Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2015-01-22

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

All times shown according to UTC.

Time Nick Message
00:03 kcranstn joined #opentreeoflife
01:06 jar286 joined #opentreeoflife
01:30 jar286 joined #opentreeoflife
01:56 jar286_ joined #opentreeoflife
02:46 jar286 joined #opentreeoflife
10:38 64MABT4CU joined #opentreeoflife
11:59 jar286 joined #opentreeoflife
12:13 jar286 joined #opentreeoflife
12:42 kcranstn joined #opentreeoflife
13:34 kcranstn joined #opentreeoflife
13:36 jar286 joined #opentreeoflife
14:05 jar286 joined #opentreeoflife
16:00 jimallman joined #opentreeoflife
17:43 guest|67001 joined #opentreeoflife
18:09 mtholder joined #opentreeoflife
18:47 kcranstn joined #opentreeoflife
18:56 pmidford2 joined #opentreeoflife
19:02 jimallman PR review?
19:02 kcranstn here!
19:03 jar286 here
19:03 pmidford2 here
19:03 jar286 going from right tab to left tab in my browser…
19:04 kcranstn does that count as ‘technical difficulties’?
19:04 jar286 hmm. only one repo has PRs… that’s opentree
19:04 jar286 https://github.com/OpenTreeOfLife/opentree/pull/536
19:05 jar286 deployed?
19:05 kcranstn I see more than that in my github notifications
19:05 kcranstn https://github.com/OpenTreeOfLife/opentree/pulls
19:05 jar286 right, I was doing one at a time, starting with oldest
19:06 jar286 per usual
19:06 kcranstn oh, only one repo (not only one PR)
19:06 jar286 sorry.
19:06 jar286 one repo
19:06 kcranstn ok, then, 536
19:06 jar286 jimallman, can we look at the page now?
19:07 jimallman #536 is on devtree, but needs lots of discussion and fresh stats
19:07 kcranstn link?
19:07 jimallman just a sec
19:07 kcranstn https://devtree.opentreeoflife.org/about/progress
19:07 jar286 shall we table this pending real stats?
19:07 kcranstn don’t see anything
19:07 jimallman ! not sure why this is blank, sorry
19:08 kcranstn let’s talk abotu the stats
19:08 jar286 ok.
19:08 jar286 pmidford2, status?
19:08 pmidford2 checking
19:09 jimallman i tried to start this conversation via email (opentree-software, Jan 16)
19:09 jar286 I think the two of you need to talk.
19:09 jimallman yes, and mtholder as well
19:09 kcranstn I think it was hard to get feedback on the vis when there was no data
19:09 jimallman (and jar286, mostly regarding taxo-release artifacts)
19:09 jar286 why mtholder?
19:09 pmidford2 the server is working
19:10 pmidford2 stats are going where I think they should be
19:10 jar286 does jimallman know where the files are being deposited?
19:10 jimallman ok. just be a problem in the page.
19:10 jar286 and did you coordinate on format? maybe there should be a wiki page
19:10 jar286 describing format
19:11 pmidford2 Agree about format wiki page
19:11 pmidford2 There is format information in an issue discussion (need to find it)
19:12 mtholder hi
19:12 mtholder sorry I'm late
19:12 jar286 hi!
19:12 jimallman i believe this was working before (stats showing on the page), looking for problems now…
19:12 pmidford2 The stats are going to devtree://home/opentree/web2py/applications/opentree/static/statistics
19:13 jimallman hm, looks like i am fetching from the wrong place. it’s an empty list in the page.
19:13 pmidford2 Where are you fetching from?
19:14 jimallman mtholder: my more general stats questions are in email to opentree-software, Jan 16
19:14 jimallman pmidford2: checking now...
19:14 mtholder thanks
19:14 jar286 (would be nice if ‘show server details’ showed branch name)
19:14 jimallman noted
19:15 jar286 will issue that
19:16 kcranstn why don’t we go through the rest of the PRs, and then if there is time left, we can talk about the Qs in Jim’s email abotu the progress page?
19:17 jar286 yes, that was a long pause.
19:17 kcranstn as opposed to debugging now
19:17 jar286 https://github.com/OpenTreeOfLife/opentree/pull/556
19:17 jimallman agreed, thanks. i do see a problem with stats paths, will address this when we’re done
19:18 jar286 is this PR isolated from the bulk of the code? i.e. any risk in deployment?
19:18 jimallman #556 is the cleanest solution to stale/misleading MRCA-test results in Nexson. taking them out!
19:18 kcranstn looks fine to me
19:18 jimallman yes, this should be isolated and compatible with master
19:18 mtholder looks good to me. I need to script deletion of the cruft from phylesystem-1
19:18 mtholder at some point.
19:19 pmidford2 looks good
19:19 jimallman yes, no problem having it there on client side fwiw
19:19 jar286 and you tested it in development servers I presume
19:19 jimallman yes
19:19 jar286 ok, i’ll merge
19:19 jar286 https://github.com/OpenTreeOfLife/opentree/pull/559
19:20 mtholder jimallman in JS you can delete obj[prop] even if prop is not in obj?
19:20 jar286 looking at it…
19:20 jimallman yep!
19:20 mtholder cool.
19:20 mtholder in a sloppy sort of way :-)
19:20 jimallman yeah, scary-cool
19:20 jimallman note the color-scheme variations in comments for #559. opinions?
19:21 jar286 clicking on them still doesn’t work in the lineage
19:21 mtholder (the color blind guy won't weigh in on that...)
19:21 jimallman lineage “rootward” from target node is still missing some data… will need to address this separately
19:22 kcranstn I like blue over red for info
19:22 jimallman mtholder: your input is actually very welcome… contrast and all that.
19:22 jimallman agreed, red was alarming/overkill
19:22 jar286 right about lineage. but maybe they shouldn’t have ‘i’s given that
19:22 pmidford2 I think blue is more standard as well
19:22 mtholder looks fine to me.
19:22 jar286 but it looks great, I want to merge it
19:22 jimallman blue on white? or white on blue?
19:22 kcranstn blue on white
19:23 jar286 yeah...
19:23 jimallman (white on blue is what’s in the PR code, so that’ll take a very quick change)
19:23 pmidford2 fine
19:23 kcranstn #5
19:23 jar286 make the change and merge it
19:23 kcranstn of the 9 samples in the PR
19:23 jimallman jar286: will do...
19:23 jar286 https://github.com/OpenTreeOfLife/opentree/pull/560
19:23 jimallman kcranstn: it’s three sets of three, but i get you
19:24 kcranstn ah, the red dot only appears when you hover?
19:24 kcranstn yes, I get it now
19:24 kcranstn cool
19:24 jimallman yes, it signifies the big jump
19:24 jimallman again, maybe too URGENT?
19:24 kcranstn nah, it’s fine
19:25 jimallman #560 covers a number of minor, related annoyances.
19:25 kcranstn mtholder - is the red / blue distinguishable?
19:25 mtholder yes
19:25 jimallman ! ok then
19:25 kcranstn as our token colour-blind dude
19:26 jar286 re #560, curious how it’s implemented, where the state is. a variable in the javascript? or in the session?
19:27 jar286 but it sounds fine.
19:27 jar286 (we’ll skip the other two PRs per usual, then we’re done)
19:27 jimallman this uses the client-side history object (or falls back to current URL) for state
19:28 jimallman i’ve bumped the Info icon branch, should be ready to merge now (tested colors on my local server)
19:28 jimallman #559, that is
19:28 jar286 oh well. I think I’ll ask you about state later, if I still care
19:28 jimallman see above.
19:29 kcranstn I merged 559
19:29 jar286 so you’re going to merge 559?
19:29 jar286 oh.
19:29 jimallman for state, #560 uses browser history by default, or scrapes the current URL
19:29 jar286 oh, ok.
19:29 jar286 any risk in merging 560?
19:30 jimallman each link implies certain behavior, so i make a few URLs and apply them accordingly
19:30 jimallman no risk IMO, i’ve tested it pretty thoroughly on devtree
19:30 jar286 I had noticed that this was a problem. a lot of web sites show this behavior
19:30 jar286 ok, I will merge
19:30 jar286 done with PR review, thanks.
19:30 jar286 jimallman, can we set a time for deployment to production?
19:31 jimallman sure!
19:31 jimallman i’m flexible here, though tonight i have a python meetup and will need to leave home ~6pm
19:31 jimallman any time tomorrow or this weekend is fine
19:33 jar286 I’m pretty flexible too. I hate to ask you to attend to this on a weekend… maybe Friday afternoon. I need to check in with pmidford2, hang on…
19:33 jar286 pmidford2, you said 1, and then I said 11…
19:34 jar286 is 11 ok?
19:34 pmidford2 Right - I'd prefer 1, but 11 will work
19:34 jar286 then I can meet with Jim at 1 or 1:30… I have a midafternoon appointment in Harvard Square or I would do deployment later
19:35 jar286 or we could meet 1-1:30
19:36 jar286 up to you
19:36 jar286 I’m probably talking nonsense.
19:36 pmidford2 I would like to sit in on another meeting at 11, but I can skip it
19:38 jar286 no, let’s talk at 1 to 1:30. should keep us focused
19:38 jar286 jimallman, can you do 1:30 on Friday?
19:38 jimallman certainly
19:39 jimallman hopefully it will be a non-event.
19:39 jimallman (no kaboom, i mean)
19:39 jar286 right, but last time I needed you, as I remember
19:39 jimallman true
19:39 jar286 maybe this time we should test ‘master’ on the dev system
19:40 jar286 although that wouldn’t have found the problem I introduced last time
19:40 jimallman pmidford2: looks like i was still trying to fetch statistics from /static/stats/  … fixing this now.
19:42 pmidford2 sorry, I may have broken that last week.  Do the keys make sense to you?
19:45 jimallman let’s see.. looking better now: https://devtree.opentreeoflife.org/about/progress
19:46 kcranstn joined #opentreeoflife
19:47 pmidford2 right - though the points aren't releases, just crontab updates (or did you fake those?)
19:48 jimallman i’m making some very spurious assumptions about release dates.
19:48 jimallman pmidford2: mtholder: now that the stats pages are back online on devtree, is there a good time to talk through the issues from my Jan 16 email? here it is on google groups: https://groups.google.com/forum/#!topic/opentreeoflife-software/HEfdy5MAR0Y
19:48 pmidford2 Sure, makes sense
19:48 * jimallman thinks we may have lost mtholder… he’s grayed out for me
19:49 mtholder back now. hold on..
19:49 jimallman if the UI makes sense, i’m pretty sure we’ll need to add more information to stats or other “transcripts” regarding release dates, taxonomy per synthesis release, etc
19:50 kcranstn I am confused - this still looks like fake data (or at least, woefully incomplete data)
19:50 mtholder I can generate phylesystme stats, by checking out old version of the repo...
19:50 pmidford2 <kcranstn> It's live data from dev, not much going on
19:50 jimallman cool! i wasn’t sure if we could recover much in the way of history, but that would be great. same for synth and taxo releases (recover components, etc)
19:51 mtholder I asssume that we just have 2 synth releases and just want ott 2.7 and 2.8?
19:51 kcranstn sounds good
19:51 jimallman kcranstn: there’s a considerable amount of fakery here.. i’m hoping to keep the UI but switch to real data once we track it all down or create it
19:52 kcranstn let’s talk about what is a reasonable granularity (i.e. ticks on the x-axis0
19:52 kcranstn weekly? monthly?
19:53 mtholder only the phylesystem changes rapidly (and not that rapidly these days)
19:53 jimallman in principle, OTU stats could change with every saved study, right?
19:53 jar286 weekly at this point would be depressing
19:53 jar286 monthly I would say
19:53 mtholder yes we could recalculate with every commit by registering another webhook
19:53 jar286 but maybe configurable…
19:54 jar286 there are different audiences for this feature
19:55 jimallman note that this is assembled from at least two sets of stats files.. phylesystem and synthesis can have different granularity, but i assumed that synth (and ott stats, if any) would only update on release dates.
19:56 pmidford2 Agree, frequency question should be mostly about phylesystem
19:57 jimallman we’ve talked before about scenarios where we might release synthesis (or taxonomy) twice in rapid succession. if it’s important to distinguish between these visually, we either need similar granularity here, or i’ll need to tweak to UI to handle this.
19:57 jimallman ie, “There were two releases in this week. Do you want {A} or {B}?”
19:58 kcranstn that seems like an edge case
19:59 jimallman hopefully, yes. and there is a stepwise navigation for each type of release: https://devtree.opentreeoflife.org/about/synthesis-release/2015-01-21T12Z
20:00 jimallman (that way, they definitely won’t miss a release)
20:01 kcranstn we can still have data points between the tick marks
20:02 mtholder re: point 6 of that email. I think we need to expand the set of static files that we generate with each release.
20:02 jimallman absolutely, it’s a continuum. the overall date range is set by the earliest and latest dates found
20:02 kcranstn +1 for what mtholder said
20:02 mtholder lots of content could be served directly from files generated with each release (see last comment on https://github.com/OpenTreeOfLife/treemachine/issues/127#issuecomment-70860291
20:03 jar286 I think we should flush the T12 business
20:03 jar286 at least in the UI
20:03 jimallman T12?
20:03 kcranstn I didn’t understand that, jar286
20:03 jar286 2015-01-21T12Z
20:03 jimallman ah, in the ISO timestrings
20:03 kcranstn +1 for that, too
20:03 jar286 in the back end it doesn’t matter, but you certainly don’t want users seeing that
20:04 jimallman gotcha. that was to deal with intra-day events (rapid releases).. see edge cases above
20:04 jimallman agreed that it’s ugly
20:04 jar286 yes I remember.
20:04 jar286 I was the one who suggested it.
20:04 jar286 maybe the whole 6-digit time should be in the filenames
20:05 jar286 by the way I think 2.7 was ephemeral. The april synthesis used 2.6.
20:05 kcranstn I was going to post a link to the OTT release page, but I see that the link to ott2.8 is not what I expected on http://files.opentreeoflife.org/ott/
20:06 kcranstn http://files.opentreeoflife.org/ott/ott2.8/
20:06 jar286 oh that’s wrong
20:06 jimallman T12 is kind of important in the URLs, to find and display the correct release. i’ll see if i can find a cosmetically nicer way to handle this.
20:07 jar286 there’s documentation in github…
20:07 jar286 I guess I need to run markdown and store the result on the server
20:08 jimallman fwiw, i can render markdown to HTML on the release pages, etc
20:09 jimallman inside the pages, i mean
20:11 jar286 ok. fixed now. whew.
20:11 jar286 http://files.opentreeoflife.org/ott/ott2.8/
20:12 jar286 I thought I had done that long ago.
20:12 kcranstn pmidford2 - is there a human readble description of the stats you are generating?
20:13 pmidford2 Somewhere, but I could reconstruct it somewhere - is there a wiki page I should use?
20:13 jimallman jar286: it’s interesting (to me, at least) that http://files.opentreeoflife.org/ott/ott2.8/index.html is quite different from http://files.opentreeoflife.org/ott/ott2.8/
20:13 kcranstn you can create one at https://github.com/opentreeoflife/opentree/wiki
20:14 jar286 clear browser cache
20:14 jimallman doh! thanks
20:14 pmidford2 kcranstn - will do
20:15 kcranstn thanks!
20:16 mtholder gotta run. ciao.
20:16 jar286 If anyone would like to make a style sheet for files.opentreeoflife.org, that would be just peachy
20:16 jimallman jar286: ok, i have a basic one we use for http://opentreeoflife.github.io/
20:17 jimallman we can probably adapt that one pretty easily
20:17 pmidford2 kcranstn - Apparently I started one already in germinator: https://github.com/OpenTreeOfLife/germinator/wiki/Overview-of-repository-statistics
20:18 kcranstn cool
20:18 kcranstn are there OTT stats, too?
20:19 jimallman not at the moment
20:19 jar286 jimallman, just point me at it and I can incorporate it
20:19 kcranstn https://github.com/OpenTreeOfLife/opentreeoflife.github.io/tree/master/css
20:20 jar286 tnx
20:20 kcranstn is there an open issue for OTT stats?
20:20 jimallman so quick!
20:20 jimallman no issue for this, it seemed premature. just the discussion in opentree-software
20:21 jar286 looks like all i have to o is copy the mini-logo into the file hierarchy
20:21 jar286 do
20:21 jimallman yes, and probably the css/ and images/ directories
20:22 kcranstn pmidford2 - are the stats scripts in that germinator repo?
20:23 pmidford2 The python calculation scripts are - I should add the scripts that push via rsync as well
20:23 kcranstn so the latest versions are from november?
20:23 pmidford2 Yes, these haven't changed
20:24 kcranstn what the broken piece preventing that chart from having real numbers?
20:26 pmidford2 We aren't getting stats from the production system because jar needs to do some ssh configuration, similar to what I did for dev last week
20:26 kcranstn but don’t most of the stats come from phylesystem?
20:27 kcranstn and OTT
20:27 pmidford2 yes, but they need to go to the web2py server on production (at least as the design currently stands)
20:27 jimallman both mtholder and i have code in place that adds fake data if the expected stuff is missing. i’ll need to review in depth to see where and why this is happening, but you can see an overview by clicking the red FAKE DATA link here: https://devtree.opentreeoflife.org/about/progress
20:28 kcranstn but the chart on dev doesn’t have data
20:28 kcranstn sorry, I am confused
20:28 pmidford2 There is data there - I reset it last week, but there is real data.
20:28 jimallman agreed, devtree has a *very* short history (6 days?)
20:29 kcranstn I have to run to a meeting. More later
20:29 jimallman ok, will try to gather some answers here
20:30 jimallman pmidford2: i’m looking at the stats files on devtree, and it’s just the last few days (data starts Jan 17). is that expected?
20:30 jimallman oh, i think you answered this (yes) above
20:31 kcranstn why are the stats linked to the history of the server?
20:31 pmidford2 If we want them I think I can get the previous data and try to merge it in
20:31 jimallman different phylesystem instances on tree vs. devtree?
20:32 jimallman (that was in reference to: Why are the stats linked to the history of the server?)
20:33 kcranstn why don’t the stats scripts simply use a local copy of the phylesystem repo
20:34 kcranstn ok, really leaving now
20:34 pmidford2 jar wanted the scripts to run on ot10
20:37 blackrim joined #opentreeoflife
20:42 jimallman back in a bit..
20:42 pmidford2 me too
21:17 kcranstn joined #opentreeoflife

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