Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2015-01-26

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

All times shown according to UTC.

Time Nick Message
02:47 ilbot3 joined #opentreeoflife
02:47 Topic for #opentreeoflife is now Open Tree Of Life | opentreeoflife.org | github.com/opentreeoflife | http://irclog.perlgeek.de/opentreeoflife/today
05:55 mtholder joined #opentreeoflife
09:17 mtholder joined #opentreeoflife
10:58 1JTAAGL95 joined #opentreeoflife
12:53 mtholder joined #opentreeoflife
14:31 jar286 joined #opentreeoflife
14:41 mtholder jar286, do you know if we post OTT in newick anywhere?
14:41 jar286 I don’t think so
14:41 jar286 could be done without to much difficulty… smasher has a newick generator
14:41 jar286 except the tip labels are very poor quality
14:42 mtholder I can do it, too. just wanted to check my code.
14:42 jar286 but that’s easily fixed if I know what to do
14:42 mtholder we should post the pruned version that treemachine uses somewhere, as that is harder to replicate.
14:43 jar286 that would be a stephen thing, i think
14:43 jar286 although actually it should be possible to replicate, without too much difficulty
14:44 jar286 maybe… well maybe it’s not just the flags, it might have other criteria as well
14:44 mtholder I may have it in peyotl now. but it is not easy to check
14:44 jar286 right
14:44 mtholder (the peyotl version just culls based on flags)
14:44 jar286 taxomachine has a different set of flags for culling, you know
14:45 mtholder I didn't know that.
14:45 mtholder thanks for pointing that out.
14:45 jar286 yep. there’s a flags file in the sources, check it out
14:45 mtholder will do.
14:45 mtholder right now I'm trying to figure out: https://github.com/OpenTreeOfLife/treemachine/issues/78
14:46 jar286 but treemachine doesn’t get its taxonomy from taxomachine. at least I don’t think so
14:46 mtholder I think that only the treemachine behavior matters for that bug.
14:46 mtholder treemachine definitely has some functions for reading the taxonomy.tsv
14:48 jar286 re #60, edges in the synthetic tree (and maybe the GOL) can correspond to *paths* in the input when there are conflicts
14:48 jar286 stephen calls it ‘sinking’
14:49 jar286 oops i didn’t mean #60, I meant ‘the 60’
14:49 mtholder but there should not be an edge in the synthetic tree that conflicts with every input, right?
14:50 jar286 don’t think so
14:50 mtholder I think that is what Ruchi is claiming to have found.
14:51 jar286 but how does ‘groupings in the synthetic tree that do not appear to be backed by any source tree.’ imply conflict?
14:51 mtholder that is from email from her.
14:51 mtholder "conflict" that is
14:52 mtholder but it should also be the case that every edge in the synthetic tree can be mapped to at least 1 edge in an input that it corresponds to, right?
14:52 mtholder in the supplemental Ruchi and Gordon describe their calculation in a little more depth.
14:53 mtholder but it is not terribly clear to me, and I don't think that the code is open source. Or is it?
14:54 jar286 like I say a synth tree edge will always map to (or come from) a path in a synthetic tree…
14:54 jar286 it’s open tree of life, so it must be open :)
14:54 mtholder s/must/should/
14:54 mtholder :-)
14:58 josephwb joined #opentreeoflife
15:01 jar286 mtholder, i’ve been thinking about 3-way merge a lot - the test case it trivial
15:01 jar286 s/it/is/
15:01 mtholder where is that test case?
15:02 jar286 I can describe it easily - take any two trees where the alignment is ambiguous. tree #3 is simply one of the merges
15:03 mtholder josephwb, do you have a newick of the taxonomic tree after treemachine have pruned off the parts that it does not use?
15:04 josephwb hmm. somewhere.
15:04 josephwb or i can make it anew.
15:04 josephwb is this about the newick labelling?
15:04 mtholder If it is not too much trouble that would be great. I tried to mimic, but need to check.
15:04 mtholder no. not about labeling
15:05 mtholder it is about  https://github.com/OpenTreeOfLife/treemachine/issues/78
15:05 josephwb okay, give me a sec to look around.
15:07 josephwb mtholder: i can't find it immediately, so i will just regenerate it. will take a bit, as it has to do the whole thing.
15:08 jar286 ‘whole thing’ = synthesis?
15:08 mtholder I can probably get my version of treemachine running.
15:08 josephwb import of taxonomy + synthesis + export of newick
15:08 mtholder it'll get spit out when I load the taxonomy?
15:08 mtholder gotcha.
15:09 josephwb i found an old copy. sending it now while i run a fresh copy.
15:09 mtholder thanks.
15:10 josephwb sent.
15:14 josephwb mtholder: that should reflect what is in the synthetic tree.
15:14 mtholder thanks
15:14 josephwb (i think)
15:15 josephwb i am happy to do the newick label thing once we have rules. very easy to do.
15:15 jar286 is newick label rules a hot potato? i suppose I should make sure they get decided & written up
15:17 josephwb rules are pretty established: http://evolution.genetics.washington.edu/phylip/newick_doc.html
15:18 josephwb but we have issues: https://github.com/OpenTreeOfLife/treemachine/issues/147
15:19 mtholder it is not a blocker for me. I'm just going to use the ott ID. We have a few ways of encoding that, so I'll have to check for all of them - which is a bit tedious, but not difficult.
15:19 josephwb problem: repos do their own thing at the moment (i.e. treemachine, taxomachine)
15:20 jar286 when I said ‘rules’ I meant ‘rules specific to open tree’ e.g. how to decide labels for taxonomy nodes
15:20 josephwb i don't think those two will produce identical label sets
15:20 jar286 also information about tools that incompletely implement the washington rules
15:20 josephwb treemachine does a weird thing to get dendroscope to work, too.
15:21 jar286 that’s the kind of thing we need to document all in one discoverable place
15:21 josephwb right now, the "jade" class is duplicated across repos. want a single copy to live in OT-base.
15:21 josephwb right, i agree.
15:22 mtholder docs would be good. But we should also just emit the legal newicks and write tools to produced wonky version for tools like dendroscope as post-processing.
15:25 josephwb yup.
15:26 jar286 https://github.com/OpenTreeOfLife/germinator/issues/13
15:28 pmidford2 joined #opentreeoflife
15:36 josephwb mtholder: newick labels can contain curly brackets, yes?
15:38 josephwb the "standard" doesn't mention them.
15:38 mtholder yes. but they must be quoted in nexus. I usually try to emit newicks that can be easily converted to NEXUS.
15:38 josephwb ah, there's the rub.
15:39 mtholder any label can be expressed in newick and nexus if you quote correctly.
15:39 mtholder with the same quoted string, I should say.
15:39 mtholder A little extra quoting is not a huge problem.
15:39 josephwb right, but you are saying they have different quoting rules with regards to curly brackets?
15:40 mtholder they can be quoted in newick (anything can). they must be quoted in nexus.
15:40 mtholder so if you quote them, you're fine in both.
15:40 josephwb that is what is being done now.
15:40 josephwb any other weird stuff?
15:41 mtholder they main weird stuff now is removing the punctuation.
15:41 mtholder and the handling of _
15:41 mtholder once you quote a string, you can leave the punctuation in it.
15:41 josephwb right. i meant with quoting rules.
15:41 josephwb between newick and nexus.
15:42 mtholder the nexus token breakers are )[]{}/\,;:=*'"`+-<>~-
15:42 josephwb i wonder, tho. the function we are calling is "newickName". should this be *just* straight-up newick? probably, right?
15:43 josephwb have separate for nexus.
15:43 mtholder that would be OK with me, but if we just emitted newick that was also parseable as nexus from the same function it would be OK.
15:43 josephwb i guess this is what needs to be decided jar286.
15:44 jar286 my view is reflected in the issue I filed… but I’m not sure it should be up to me
15:44 jar286 I added ‘nexus’ to the issue description
15:47 jar286 I’m not keen on a postprocessor, since it makes things more fragile
15:47 mtholder the problem is that some tools require illegal newick strings.
15:47 jar286 oh, ouch!!
15:48 jar286 how can that be? there has to be a common subset
15:48 mtholder bugs
15:50 mtholder I guess it is more common for the buggy tools to disallow some legal strings rather than requiring illegal forms. but I think that some of the things done to accommodate dendroscope would break standard-compliant parsers.
15:51 pmidford2 joined #opentreeoflife
15:54 jar286 grumph.
15:54 pmidford2 joined #opentreeoflife
15:55 josephwb mtholder: is there a reference for the nexus standard beyond the maddison 1997 paper? extended nexus?
15:57 mtholder no.
15:57 mtholder I'm afraid Mesquite broke some things.
15:57 mtholder in the sense of introducing new syntax for old construct.s
15:58 mtholder but the trees haven't changes AFAIK
15:58 mtholder or the TAXA
15:58 josephwb ok, thanks.
16:00 josephwb so, if spaces are the only "illegal" characters, replace with underscore, but if more illegal characters are present, quote and keep the sapces?
16:00 mtholder yes.
16:01 josephwb great, thanks again.
16:01 mtholder and there is the obnoxious ' issue
16:01 josephwb true dat.
16:01 mtholder if the string has a ' then it has to be quoted and the internal becomes 2 ''
16:01 josephwb yes, that should be working
16:01 josephwb that was set a while ago.
16:01 mtholder yeah. I think that it is.
16:02 jar286 can’t find maddison on line so far…
16:02 jar286 1997 that is, not dave
16:03 mtholder I've got a searchable copy (a preprint) that I can mail you.
16:03 josephwb http://sysbio.oxfordjournals.org/content/46/4/590.full.pdf
16:11 jar286 just want to link it from the issue description
17:47 mtholder joined #opentreeoflife
18:22 jimallman jar286: not sure if you’ve had time to review the CA cert-bundle yet, curious about your opinion on the cleanest fix for opentree#553: https://github.com/OpenTreeOfLife/opentree/issues/553
19:20 mtholder joined #opentreeoflife
20:40 jar286 jimallman, I read the thread but have no particular suggestion, especially if there’s no way to satisfy the recommendation. I’m really glad you pursued this
20:41 jimallman i was just checking to see if you have a preference regarding building a certificate “bundle” (single file with all intermediate CA certs) versus using the separate files as provided in the Comodo .zip)
20:42 jar286 no opinion.  i’ve never done this kind of thing
20:42 jar286 i would look to see how it’s done elsewhere…
20:42 jimallman OK. i’d like to stick with the bundle file, then. i’m under the impression that the separate-files approach needs be set up with numbered symlinks, etc.
20:43 jar286 right. whatever’s most direct
20:43 jar286 do you need anything else from namecheap? a new bundle or anything?
20:44 jimallman (by the way, i’m still not sure why this is suddenly a problem. apparently lots of intermediate CA certs are held in browsers and OS cache. my guess is that Comodo might have changed or retired an intermediate cert, so suddenly various clients are confused..?)
20:44 jimallman i don’t think we need anything new, no. these certs seem to work fine once we provide the full chain.
20:45 jar286 ok.
20:45 * jar286 sorry, i’ve got to take a break, should be back in about an hour
20:45 jimallman no problem, i should be able to wrap this up.
21:52 josephwb hey jimallman
21:53 josephwb is the study "score" communicated in the nexson?
21:54 josephwb e.g. this study is 92% complete: https://tree.opentreeoflife.org/curator/study/view/pg_440/?tab=metadata
21:54 josephwb maybe pmidford2 knows?
21:56 pmidford2 Not off the top of my head, though I though those were calculated by the curation tool but not saved.  I don't remember seeing an annotation tag for them.
21:56 pmidford2 though I thought...
21:57 josephwb yes, calculated in the curation tool. i'm trying to find "sufficiently curated" studies via the phylesystem api
21:57 josephwb i can check individual properties, but a score would be fatser\
21:57 josephwb faster
21:58 josephwb more curious than anything at this point since I do in fact individual properties
21:58 pmidford2 Sounds like you want to grab jimallman's scoring code, don't think there's an obvious other way
21:58 josephwb ok, thanks.
22:19 mtholder joined #opentreeoflife
22:23 jar286 Just discovered lobste.rs
22:24 jar286 Ferrier is no fan of Scrum. “Sprints are utterly ridiculous ways to produce software,”…
22:39 * jimallman is back now..
22:42 jimallman josephwb: the scoring is strictly client-side. here’s the procedure:
22:42 jimallman https://github.com/OpenTreeOfLife/opentree/blob/c93821770c7ad93258b2712c716ca62fd7340f7a/curator/static/js/study-editor.js#L2618
22:43 jimallman …and more importantly, the criteria (logic and weighting for each): https://github.com/OpenTreeOfLife/opentree/blob/c93821770c7ad93258b2712c716ca62fd7340f7a/curator/static/js/study-editor.js#L2292

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