Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2014-07-23

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

All times shown according to UTC.

Time Nick Message
00:14 kcranstn joined #opentreeoflife
00:21 kcranstn why would I be getting an “You have provided an invalid or expired authentication token” error?
00:21 kcranstn same token that worked earlier today. These don’t have a short expiry, do they?
01:55 josephwb joined #opentreeoflife
01:57 josephwb joined #opentreeoflife
03:24 josephwb joined #opentreeoflife
03:48 josephwb joined #opentreeoflife
04:07 josephwb goodnight ilbot3. see you in the morning.
04:07 josephwb you too jimallman
09:50 scrollback joined #opentreeoflife
10:46 josephwb joined #opentreeoflife
10:52 josephwb morning, ilbot3
10:59 josephwb joined #opentreeoflife
11:23 josephwb joined #opentreeoflife
12:18 josephwb joined #opentreeoflife
13:08 josephwb jimallman?
13:15 jimallman josephwb: howdy
13:22 josephwb hey. stephen was wondering which machine was the dev machine at the moment.
13:23 josephwb ot10?
13:23 josephwb he wants to update the treemachine db
13:31 jimallman let’s see. i lose track, so dig keeps me straight:   $ dig devtree.opentreeoflife.org
13:32 jimallman oh wait, the db goes to devapi…. so that’s ot10
13:52 kcranstn joined #opentreeoflife
13:55 kcranstn @jimallman?
13:56 jimallman kcranstn: hi!
13:56 jimallman i haven’t chased down the expired-token problem yet. is that a blocker for you?
13:56 kcranstn morning. Question about the API docs… why do these calls point to localhost? https://github.com/OpenTreeOfLife/phylesystem-api/tree/master/docs#updating-a-study
13:56 jimallman (they should have a long life, as you said)
13:56 kcranstn and expired-token is not blocking me, just concerning
13:57 jimallman agreed.
13:57 jimallman localhost looks like someone just plugged in URLs that worked on their test system. we should certainly change these to a single domain in all examples.
13:57 kcranstn ok, will do
13:58 jimallman also any URL with /api/ should now read /phylesystem/ instead.
13:58 kcranstn yup - got that
14:04 josephwb wonder if we will get any cross-subscribers from this "TOL": http://www.theologyonline.com/forums/showthread.php?t=96104
14:04 josephwb *was searching for TOL APP*
14:05 josephwb kcranstn: did you get to check out the ToLApp at evol2014?
14:06 kcranstn no, I didn’t
14:07 josephwb oh, maybe he wasn't there. not clear.
14:07 travis-ci joined #opentreeoflife
14:07 travis-ci [travis-ci] OpenTreeOfLife/phylesystem-api#573 (master - a860821 : kcranston): The build passed.
14:07 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/phylesystem-api/compare/7a45cf62eb8c...a86082156a5a
14:07 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/phylesystem-api/builds/30654257
14:07 travis-ci left #opentreeoflife
14:09 towodo joined #opentreeoflife
14:29 * jimallman is trying to find towodo in another window..
14:29 * towodo here
14:30 * towodo (Funny that the IRC bot logs all of the /me lines. The W3C IRC bot doesn’t do that, which means that /me is available there for private (unlogged) conversation)
14:31 towodo jimallman, good morning
14:32 jimallman howdy. i didn’t get to that last to-do yesterday, will try this morning
14:32 towodo ok
15:48 josephwb ack. running into problems because treeIDs are no longer unique.
15:51 kcranstn joined #opentreeoflife
15:53 kcranstn ^ josephwb - explain?
15:54 josephwb kcranstn: a bunch of treemachine code deals with treeIDs alone. Fine when unique, but now curator gives all trees treeID = 1
15:54 josephwb so, if I list trees in the DB, I get: 1 1 1 1 1 1 1 1 1 1 1 ...
15:54 josephwb even though they are distinct
15:57 josephwb PG gave a unique ID to each tree, so you could retrieve trees with that value alone (i.e. not knowing the studyID, or the SHA)
16:15 kcranstn I can’t recall that was ever discussed as a design decision
16:15 jimallman josephwb: in other places, we’ve used concatenated {STUDY_ID}_{TREE_ID} to keep these distinct. it’s not ideal, but could that work as a short-term fix?
16:16 jimallman kcranstn: it’s come up a number of times (abandoning unique tree ids across the docstore). i don’t recall any strong argument against it in principle, but it does break code that assumes unique tree ids.
16:16 josephwb kcranstn I  brought it up but was denied
16:17 jimallman you saw this day coming…
16:17 josephwb it is fine. we'll just change the treemachine code
16:18 josephwb but I think unique treeIDs are useful
16:18 kcranstn we could assign UUIDs to trees rather than unique-within-study IDs
16:18 josephwb like PG
16:18 kcranstn @towodo?
16:19 kcranstn the challenge is coordinating across curator and phylografter.
16:19 kcranstn could ensure unique-within-namespace
16:20 kcranstn do the pg ids include the namespace?
16:20 josephwb Just to be clear: it will soon be a non-issue for treemachine.
16:20 kcranstn ok, I will stop fretting
16:20 towodo hi
16:20 jimallman hm, it looks like at least one other person (temi) has been caught by this: https://github.com/OpenTreeOfLife/treemachine/issues/89
16:20 kcranstn although I can certainly see people wanting to refer to a specific tree
16:21 towodo reading
16:21 jimallman here’s some prior discussion and the concatenated-id “solution”:  https://github.com/OpenTreeOfLife/treemachine/issues/99
16:21 josephwb right; if anything, it cannot hurt things, but may be useful
16:22 towodo we coordinate study ids. I don’t see why studyid!treeid or the like wouldn’t work (separator can be any non-xml-ncname character)
16:23 towodo uuids give a false sense of security
16:24 towodo don’t use _ please
16:25 towodo josephwb ^
16:25 kcranstn it looks like cody might have some of this code for treemachine on a branch
16:29 josephwb which branch?
16:30 kcranstn https://github.com/OpenTreeOfLife/treemachine/tree/store_tree_ids_and_commit_shas
16:30 kcranstn (see issue 99 that jim posted, above)
16:33 josephwb towodo: I am naive as to why '_' is so blasphemous.
16:33 towodo because it’s ambiguous
16:33 towodo there is no deterministic way to separate the study id from the tree id.
16:33 kcranstn because we use ‘_’ in the study IDs
16:34 towodo A_B_C could be (A_B) C or A (B_C)
16:34 kcranstn a split on ‘_’ won’t separate studyid from treeid
16:34 kcranstn if we use the same separator
16:34 towodo _ could appear in a tree id. we don’t control tree id except that it has to be an ncname
16:35 towodo it’s not blasphemous, it’s just wrong.
16:35 towodo no religion here
16:37 towodo we had talked about using JSON syntax {“study”: “pg_123”, “tree”:”b_c”} but if that’s too verbose use a separator like ! or # (just not _ - .)
16:38 josephwb ok, didn't think about a treeid containing '_'
16:39 towodo when we import nexml we keep the tree ids that were there. no control.
16:39 josephwb regarding the studyids, *we* (treemachine/gcmdr people) were using underscores first :-#
16:40 josephwb so, what is better than '_'?
16:40 josephwb what character will not appear in a studyid/treeid?
16:41 josephwb towodo kcranstn ^
16:41 towodo I said above. anything that’s not an ncname char is fine. ncname chars are alpha, numeric, _ - .
16:42 towodo (including dot.)
16:42 towodo probably not : either
16:43 towodo http://stackoverflow.com/questions/1631396/what-is-an-xsncname-type-and-when-should-it-be-used
16:43 josephwb *looking up what ncname is*
16:46 josephwb Looks like the following are kosher: :, @, $, %, &, /, +, comma, semicolon
16:47 josephwb and a whack of others
16:55 travis-ci joined #opentreeoflife
16:55 travis-ci [travis-ci] OpenTreeOfLife/phylesystem-api#574 (master - adf5842 : Mark T. Holder): The build passed.
16:55 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/phylesystem-api/compare/a86082156a5a...adf5842b770b
16:55 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/phylesystem-api/builds/30671009
16:55 travis-ci left #opentreeoflife
17:03 josephwb ok towodo, found it: [\i-[:]][\c-[:]]*
17:03 josephwb so, we can select from:
17:03 josephwb , !, ", #, $, %, &, ', (, ), *, +, ,, /, :, ;, <, =, >, ?, @, [, \, ], ^, `, {, |, }, ~
17:04 josephwb Since "NC" stands for "non-colonized", seems a colon is perfectly fine.
17:05 josephwb I'll talk to Stephen about overhauling things.
17:06 josephwb I inclined to go with ':'
17:06 josephwb '#' won't work e.g. if processed in python scripts (as in gcmdr)
17:07 josephwb *thinking out loud, no need to respond*
17:09 josephwb treemachine has 21 branches! Maybe we should purge the deprecated ones.
17:16 * towodo back from lunch
17:16 towodo I don’t recommend colon.
17:17 towodo too close to what XML already does.
17:17 towodo but I think it would work.
17:17 towodo I think I would use !
17:18 towodo *responding out loud, no need to think*
17:36 jimallman fwiw, i find that the tilde (~) works well as a visual separator
17:38 josephwb jimallman: yes, i was thinking that too. '!' is harder to see.
17:39 towodo space ?
17:39 jimallman you might also consider multi-character separators like “::”, but that can be weird.
17:39 jimallman space? hmmm, why not?
17:39 towodo good separator
17:39 josephwb doesn't work with commandline analyses
17:40 towodo ouch. so a requirement is not “” in the shell.  well neither does ~, right?
17:40 towodo oh, ~ is only special at beginning
17:40 josephwb i think so
17:41 towodo ! is special
17:41 towodo + works well
17:43 josephwb i like +
17:43 josephwb +1
17:45 towodo even looks like a tree
17:45 josephwb right
17:47 josephwb Dail and Laura are wondering why no one responded to their "dark area of the tree" email.
17:47 josephwb Oh, maybe this is a question for kcranstn
17:49 towodo looks like 1) is a question for stephen, 2) a question for kcranstn.
17:49 towodo if I speak up people will start expecting me to speak up, and that wouldn’t be good.
17:52 towodo figuring out where curator effort should go is top priority
18:00 kcranstn joined #opentreeoflife
18:02 josephwb towodo: where is that list of Genera you pointed me to the other day?
18:03 towodo what, the short names in alphabetical order?
18:03 towodo why?
18:03 josephwb oh, is it just short names?
18:03 towodo yes, why?
18:04 josephwb was just going to bulk add it to my dictionary so my computer didn't think every scientific name was a typo
18:04 josephwb not important
18:05 towodo you could just use the taxonomy instead.  only 6 million words
18:05 josephwb right
18:06 josephwb easy enough. just if you had the text file up it would save me a little typing
18:07 towodo the list I showed you is just genus names up to about 7 or 8 characters
18:07 josephwb ok
18:08 josephwb alright, back to work
18:09 josephwb oh, did you ever return to communicating without Cody's RepresentationConverter?
18:09 towodo no
18:11 scrollback joined #opentreeoflife
19:01 jimallman joined #opentreeoflife
19:57 towodo joined #opentreeoflife
21:11 towodo jimallman, how about those pems? all you have to do is ssh -i ~/.ssh/opentree/production.pem admin@ot18.opentreeoflife.org
21:42 jimallman towodo: working for me now on ot18, as user opentree and admin.
21:42 jimallman thanks
21:42 towodo ok, thanks for checking.  will set key for ot14 and ot15 now
21:43 * jimallman has clobbered the .emp file
21:43 jimallman s/.emp/tmp
21:43 towodo there’s a Cavia pull request waiting for your approval
21:43 towodo no hurry
21:46 jimallman towodo: did you intend to specify branch Bos in the server-config files? i assumed we’re planning on making Cavia branches, even if not much has changed since Bos.
21:46 towodo no. just saw that.
21:47 towodo fix pushed
21:47 towodo tnx
21:48 jimallman i see it. it looks great otherwise, merging now…
21:50 towodo can we set a time? I wish I could do Saturday or Sunday but busy.
21:50 jimallman hm, one other thing: for Bos, we started with the dev data and docstore. i’m not expecting trouble, but should we do the same here to begin, then switch over once we’re sure (and have pulled Bos offline)?
21:50 towodo boy, file transfers from MIT to EC2 are really slow.  and not at the MIT end
21:51 towodo oh. hmm.
21:51 towodo sure. let me see if there’s a comment to guide me
21:52 jimallman i think the TODO comments cover these last pivots in each config file
21:52 towodo maybe I deleted them.
21:52 jimallman a couple  :)
21:52 towodo oh i see
21:53 towodo # **** Use phylesystem-0 for testing, phylesystem-1 for release
21:53 towodo OPENTREE_DOCSTORE=phylesystem-0
21:53 jimallman i’ll check my notes, but i think the switch from phylesystem-0 to -1 (and attendant .pem) was one TODO (still there). the other was the value of $OPENTREE_API_HOST, which starts as ot18 so we can test before the DNS switch
21:53 towodo what else? same for identity?
21:54 josephwb jimallman: I'm going to bug you about this again: how do I get newick from the curator?
21:54 josephwb nothing under "Files"
21:54 jimallman towodo: OPENTREE_GH_IDENTITY should switch along with OPENTREE_DOCSTORE, yes
21:55 josephwb nevermind; I'll just pass the nexson through treemachine
21:55 jimallman towodo: … also OPENTREE_API_BASE_URL  (see my comment above)
21:55 jimallman josephwb: sorry, trying to complete a thought.
21:55 jimallman reading your question now
21:56 jimallman ah yes, Newick export. i need to revisit these options with mtholder, and see which ones can work with the current ncl. (that one might have been wishful thinking on my part, sorry)
21:56 jimallman josephwb: if this is a pressing thing, i can move that to the front burner
21:56 josephwb no problem at all; i've got an easy workaround
21:57 josephwb how many burners do you have?
21:57 jimallman cool, thanks. this is definitely on my TODO list
21:59 jimallman josephwb: http://onefatfrog23.files.wordpress.com/2011/09/imperial1.png
22:00 jimallman (but some days it’s more of a hot plate with a flaky cord)
22:02 jimallman towodo: i’m trying to think of a better mnemonic than TODO for these last-minute config switches… PIVOT ? RELEASE ? PREFLIGHT ?
22:02 towodo oh, I’ve already done something.
22:02 towodo I’m using ***** plus prose.
22:02 towodo just committing & pushing for your review now
22:02 jimallman thanks, that sounds great
22:03 towodo ok look now
22:04 jimallman re: when to do the migration, i’m heading to Asheville for a few days so my schedule is also weird. we’re driving up tomorrow, so i can make some time Friday, preferably early morning or evening..? i’m at the dentist Monday AM, but i’ll be back in the saddle here bt 11am if that works better for you.
22:05 jimallman that hint for phylesystem-0.pem won’t work for me, unless i rename the file (which i’m glad to do, but we’ll need to let others know to do the same)
22:05 towodo oh. I’m just thinking josephwb has made this sound urgent, so I’m thinking Friday, but it can wait if you’re preoccupied with having a good time (wish I were going to asheville?)
22:05 towodo let me look.
22:06 towodo add .pem to the comment?
22:06 jimallman agreed that the sooner, the better. let me check with my social secretary here, maybe Thursday night?
22:06 towodo no, I’m busy Thursday night
22:07 jimallman i’m tempted to add both versions, pre-filled in and commented  ## FOR TESTING:  and  ## FINAL RELEASE:  or somesuch
22:07 jimallman that way it’s just a matter of (un)commenting the right line
22:07 towodo hmm, that’s a good idea
22:07 jimallman oh, tonight also works for me
22:08 jimallman If We Dare
22:09 towodo hmm… I still need to do the oti initialization, and do basic checks. i think tonight is too soon
22:10 jimallman ok. checking my schedule here.
22:10 towodo new version of Cavia for your review
22:11 towodo do you happen to remember where Mark’s API tests are?  (before I email him?)
22:15 towodo never mind, emailed Mark
22:28 jimallman towodo: i’ll email regarding schedule and possible working windows
22:28 towodo ok.  I’ll work on neo4j db installs (copying now)
22:28 towodo headed home now, ttyl
22:50 josephwb joined #opentreeoflife
23:42 josephwb joined #opentreeoflife
23:52 josephwb joined #opentreeoflife
23:53 josephwb joined #opentreeoflife

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