Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2014-07-02

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:19 jimallman joined #opentreeoflife
00:43 josephwb joined #opentreeoflife
01:09 josephwb joined #opentreeoflife
01:18 jimallman joined #opentreeoflife
01:23 mtholder joined #opentreeoflife
01:32 jimallman joined #opentreeoflife
01:32 towodo joined #opentreeoflife
01:34 scrollback joined #opentreeoflife
01:50 josephwb joined #opentreeoflife
04:31 jimallman joined #opentreeoflife
12:02 josephwb joined #opentreeoflife
12:04 jimallman joined #opentreeoflife
12:52 towodo joined #opentreeoflife
13:05 jimallman joined #opentreeoflife
13:07 josephwb joined #opentreeoflife
13:12 jimallman joined #opentreeoflife
13:38 josephwb are you there towodo?
13:48 jimallman joined #opentreeoflife
14:02 towodo josephwb: sort of
14:03 josephwb i'll bug you later then
14:03 josephwb but perhaps i'll just put the scenario in your mind.
14:03 josephwb or, not
14:07 towodo joined #opentreeoflife
14:27 jimallman joined #opentreeoflife
15:01 rree joined #opentreeoflife
15:12 rree left #opentreeoflife
15:22 josephwb joined #opentreeoflife
15:25 rree joined #opentreeoflife
15:43 jimallman joined #opentreeoflife
16:05 jimallman joined #opentreeoflife
16:08 josephwb towodo: I just sent me query to your gmail account
16:09 towodo I put the unmanaged extension code aside… it can make arbitrary responses, so yes, that's feasible.
16:09 towodo But that's not why or how I was using it.
16:10 towodo and there was some difficulty with maven
16:10 josephwb is that how i could get my thing working: sending a gzippedoutputstream?
16:11 towodo I would think so.
16:11 towodo https://github.com/OpenTreeOfLife/treemachine/compare/unmanaged
16:11 josephwb was this the thing that had problems with the dependencies?
16:12 towodo I think so… weren't you the victim? I had to back it out
16:12 josephwb i was involved
16:12 towodo the solution is to make a plugin .jar that doesn't have the neo4j-server library in it...
16:13 towodo and instead get neo4j-server from the lib/ directory or something
16:13 towodo I haven't tried this yet, I really hate working with maven
16:14 josephwb i gathered that
16:14 towodo this may require creating a second .jar file. that would require a second source tree and another .pom file
16:15 towodo what I don't understand is how the class manager environments work. that is which classes are visible to which other classes.  i have no model of that.
16:15 josephwb maybe cody can help with this
16:15 towodo I asked him a few questions, he doesn't seem to know any more than i do
16:16 towodo he did the bare minimum to get it working
16:16 josephwb understood
16:17 towodo now you can make unmanaged plugins that don't suffer from this dependency problem.
16:18 towodo the reason I ran into trouble was because I particularly wanted to wrap the existing GoLS plugin methods without changing them. this means being able to use Representations.
16:19 towodo well strike that… you will have to extract bytes of Representation objects too
16:20 towodo I think it was a mistake for Cody to make new Representation classes; it means you get objects that you can't get any information out of without linking in neo4j-server.jar. He didn't know this of course
16:20 josephwb yes
16:21 towodo does that make sense?  you can create Representations without neo4j-server, but you can get any output out of them without neo4j-server.  I need the output to put in GET responses, you need it so you can compress it.
16:21 josephwb yes, makes sense
16:22 towodo actually the cleanest solution I think is to get rid of the dependency, or to introduce special methods on the extended Represenation class that get the bytes through a side channel
16:22 towodo hmm...
16:22 towodo hadn't thought of that before
16:22 josephwb i'm afraid that is beyond my capabilities at the moment; very new to plugins
16:23 towodo not that complicated, just a bit messy.  A plugin returns an object of type Representation. That's a neo4j class
16:23 towodo Cody subclasses it.
16:23 josephwb yes, right
16:24 towodo If all the plugin methods returned YummyRepresentation instead of Representation, then a special method on YummyRepresentations could expose the bytes
16:25 towodo in theory
16:27 towodo hmm. looking at it. I think there may be snags
16:29 towodo we would have to create a MappingSerializer… might not be possible, class could be final require something that's not exposed
16:31 towodo treemachine/src/main/java/org/neo4j/server/rest/repr/ArgusonRepresentationConverter.java
16:32 towodo darn, i deleted neo4j sources to save disk space.
16:32 towodo lunch now
16:32 josephwb ok, thanks for thinking about this
16:33 towodo I have a feeling Cody rewrote the code to change ad-hoc json syntax generation to use the neo4j MappingSerializer class, which generates correct json by design
16:34 towodo we might have to convert it back, or reimplement or copy what's in neo4j.
16:34 towodo the next step on this path would be to traipse around the neo4j sources looking to see if we can get our hands on our own MappingSerializer.
16:47 jimallman joined #opentreeoflife
17:29 josephwb1 joined #opentreeoflife
17:32 jimallman joined #opentreeoflife
17:51 josephwb joined #opentreeoflife
18:05 rree left #opentreeoflife
18:06 towodo new laptop arrived!  16G
18:31 blackrim joined #opentreeoflife
18:31 blackrim josephwb: ok, arcweights=2 seems to "fix" it. it also creates some weirdness in the names
18:33 josephwb i will try that
18:34 blackrim if you ignore internal names, that seems to take care of it.
18:34 blackrim on my weak example
18:35 josephwb how to ignore internal names?
18:36 blackrim oh,we already do with the tree afterwards. so all is good. i am going to try on the mammals
18:36 josephwb just arcweights=2?
18:39 blackrim change from 1 to 2
18:42 josephwb will do
18:49 josephwb blackrim: yup, works for me too
18:49 blackrim cool. i am going to keep on this track for the mammals. then we can tqake to synth
18:50 AndroUser2 joined #opentreeoflife
18:51 josephwb yeah, i didn't have to mls_posttree_fix
18:51 blackrim ok, no knees at all?
18:51 josephwb my names look ok
18:52 josephwb i have knees
18:52 josephwb we need to prune these out prior to synthesis?
18:52 blackrim yeah, we still need to delete the knees for upload into the graph, we could possibly leave out the checktaxhier though
19:09 blackrim ok, loading into the graph. i will synth and extract after this
19:09 josephwb mine is still going
19:09 josephwb in mls
19:11 blackrim it is loading fine in the graph.
19:15 josephwb justtrees?
19:15 blackrim no, pgloadindnew graphdb newicktree
19:16 blackrim requires tips be ottolids
19:16 josephwb go it
19:16 josephwb got it
19:24 blackrim ok, mammals look ok now
19:24 towodo are you keeping a list of the nodes you check every time? ...
19:25 towodo or is it eyeballing a screenful of orders?
19:25 blackrim i didn't check any nodes. just general overview (we had low bar for these tests on new supertree methods)
19:25 blackrim i am going to get more serious about the plants and will get a list and will put it in gcmdr
19:25 towodo you said 'mammals'… I was considering that to be a node… but good
19:26 blackrim that is a node, but it was also the only set of trees included
19:26 towodo sorry to interrupt
19:27 blackrim no problems and thanks for reminding me to note the tests. making that file now so we can start to keep track and have a regular test
19:36 blackrim ok, my load_mls is broken because the fixNamesFromTrees isn't working. have you noted this josephwb
19:37 josephwb what step are you at?
19:38 blackrim just loading normal nexson files
19:38 blackrim it is turned off actually. so something else i guess
19:39 josephwb with load_mls?
19:39 josephwb that just makes the MLS Nexus input
19:39 blackrim yeah, but it gets the ottol ids
19:39 josephwb right. is it not getting them?
19:40 blackrim i think i got it. something silly i was doing.
19:40 josephwb ok
19:40 josephwb my mls is still running. it should be alright
19:40 josephwb right?
19:40 blackrim ok, fixed.
19:41 blackrim i am going to run rosids and send to plant people. then back to fungi!
19:51 josephwb blackrim: looks like arcweights fixes things
19:51 josephwb but i am getting some weird internal node labels
19:52 blackrim yeah, that isn't a problem (it is where there is the disagreement). won't be a problem for loading because we ignore those labels
19:52 josephwb it seems that if things get sunk, MLS names the node XXXXX-YYYYY
19:52 josephwb right
19:52 blackrim yeah
19:53 josephwb can treemachine read in a tree with knuckles?
19:53 blackrim it can read it but the taxon mapping fails
19:53 josephwb yes, of course
19:53 blackrim you have to delete knees at least (i still run it through mls_posttrree_fix but maybe delete knees is all it needs)
19:55 jimallman joined #opentreeoflife
20:00 josephwb ok, trying to do final bird step
20:02 josephwb seems to have stalled after mapping ottIds to names
20:03 josephwb ok, still going
20:03 blackrim for pgloadindnew?
20:03 josephwb yes
20:03 blackrim it is a big tree right ? my guess is that it is just taking a bit because of that
20:04 josephwb the bird supertree
20:04 josephwb 3426 names
20:05 josephwb ~40 trees
20:11 josephwb ok, i've got a tree
20:12 josephwb looks pretty good
20:12 josephwb will run a proper gcmdr+treemachine synthesis so i can compare them
20:13 josephwb why the hell is arcweight=1 the default?!?
20:13 blackrim no idea
20:17 blackrim sent the rosids to doug and bryan
20:19 josephwb ok, will compare the bird trees 2nite
20:19 blackrim sounds good
20:19 josephwb BTW: doesn't seem like A/C is working out here
20:20 blackrim arg
20:20 josephwb set for 66, but it is ~73
20:20 blackrim i will send another email. what is the temp?
20:20 blackrim ok
20:20 josephwb not unbearable
20:21 blackrim yeah, sometimes making it go up a bit and then down fixes it. we will see
21:27 jimallman joined #opentreeoflife
21:51 josephwb joined #opentreeoflife
22:09 josephwb joined #opentreeoflife
23:06 josephwb joined #opentreeoflife

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary