IRC Logs for #cmvt Thursday, 2013-06-20

prologicwhat's with the 100%00:00
prologicand 5600:00
prologiclogos on the right?00:00
prologicoh I see00:00
prologiccool so local dev deploy worked00:01
prologiclemme get this on test bbs00:01
DanielBairdback.. so, yes those are the selection graph thingies.  they'll get text at some point.  And yes, if a vars endpoint sounds plausible to you, i'll add docs for it in the doc page00:16
prologicyeah it does00:19
prologicI mean if you need or think we need that kind of data00:19
prologicgo for it :)00:19
prologicI'm almost done here00:20
prologicwill start on this tool shortly00:20
prologicI've created an issue for it00:20
prologictest and prod have been updated00:42
prologicand I got a coffee00:42
prologicok let's have a look at your JS that produces this fake data00:43
prologiccan you walk me through this data that's needed? :)00:45
DanielBairdsorry, back again00:50
DanielBairdi'll dump some json, stand by00:51
DanielBairdactually that's a raw array, your response should have { "data": <<stuff here>> } or whatever, wrapped around it00:57
DanielBaird"state" can be whatever string you like, but probably it should be an id into the region api.  "region" is the region name, you might also want to add an "id" for a url-acceptable region identifier.01:00
prologickk lemme have  a alook01:01
prologiccool ok01:01
prologicwhat am I seeing here?01:01
prologicmax values?01:01
prologicI can see it's an intersection against states01:02
prologicor state regions01:02
DanielBairdno that's just the state the region is in.01:02
DanielBairdthe pastebin is an array of summaries, one summary for each region01:02
prologicoh ok01:02
prologicare the number just ids into a summary data?01:03
DanielBairdthe bioX values? that's what gets graphed, so it's your summary for that region for that bioclim.  I guess it should be a mean across the pixel vaues for all pixels in the region.01:04
prologicoh right01:04
prologicyeah that's what I was asking01:04
prologicso it's a mean01:04
prologictoo easy01:04
prologicand as we agreed this should be access via01:04
prologicGET /api/v1/summaries/ibra01:05
DanielBairdyes that right..01:05
DanielBairdwhatever that url was :)01:05
DanielBairdthe pastebin array, inside of { data: ARRAY }, is what you serve once the url has specified the region type (ibra), the year, the model, and the emission scenario.01:06
prologicGET /api/v1/summaries/projections/scenario/model/year/var01:06
DanielBairdnot var, in this case01:06
prologicyes sorry01:06
prologicyour example is all vars01:07
DanielBairdi guess the URL that includes the var would have just that var in each summary..01:07
prologicwrong uri :)01:07
prologicyup yup01:07
prologicthis should be interesting01:07
prologicI have to build a tool to precompute this stuff01:07
prologicthere is no way to compute the data fast enough dynamically I don't think01:07
prologicnot when (in testing) intersecting the largest region takes ~6s01:07
prologicso intersecting all regions will add up01:08
DanielBairdyou're probably right, across all vars especially01:08
DanielBairdin previous projects i've had academics to do the pre-computing of summary stuff like this.  that's much easier..01:08
DanielBairdalthough I do have Jeremy and his proteges01:09
prologicthat would be nice I guess01:09
prologicbut then I'd still have to get it in some structure01:09
prologicor format I can deal with pragmatically :)01:09
prologicbut it doesn't matter01:09
prologiccomputing the values isn't the hard part01:09
prologicserving it up in a sensible way is01:09
prologicok bbs let's see how this goes01:10
prologicwell right now I'm just going to concentrate on computing the data required01:24
prologicand serving as much as I can to your js code01:24
prologicthe rest I'm hoping will fall into place later on :001:24
prologicso I've committed some basic structure in01:28
prologicwith some sample (smaller subset) of data01:29
prologicgoing to test this01:29
prologicnot computing summary values/data yet01:29
prologicthat's next01:29
prologicoh I've got to add to the data tasks01:29
prologicdownload and extraction of region data01:29
prologicI'll just be using one for testing IBRA701:30
prologicI've uploaded sample (small subsets) of raster data to my web server in the US for testing/development01:30
DanielBairdah right, yes carry on sir01:33
prologicDanielBaird:  ping?04:35
prologicyour vm up-to-date?04:35
DanielBairdi haven't tried it for a while.. want me to give it a from-scratch test?04:36
prologicyes you'd better I think04:36
prologicwhat I want to ensure works for you is the basic shitty api stuff working for you04:36
prologicstill working on the actual summary generation though04:37
prologicbut I at least have a /api/v1 endpoint04:37
DanielBairdokay i'll kick it over. i think i've started a fresh build a few times, but never gotten back to testing it :)04:38
DanielBairdoh wait i suppose i should fetch first.04:39
prologicthat would be good :)04:51
prologicyou're doing a fresh vm?04:51
prologicrunning playbook first?04:51
prologicthen doing a remote deploy of the app?04:51
prologicfab should work04:52
prologicif your vm's ip is also
DanielBairdfresh vm.. i'm tryna run teh playbooks, have to dick about with ips04:55
prologiccan't you just make your vm's ip the same as mine? :)04:57
prologicso you don't have to dick about with ups :)04:57
DanielBairdyep changing it.. vm is restarting now into it's new ip04:57
prologicahh cool04:57
DanielBairdseems to be taking a while to start up04:57
prologicthen it should be a two step process :)04:57
prologicvagrant blah04:58
prologicand fab deploydev04:58
prologicthe playbook takes a while to complete04:58
prologicabout 10-15mins here I think04:58
DanielBairdscrew it i'll start over.. no idea why the restart was taking forever.04:58
prologichas to compile a bunch of stuff from source04:58
prologicVirtualbox + Vagrant, Proxmox VE and NECTAR04:59
prologicgood testing/development platforms all round :)04:59
DanielBairdansible-playbook site.yml ?05:00
DanielBairdoh right i have to do the sudo thing.05:01
prologicwait a minute :)05:01
prologicisn't this all setup via vagrant already? :)05:02
prologicI remember you were integrating the two :)05:02
prologicbut yeah depending on how your vm is setup05:02
DanielBairdi have a vagrantfile where i was trying to get it working, but haven't gotten back to it for ages.05:02
DanielBairddecided to do it manually this one last time05:02
prologicfor local I do ansible-playbook -u root site.yml05:03
prologicwhich defaults to ccav.local in hists05:03
prologicand ansible-playbook -s -u ec2-user -l jmills.nectar siteyml05:03
prologicfor nectar05:03
prologicI thought you had got that working :)05:03
DanielBairdi think i'm just about there, just not quite over the line05:04
prologicahh :)05:04
DanielBairdworking now. looking at my ansible vagrantfile, i think i'm just missing the username (which is 'vagrant').  i'm specifying the su-user as 'root' but i think i also need to specify the user it ssh-es in as05:06
DanielBairdi'll try that out of this current manual thing works05:06
DanielBairdof = if05:06
DanielBairdplaybooks are downloading python source.. we have to compile python from source??05:07
prologicyes we do05:07
prologicas it's all designed for CentOS 6.3 x86_6405:07
DanielBairdfricking centos05:07
prologicI completely agree :)05:07
prologicwhy they would not officially maintain a Python 2.7 or any other version in their upstream packages05:08
prologicis beyond me05:08
prologicI myself maintain python 2.5 2.6 2.7 3.1 3.2 3.3 and pypy for CRUX05:08
prologicit's one possible configuration management tool(s)05:08
prologiccould have used puppet05:08
prologicor any other05:09
prologicbut I quite like ansible's simplicity and flexibility05:09
prologicand minimal bootstrapping of the target nodes05:09
prologici.e: nothing05:09
DanielBairdyeah I agree, having to not install stuff on the remote nodes is pretty awesome05:09
DanielBairdnow it's installing pip05:10
prologicI'm gonna go get some coffee05:11
DanielBairdnp i'll post here when it's playbooked all the way to the end.05:11
DanielBairdplaybook all done..05:45
DanielBairdi should be able to fab stuff right?  fab not found05:48
prologicdo it from your local (on your desktop) repo05:56
prologiccd /path/to/ccav05:56
prologicfab deploy.dev05:56
prologicfab -u <user> deploy.dev05:56
DanielBaird(ui-ccav)C02JKVMWDR53:ui-ccav pvrdwb$ fab -u vagrant deploy.dev06:19
DanielBairdDeploying dev instance to ... [] Login password for 'vagrant':06:19
DanielBairdFatal error: put() encountered an exception while uploading '<StringIO.StringIO instance at 0x105d0da28>'06:19
DanielBairdUnderlying exception:06:19
DanielBaird    Permission denied06:19
DanielBairdDisconnecting from done.06:19
DanielBairdtook ages to get to that point06:19
DanielBairdwhen i run it again, I get the same error (different object id) immediately.06:22
prologicand can you ssh in as the vagrant user?06:37
prologicyou probably don't want to do that though I think06:38
prologicwhat the playbook sets up is a web apps account06:38
prologicand several developer/user accounts06:38
prologicin fact throw me your ssh key06:38
prologicI'll update the playbook with your key06:39
prologicwhen deploying you should only need to do so with your own user account06:39
prologicyou basically get added to the webdevs group and get some sudo rights, etc06:39
prologicwe'll sort it out in the morning06:39
prologicwife's here to pick me up06:39
prologiccya then06:39
DanielBairdnp see you06:39
*** DanielBaird has quit IRC06:40
*** irclogger_ has joined #cmvt23:27

Generated by 2.11.0 by Marius Gedminas - find it at!