Closed PJpb closed 5 years ago
drivenbyentropy
commented
5 years ago Hi, Thanks for reporting this. Does his affect only data sets created with previous versions of AptaSuite or also those created with the current one? In addition, could you provide me with the OS and version you are running? If you have the chance, would you be able to see if you can replicate this error on a Linux machine?
I will see if I can reproduce it locally in the meanwhile.
Thanks again! Jan
PJpb
commented
5 years ago Hi, OS is Windows 10 Home, version 1803, compilation 17134.345. I do not have an access to a Linux machine at the moment, I'll see what I can do next week.
This situation happend for the data parsed from scratch and clustered with the v0.9.3. When I restarted the AptaSuite, I could load a cluster analyzed previously rather quickly (from LazyCache), and I could also load new clusters (not previously analyzed at any timepoint). However, when I open a previously loaded cluster from the cache, I get the same bug as described (bug_report2.txt ). I did some further testing and observed that:
Hope that helps to locate the bug!
edit: I've reopened the analysis, and I could open both clusters A and B. I was also able to switch from cluster A to B and B to A multiple times. (The Cluster Cardinality Chart was recalculating every time - I've opened a new ticked on that.) When I've opened cluster C - also already previously computed, it went fine, but the bug appeared when I reopened cluster B again (from C).
drivenbyentropy
commented
5 years ago Thanks for the detailed report. It seems that "something" is closing your channels to the mapdb files. Before I go down that route, could you please verify for me what JVM your are running? Specifically, if you are running 32 or 64 bit? I believe its the latter but I'd like to make sure.
The easiest would be to paste the output of java -version here. Thanks!
PJpb
commented
5 years ago Here you go:
java version "1.8.0_151"
Java(TM) SE Runtime Environment (build 1.8.0_151-b12)
Java HotSpot(TM) 64-Bit Server VM (build 25.151-b12, mixed mode)Thanks! I have one more question. Could you please check if this issue only occurs when you are changing clusters while some members of the previous cluster are still being computed or if this also happens if you switch after everything his finished computed?
To test this, please first delete the cache. This can be done manually by going to the project folder corresponding to the experiment on disk, and then navigating to the clusterdata folder. Here, please delete all files but clusters.mapdb and then start AptaSuite.
Now, click on any cluster, let it finish completely, and then switch to a new one. Let it finish again, and switch back. If this does not produce the error above, we are one step closer to solving the issue.
PJpb
commented
5 years ago It seems that this is a good lead - it worked nicely only when I allowed for all the computations to be done, including cluster cardinality, logo and mutations analysis. I did it a few times with a few clusters and the LazyCache worked well (as long as I waited long enough - it does take a bit of time for my laptop to compute the mutations and cardinality, especially for bigger clusters). Then I tried to replicate the error by changing the cluster before the additional analysis in the right panel got computed and it did crash in the same way as described above.
Good hunting! ;)
drivenbyentropy
commented
5 years ago @PJpb Sorry for the lack in response. I have issued a new version `v0.9.4. Please let me know if this fixes the problem.
Hi, The LazyCache doesn't work properly - after getting back to a cluster that has been previously analyzed and stored in the cache, the sequences are not displayed at all. Windows PowerShell log attached. bug_report.txt
Edit: What is more, no new (not analyzed before and stored in the cache) clusters can be displayed - the loading goes fine, but no sequences are displayed. In both the cases described the charts on the right hand side (e.g. Cardinality Chart) do not display as well.