in there to prevent an emergence of a zoo of different flavors as there is for HDF5?
   * how to do time series where each timestep / angle could have different image size
   * if we want a general "N5 viewer" for images, we'd have to add calibration data
   * put this information around the N5 dataset, because it behaves more like a dataset within an H5 file.
   * at least add a version number!
  • why another file format?
   * parallel writes (awesome for clusters with shared filesystem)
   * provides 
   * couldn't this just be another flavor of HDF5?

Hackathon Progress