These are chat archives for thunder-project/thunder

20th
Nov 2015
Jeremy Freeman
@freeman-lab
Nov 20 2015 01:36
@kkcthans can you try adding the argument nplanes=1?
without that argument set to 1, a multiplane tiff is interpreted as one single record (i.e. a volume at one time point)
i definitely think you're getting a memory error
and not sure this will fix it, but definitely something to try
you can also try adding export _JAVA_OPTIONS="-Xms512m -Xmx4g” to your ~/.bash_profile
Kyle
@kr-hansen
Nov 20 2015 02:25

@freeman-lab : Thanks! adding nplanes=1 did the trick. It loaded up just fine and how I expected.

One thing I did notice was that after adding nplanes=1, I also tried the arguments "startIdx=0" and "stopIdx=10" to no avail. The whole file was still loaded when I went to inspect an element. Is it not feasible to open a subset of multiplane tiffs? Basically, to open part of the file you have to open the whole file?

Jeremy Freeman
@freeman-lab
Nov 20 2015 02:26
ah interesting, so that’s not currently supported because implicitly startIdx and stopIdxrefer to files, not time per se
obviously, with one file per time point, they are the same, and loading a subset works fine
Kyle
@kr-hansen
Nov 20 2015 02:27
Good to know. It isn't particularly necessary for me at this time. If I need it, I can always convert my multiplane tiff to a series of tiffs.
Jeremy Freeman
@freeman-lab
Nov 20 2015 02:27
ok cool, it could definitely be added if you want to open an issue
probably should’t tackle it until this is done though thunder-project/thunder#237
as it involves the io stuff (which is one of the main things being refactored / hopefully simplified!)
Kyle
@kr-hansen
Nov 20 2015 03:41
Sounds good. I'll look forward to it and keep my eye on that refactoring. I'll open an issue about it.
Jason Wittenbach
@jwittenbach
Nov 20 2015 22:13
@26tarun I’m not sure I completely understand what you mean by ‘batch feature extraction’. I tried to take a look at the notebooks, but the links seems to be broken