Mateusz - if you want to change this it is straightforward - simply create a new XMLSchema spec or add optional components to the existing wsparamset storage spec. Alternatively you can simply reuse Slivka's existing JSON serialse/deserialise ?
you can decide YAGNI for external representations if you like. There will still be client specific glue code for mapping between jalview parameter objects and the client's parameter representation. Its a normative refactoring - which may or may not yield simpler code with fewer declarations.
@foreveremain:matrix.org I'm trying to get my API branch up to date with develop. I think I've got the three overlapping classes merged okay (i.e. the ones with conflicts) and have those safely saved out of the repo, but I'm not getting a compiling Jalview! I remember you saying that when trying to merge the latest release to master you had to incant in a particular way -- was it merge release into master using a ... "ours" strategy? I've tried both ways but only with default strategy.
4 replies
@benzoid: is there any reason for keeping both FEATURE-3851 branches ? you seemed to have pushed your merge branch with develop to remote
1 reply
Accessing https://www.compbio.dundee.ac.uk/slivka/ throws 502 Proxy Error. I checked www-slivka on www-2 and all processes including apache daemon were dead. I restarted apache and slivka, but there is still something wrong in the upper layers.
7 replies
@benzoid: feature/JAL-3982_mouseover_highlighting_viaJAL-3860 will give you structure highlighting. passed to you to test
1 reply
how is the start of the week looking @benzoid @warownia1 ?
11 replies
@warownia1 @benzoid remember - usual 2pm catchup today ! (and ben - meeting 30 mins before to talk genomes/jalview!)
3 replies
@benzoid: patch wont quite happen tonight (no surprises I guess!) - most of the logic is extracted from VCFReader now but need a target pattern for the final step. Possibly something like StructureSelectionListener.mouseOverGenomic(species,assembly,chromosome,long..)
👏 @benzoid ! you survived the ensembl rest api going down in the middle of your demo. What are the chances ?
1 reply
@benzoid: I've pushed a prototype for highlighting chromosomal location, but it needs a bit more work, and lots of testing. For human/Ensembl you need to use homo_sapiens in the species field, and I've verified it shows positions on an ENSG sequence. What doesn't work right now is highlighting of all regions mapped to the location (ie positions in the transcripts, CDS and peptide products) - the code is there (more or less) but should really be moved into StructureSelectionManager.
19 replies
@benzoid: it sounds like you've considerably broadened the scope of the API project.. that's good but remember that you also need to broaden your research about the use cases in that case, and it would be nice to at least stick to the roadmap in terms of first targetting what's needed for jalview+genome browsers, and subsequently what's needed for Jalview+Jupyter (that was the other REST API application)
6 replies
In SeqAnnotationServiceCalcWorker#updateResultAnnotaiton line ~704 what was this block supposed to do?
SequenceI aseq = null;
/**
* transfer sequence refs and adjust gapmap
*/
if (ala.sequenceRef != null)
{
SequenceI seq = running.getSeqNames()
.get(ala.sequenceRef.getName());
aseq = seq;
while (seq.getDatasetSequence() != null)
{
seq = seq.getDatasetSequence();
}
}It looks like it was meant to get the original sequence dataset, but the variable seq, which the dataset is assigned to, is gone after leaving the if block. I'm not sure whether aseq should be a dataset and it's a bug, or an original sequence and the loop is redundant.
3 replies
great talk @warownia1 ! I'm sure it'll be referred to by a few people in the group :D
@benzoid: I saw an update via email to https://issues.jalview.org/browse/JAL-3288 about you having found a workaround - but it doesn't seem to actually be on the issue tracker.. is there something wrong with the server ?
16 replies
@benzoid: I'm considering pushing a patch release for tomorrow's training that addresses https://issues.jalview.org/browse/JAL-3365 ? there's a branch with a fix that I think works, but there's another with a failing test highlighting an issue with exporting and re-importing sequence associated secondary structure annotation that I couldn't fathom. Could you have a look at the failing test ?
I'd like to have a very short meeting today @warownia1 @benzoid 2pm-2.30pm only.
1 reply
@benzoid: I'm not going to be able to answer the question on the forum until after the training course is done this evening, but please do chat with @gjbarton about a good answer :)
1 reply
hi @warownia1 we're meeting at 3pm UK - 4pm Europe time today. Hope that works.
afternoon(nearly) @warownia1 and @benzoid - hope all is well. as promised I'm working on the TODO for the website today.
2 replies
Morning/afternoon @benzoid & @warownia1 .. meeting as usual at 2pm UK/3pm PL today.
2 replies
@warownia1: meeting today at 4pm UK time as usual.
3 replies
greetings from a misty Tay. I'm focusing on catching up on writing and (of course) the website this week. let me know if I've forgotten to do something I said I would :)
Hi, what are the best practises of organising tests into classes. I have a
Foo class that I want to test. The consensus is to put all Foo-related tests into a public TestFoo class. Now, let's consider a following scenario, I want to perform two distinct sets of tests on two Foo instances one initialised with dataset A and the other with B. I can't have them in a single class with a single @BeforeClass any more. What's the right way to group tests in testng so they have different setups?
1 reply
@benzoid: just saw https://issues.jalview.org/browse/JAL-4007 - FWIW you might want to edit out the bit about "Theoretically this may lead to a problem with the actual build, though since nothing's actually changed this is unlikely, but it most likely leads to tasks being re-run that don't need to be leading to a less efficient build process." - since the whole point of that 'optimisations disabled' message is that gradle has disabled incremental caching so the build runs correctly.
7 replies
@benzoid: slight feature creep for the website - I've noticed the menu minibar that appears when you scroll down the page in widescreen doesn't get shown when you are in mobile mode. Sidebar based navigation is also lost - I guess we can ensure the text on each page duplicates the links to the sidebar (it mostly does, but I would love to thin it down), but it would be cool to still have the minibar for navigation, since the header looks a bit naff right now
6 replies
@benzoid: is the build server stuck ? Tests are taking more than 10 mins to run, and appear to be failing.
3 replies
but.. it looks like
Gradle suite > Gradle test > jalview.gui.SeqPanelTest.testFindMousePosition_wrapped_scales_longSequence FAILED
10-May-2022 18:34:48 java.lang.AssertionError at SeqPanelTest.java:1037 is a reliable issue.
1 reply
sorry - no mice advice - but we are now on jitsi for the team meeting
1 reply