Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Nov 10 22:12
    SarenT opened #139
  • Nov 09 15:13

    ctrueden on master

    Bump to next development cycle … (compare)

  • Nov 09 15:13

    ctrueden on TrackMate_-5.0.1

    [maven-release-plugin] prepare … (compare)

  • Nov 08 20:00
    etarena commented #613
  • Nov 08 12:35

    tinevez on master

    Fix important TrackMate bug in … Add an error message in case so… Add the possibility to save a d… and 1 more (compare)

  • Nov 08 09:17
    imagejan review_requested #366
  • Nov 08 09:17
    imagejan ready_for_review #366
  • Nov 08 09:17
    imagejan edited #366
  • Nov 08 09:14
    imagejan synchronize #366
  • Nov 08 09:14

    imagejan on delegate-converters

    Add test for AbstractDelegateCo… (compare)

  • Nov 08 08:02
    imagejan edited #366
  • Nov 08 08:00
    imagejan opened #366
  • Nov 07 16:12

    imagejan on delegate-converters

    Add abstract delegate converter… WIP: add test (compare)

  • Nov 06 19:11
    etarena commented #613
  • Nov 06 19:04
    etarena commented #613
  • Nov 06 18:59

    ctrueden on coloc

    WtKendallTau: allow for stable … Add test of SACA Co-authored-b… Convert SACA algorithm to ImgLi… and 1 more (compare)

  • Nov 06 17:19

    etarena on coloc

    Add SACA placeholder to Colocal… Add Shulei's original SACA code WtKendallTau allow for stable r… and 2 more (compare)

  • Nov 06 16:02
    etarena commented #613
  • Nov 06 16:01
    etarena commented #613
  • Nov 06 15:55
    etarena edited #613
tpietzsch
@tpietzsch
It seems to me that DataSource is more a version of Source instead of a replacement for SourceAndConverter. That is, you still need a Converter to display it.
to work with for multi-resolution image analysis and display
for display, yes
for image analysis probably not, because you will not care about converting to ARGB, and also not about volatile access
for that the non-volatile Source is enough
Philipp Hanslovsky
@hanslovsky
Yes, @tpietzsch is right. DataSource still requires a converter (which is part of a SourceState (not the same as in BDV) in Paintera).
tpietzsch
@tpietzsch
but in general, yes, if you want something for both display AND analysis, then SourceAndConverter is the thing
@hanslovsky same idea in BDV. BDV's SourceState< T > extends SourceAndConverter< T >
Philipp Hanslovsky
@hanslovsky
:plus1:
Christian "Tischi" Tischer
@tischi

@tpietzsch @hanslovsky Thanks! Sounds great! I could look into

(1) make a way to expose adding SourceAndConverter in the vistools BdvFunctions

unless someone else wants to do it?

@hanslovsky @bogovicj @tpietzsch Some other topic: I need to serve a TB sized volume EM data set via http, streaming it into Bdv. Currently I am using BigDataServer with bdv.h5 files. This works well, however our IT-department wants to store the data as S3. Is there something useable in terms of n5/S3 for Bdv?
Philipp Hanslovsky
@hanslovsky
John Bogovic
@bogovicj
:+1:
Christian "Tischi" Tischer
@tischi
@hanslovsky @bogovicj Thanks! Do you also have examples that writes a pyramid to n5-s3 from an RandomAccessibleInterval and reads it back, e.g., into a Source (CachedCellImg backed). Or something like that?
John Bogovic
@bogovicj
I don't, but @igorpisarev might
(he's away this week, though, maybe poke again on this topic next week if you haven't already made something work)
Philipp Hanslovsky
@hanslovsky
Using n5-aws-s3 should not be any different than any other N5Reader or N5Writer. Maybe n5-spark, n5-utils, n5-imglib2, or the paintera conversion helper have some inspiration for you.
John Bogovic
@bogovicj
yea, but the "other" stuff you do will be more involved (all the aws cli stuff) https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-configure.html#cli-quick-configuration
Philipp Hanslovsky
@hanslovsky
:plus1:
tpietzsch
@tpietzsch
bdv will soon have export to N5, just needs to be merged
is it 1 TB or many?
because for many TB, you will probably want to do it on a cluster. And for that @igorpisarev stuff might be perfect
Otherwise, if it is 1TB it might still be possible from Fiji/BDV locally
Christian "Tischi" Tischer
@tischi
Awesome that you have done the work for the Bdv! It is 8 TB if you naively load it into RAM, but since there are many black pixels, lossless compressed it is 2.5 TB. @tpietzsch would your code also support export to and reading from n5-aws-s3 or only to the standard file based n5?
Christian "Tischi" Tischer
@tischi
@bogovicj I guess I wait until Igor is back in case he already has some practical experience with aws-s3. And thanks for the pointers regarding the aws user guide. We would have the data stored locally, so I have to look at this with our IT guys.
NicoKiaru
@NicoKiaru
I have a hard time understanding the link between the SharedQueue beeing passed when using VolatileViews.wrapAsVolatile(final RandomAccessibleInterval< T > rai, final SharedQueue queue, final CacheHints hints ) and the CacheControl that I see a bit everywhere when Volatile type are involved. My main issue is to make a class that makes a VolatileSource from a non-volatile source (see here), but I don't know how to deal properly with the cache.
Christian "Tischi" Tischer
@tischi
Not sure that helps, but for me, when the RandomAccessibleInterval rai is backed by a CachedCellImg it was working fine automagically just using VolatileViews.wrapAsVolatile( rai ) without specifying any of the Queues and CacheHints..
Christian "Tischi" Tischer
@tischi
@hanslovsky @NicoKiaru @bogovicj Does any of you maybe have an idea about this issue? tischi/bdv-utils#2
Christian "Tischi" Tischer
@tischi
@NicoKiaru Thanks for posting in the issue. I think you found the "bug".
NicoKiaru
@NicoKiaru
Hi, http://fly.mpi-cbg.de:8081 do not seem to be alive. Is it temporary ? Is there another bigdataserver that can be used for testing ?
NicoKiaru
@NicoKiaru
I've just pull a PR in bigdataviewer-vistools which allows for some bdv.viewer.Sourceto be wrapped as Volatile. (bigdataviewer/bigdataviewer-vistools#38) Probably not optimal, but I think it exposes a current limitation of BdvFunctions.show(Souce src): it does wrap sources that could be wrapped as volatile. I put an example of such limitation being overcomed in this small video : https://www.youtube.com/watch?v=SdngXeEatqA. Maybe it's the wrong way to solve the limitation (exposing the SourceAndConverter is probably more general), but here it is if you want to have a look.
Curtis Rueden
@ctrueden
@NicoKiaru Bene knows. He says (via private mail) that server is permanently down, but he is working on restoring it somewhere.
Oh wait, I am talking about a different server I think. The 3D Viewer docs.
@turekg Do you have power/influence over the fly.mpi-cbg.de server?
Christian "Tischi" Tischer
@tischi
@NicoKiaru Here is a BigDataServer running, but many images, takes a while: http://cbb-bigdata01.embl.de/0.6.0
NicoKiaru
@NicoKiaru
thanks @tischi and @ctrueden !
Stephan Saalfeld
@axtimwalde
@tomancak @tpietzsch I think fly.mpi-cbg.de is not very alive, can you resurrect it?
(oh, and because there might still be some useful stuff, is it possible to get a copy of my home directory there?)
...probably asking too late...
Igor Pisarev
@igorpisarev
Hi @tischi! I'm back and can help with setting up n5-s3 if you have any questions. As @hanslovsky and @bogovicj suggested, you can use n5-spark to generate multiscale exports using an instance of N5AmazonS3Writer. The best examples would probably be the tests in n5-spark and n5-aws-s3 repositories
Igor Pisarev
@igorpisarev
n5-spark can run either locally or on a cluster, including Amazon's EMR platform (elastic map-reduce) that supports Spark. There is a wiki page in https://github.com/saalfeldlab/n5-spark/wiki that shows how to run a Spark application on AWS
turekg
@turekg
@ctrueden @axtimwalde I never heard of this machine, but I can look into it if you want...
Christian "Tischi" Tischer
@tischi
Hi @igorpisarev, thanks for offering help! Just to get started, do I need a cluster and n5-spark or could I also do it one a single machine for a small test data set?
Stephan Saalfeld
@axtimwalde
@turekg this is the old @tomancak lab application server. It used to run the original CATMAID, some databases maintained by Pavel and Radek, and the personal home-pages of @StephanPreibisch and @axtimwalde
Igor Pisarev
@igorpisarev
@tischi the n5 downsampling code is in n5-spark, but you can run it on your machine. Take a look at n5-spark repo, there are scripts for building the shaded jar and running it on a local machine. When you run locally, Spark emulates the cluster environment by using threads, and it comes as a self-contained maven dependency so you don't need to install anything extra. You can also run it in your IDE by supplying the following JVM variable: -Dspark.master=local[1] that sets the number of threads to use, or -Dspark.master=local[*] for using all cores
Christian "Tischi" Tischer
@tischi
@igorpisarev Check! I will try asap. Thanks again!!
Jan Eglinger
@imagejan
An imglib2-roi question: is it possible to sample (random) points on a sphere by taking some GeomMasks.closedSphere(...).xor(GeomMasks.openSphere(…))? And/or to get a full list of points representing the sphere surface by some Bresenham-like algorithm?
Or is using imglib2-roi for this a bad idea, and I should rather take a math function of a sphere and sample points “manually” ?
(I was thinking in the direction of ImgLib2 Example 4a, but taking a sphere surface instead of the full volume…)
turekg
@turekg
@axtimwalde must be tomancak-srv1. It is not available outside MPI-CBG. I can see CATMAID in /srv/www .
I am due to retire this machine, after moving some stuff over to bds.mpi-cbg.de and archiving the rest. Tell me where I can upload your stuff and I can get a tarball of your dir to you