IRC log of #schooltool for Monday, 2012-08-06

*** menesis has quit IRC01:01
*** yvl has joined #schooltool09:28
*** menesis has joined #schooltool11:19
*** yvl has quit IRC12:04
*** yvl has joined #schooltool12:05
*** th1a has joined #schooltool15:44
*** aelkner has joined #schooltool15:55
*** replaceafill has joined #schooltool16:03
th1ahi replaceafill, aelkner, menesis, yvl.16:31
yvlgood morning16:31
replaceafillgood morning/afternoon16:31
th1aOK... I want to start by talking a little about charts in the gradebook(s).16:32
th1aIncluding CanDo.16:32
th1areplaceafill:  I hope you got the email I sent Friday evening about the "Net Stacked Distribution" charts.16:33
replaceafillth1a, yes16:33
replaceafilli have a question though16:33
th1aLet me catch everyone else up.16:33
th1aThe third one is the one we're talking about.16:34
th1aEach row would be a student.16:34
replaceafillreally nice approach16:34
th1aEach color area a grade.16:34
th1aThe "0" point in the middle the selected passing score.16:34
replaceafillth1a, if i understood correctly the passing score wouldn't show, correct?16:34
replaceafillit's the line16:35
replaceafillin the example, the line is the neutral response16:35
th1aPassing is above the line.16:35
yvlreplaceafill is right16:35
th1aWell, he's right about the example, but no, we wouldn't do it that way.16:35
yvlright :)16:35
th1areplaceafill read the example more carefully than me.16:36
replaceafillth1a, i think you want the passing score showing as a "negative" score, right?16:36
replaceafillor possitive?16:36
yvlbecause passing16:36
replaceafilldamn, i got confused by the directions again ;)16:36
th1aThere is really no "neutral" score.16:36
th1aThere is no "neither passing nor failing" score.16:37
th1aExcept "no score"16:37
th1aActually, for this example, "neither agree or disagree" is equivalent to "no score"16:37
replaceafillgot it16:38
replaceafillth1a, ok, i think i can have it done today16:38
replaceafillmy only question was that16:38
replaceafill(about not showing the passing score)16:38
replaceafilli cleaned the view to use ajax tables16:39
th1aThis also should make sense as a view in the gradebook,16:39
replaceafilland it now has sortable columns16:39
th1aalthough not as much...16:39
*** jelkner has joined #schooltool16:39
th1aAh, we can ask jelkner.16:39
jelknerthat's why he's here?16:39
jelknerask him what?16:40
th1aWe're going to try this style graph for the section competency report:16:40
jelknerreplaceafill, th1a, you saw my email, yes?16:40
th1aYes, we should be able to do that.16:40
replaceafilljelkner, i did16:40
th1aWe needed something to prod us to sign all the paperwork.16:40
jelknerthere you go16:41
jelkneri'm hoping to help prod16:41
th1aI guess this makes less sense in a point-based gradebook since activities are weighted differently.16:42
th1aSo jelkner, this would be showing the number of skills at each score level.16:42
jelknerthat rocks!16:42
th1aAbove and below the selected passing view.16:42
th1apassing score.16:42
jelkneri have a student with a question16:42
jelknerbut i love it16:42
jelknerbecause it would help me get a quick overview of which skills were lacking in the section as a whole16:43
jelknerwhich can then help guide instruction16:43
jelknervery useful info16:43
th1aYou mean if each row was a skill.16:43
th1aWe can do that AND if each row is a student.16:44
replaceafillby skills report16:44
th1aSo one thing is that this should be of all relevant skills, not just the scored ones.16:46
th1aSo if someone doesn't have many scored, the bars should be short.16:47
replaceafillyou get the overall number from the width of the "total" bar (according to the article)16:47
*** ignas has joined #schooltool16:47
th1aAlso, some random useful links I found regarding chart colors:16:47
th1aBasically, auto-generating a range of color shades.16:48
th1aThe math.16:48
th1aWe'll probably settle on a precise method, but it is something worth knowing about.16:49
jelknerthat's what i thought you were talking about16:50
jelknereach row a skill16:50
th1aAlso, I have a meeting with Schwa tomorrow morning to talk about the School page redesign and a couple other of the UX issues CL identified.16:50
th1aEach row a kid makes sense too.16:51
th1aThat's what we were thinking of.16:51
th1aI'm not sure what the timeframe will be for the design work.16:52
th1aIf they're ready to roll we might be able to jam it in, since it won't be a deep code change.16:52
th1aI'll have a better idea after the meeting tomorrow.16:52
th1amenesis: ayt?16:53
menesisI was not able to complete the packaging of cando packages on friday16:55
menesisI only really started that day16:55
th1aI think I'll just change my schooltool contact email to the new address, but in the meantime, can you try restarting sendmail on, or something?16:55
menesisI can try16:55
menesisbut what really needs to be done is you choose a hosting provider16:55
th1aAnd if that works, maybe just a cron job to do that daily as a band-aid until we have time to deal with it correctly.16:56
menesisbuy a virtual server16:56
th1aYes, but I want you to get the packaging done first.16:56
menesisand then I can move content there.16:56
menesisof course16:56
menesiswhat is the problem with your email?16:57
th1aIt is not getting through.16:57
menesismessages are not accepted by mail server?16:57
th1aWhen the server restarted last week a glut came through.16:57
th1aSo actually they may not even be queuing.16:57
menesisafter the restart, I am now getting mail from root@schooltool.org16:59
menesisbut what is sent to it.. haven't tried17:00
th1aDid you just restart sendmail?17:00
menesiswill test17:00
th1aI don't send from it.17:00
th1aWell, I forward from it.17:00
menesisno, I haven't restarted yet17:00
th1aOK, perhaps we can discuss this later.17:02
th1aSo packaging... you're started on it now?17:02
menesisyes, I started on Friday17:02
menesisbut had a shorter day..17:03
menesisthere are 5 packages to release, two of them new17:03
menesisI will cut some corners to get something out sooner17:03
th1aOh... I just forwarded a postfix error email I just got from the server.17:04
th1aPerhaps that's the problem.17:04
menesise.g. not merge translations17:04
th1aThat's fine.17:04
*** dreich has joined #schooltool17:05
menesisi don't want to promise all needed debs are uploaded today.17:06
menesisbut am working on it17:06
th1aOK.  Keep moving forward and we should be fine.17:06
th1aDo that before doing anything with the server other than kicking it a few times.17:06
menesisI will avoid doing anything else.17:07
th1aThanks menesis.17:08
th1aLet me know when these are ready.17:08
yvlnothing exciting to report actually17:09
yvlhammering through :)17:09
yvlexperimented with stuff we talked about last Wed17:10
yvlpolished the importer view a bit17:10
yvland some minor stuff17:10
yvllike adding link to tasks under server17:10
yvlonly showing links to result task server if devmode is enabled17:11
yvlplumbing :D17:11
th1aWhat's your vacation schedule at this point?17:12
replaceafillyvl can you help me with a template question :)17:12
replaceafilloh sorry17:12
yvlth1a, all next week17:12
th1aOK, so work on this this week -- get it to a good stopping point, and then ldap when you get back?17:13
yvlthat's the plan17:13
th1aSo should this machinery help with, say, syncing with Moodle?17:14
yvlwell, at least it should be used to sync with external sources17:14
th1aDoes it have built in cron-like functionality? (not that cron is hard)17:15
yvlit has both queuing and scheduling17:15
th1aI was thinking that (while reassuring myself we weren't wasting too much time on this...).17:16
th1areplaceafill:  go ahead with your question.17:18
replaceafilli'm creating a table column17:18
replaceafilland i think its rendering it's going to require some logic17:19
replaceafillso i tried to use a actual template for it17:19
replaceafillbut the viewpagetemplatefile class requires (context, request)17:19
replaceafillis there any way to get the same behaviour without this hack?17:20
* yvl is looking17:21
yvlplease go on with report for now17:22
replaceafilli really don't want to use a string for the template17:22
replaceafillth1a, last question about the charts17:22
replaceafillthey should be centered like in the example, correct?17:22
replaceafilli mean, centered in the vertical line17:22
replaceafillwe probably need more space again?17:23
replaceafillmy only concern at this point is the centering part :)17:23
th1aLet's start byassuming yes, you need 100% in both directions.17:23
th1aAnd go from there.17:23
replaceafillth1a, kk17:23
* replaceafill done17:23
th1aAlso, I think we should more away from worrying about tooltips in this one.17:24
th1aAnd just switch to a "click and you get a detailed popup" paradigm.17:24
th1aWhich for starters can just mean click on the student name.17:24
th1aThat is probably sufficient overall anyhow.17:25
replaceafilli thought you meant clicks the rect's ;)17:25
th1aDefinitely not on the individual ones.17:25
replaceafillgot it17:25
th1aI don't think that headache is necessary in this view.17:25
th1aThanks yvl, replaceafill.17:25
yvlone sec17:26
yvlreplaceafill, I think that "hack" is very acceptable17:26
yvlas in not a hack :)17:26
yvlyou could pass17:26
replaceafillan acceptable hack ;)17:26
yvlself.template(context=item, request=formatter.request, view=self.view)17:26
yvlfor the sake of some consistency17:26
yvland... consider using plain strings17:27
replaceafillbut it still would require the hack, right?17:27
replaceafillah got it17:27
yvland consider using plain simple strings ;)17:28
replaceafillhhmm sure, i think i can get away from the logic in the template17:28
replaceafilland put it in the class17:28
yvlthat crap is not designed for templates17:28
replaceafillhey, i worked really hard in my column!!!17:28
yvland the logic is better of in "view" than template anyway ;)17:28
replaceafillthanks yvl17:28
yvlthe column is nice, the rest is crap x)17:28
yvlok, let's give the mic to aelkner17:29
aelknerok, so i started to change to be runnable with or without download17:29
aelkneri encountered problems getting the download part to work however17:30
replaceafillaelkner, /occupations?17:30
aelknerhow did you know? :)17:30
replaceafilli hit the same wall17:30
aelkneri merged what i did to trunk, so we could look at the diff together17:31
replaceafilli think the API may have changed since yvl wrote the original version17:31
aelknerfor starters, i needed to get rid of the load call in __init__ of the DB class17:31
aelknerreplaceafill, ah!17:31
aelkneri needed to move the LOADED = DB('get-verso') up higher and move the update_class call to the bottom17:33
aelknerso the code is set up to work with or without the --no-download command line option17:33
aelknershould it be --no-download or --no_download, btw?17:34
aelknerthe if __name__ == '__main__' used to have all of this logic that i moved to loadAll17:35
aelknerthere were a lot of commented lines of code for some reason, so i cleaned that stuff up17:36
aelkneranyway, th1a, if you want me to, i can keep plugging away at figuring out what went wrong with /occupations17:37
aelkneri was also hoping to catch up with what you guys have been doing with reports17:37
aelknerand hopefully get a task in that area17:37
th1aWell, a) ask Lee Capps.17:37
th1ab) we don't even want occupations anyhow, right?17:38
aelkneryes, i was thinking about scrapping it17:38
aelknerthat would be the easier thing to do, and quicker than waiting for responses from Lee17:38
th1aYes, but you should send him an email right now anyhow if you think the API might have changed.17:39
th1aWe don't need to play reverse engineering games.17:39
th1aIt's not like we're trying to re-implement Active Directory here or something.17:39
th1aThere's not really much to say about the reports other than what we already said about the charts.17:41
yvloh, they did change the format17:42
yvla little bit17:42
aelknerif we don't need occupations, perhaps the thing to do would be to get rid of that class17:42
aelknerthe download will be faster and we wouldn't have to figure out what changed17:42
aelkneryvl, i don't know if you heard:17:42
aelknerthey decided they didn't need the whole hierarchy17:43
yvlI didn't follow that17:43
aelknerjust cluster, course, dutyarea (now called skillset) and Task (skill)17:44
aelknerso i'm thinking i should just excise the other classes, speeding up the download17:44
aelknerand saving us the trouble of asking Lee about the api change17:44
yvlno need to ask btw17:44
aelkneralso, the download will be faster which will be good for each site17:45
yvlit's way, way quicker just to implement17:45
aelknerwell, easier is a matter of opinion17:45
aelknerperhaps for you, it would be a quick change17:45
aelknerand if you think so, could you do it real quick for me?17:46
aelknerotherwise, what about the idea of getting rid of it and speeding the download?17:46
yvlI'm fine with either option17:46
th1aWhat's the change in the format yvl?17:46
yvlthey reorganized something17:47
yvlI think instead of having one bag of occupations, they have a list of clusters/pathways/somethings and lists of occupations under them17:47
yvlI don't know if they retained the cluster information in the occupation itself17:47
yvlif so, then it's 10 min change17:47
yvlif not, 30 min change17:48
aelknerth1a, what do you think about just getting rid of the pathway and occupation download?17:48
replaceafillimho we should download all that they have (if it's easy)17:48
th1aI'm ok with it.17:48
th1aI'm not sure what yvl is looking at that you aren't though.17:49
replaceafilli can hear the "let's put the structure back" comment...17:49
replaceafilljust my 0.0217:49
replaceafillit's easier to have it all and get rid of stuff at the xls generation level17:49
aelknerthat sounds reasonable17:49
th1aWe'll need it eventually for some reason.17:50
th1aOTOH, in the short run we'll probably use the quicker download anyhow.17:50
yvlI'm not 100% sure, but I think I fixed it :D17:50
yvlumm, where's the "trunk" now?17:50
* yvl can't recall :/17:51
aelknerno, thank you for helping with this17:51
aelknerwe don't need the atexit stuff anymore, do we?17:53
yvlmy pleasure, my duty :)17:53
yvlwell, atexit was used to store whatever is downloaded17:53
aelknerbut it's commented out17:54
yvlin both cases where download finished17:54
yvlor when download breaks17:54
yvlwell, now it does not break, in theory :)17:55
yvlso after loadAll it just saves17:55
yvlwhen I was writing this script, it broke often17:55
aelkneryes, i did remove the commenting out of the save and backup calls17:55
aelknerwhy did it break so often?17:55
aelknerwas the verso ste unreliable?17:55
yvlscript was incomplete17:56
aelknerah, ok, so if we have it polished, then we don't need the atexit17:56
yvlI just dumped whatever was there, instead of looking through the browser17:56
aelknerth1a, could we have yvl bring the script to the finish line since he would be much faster than i17:57
yvl(it just s*cks so bad that aelkner wants to have nothing to do with it :D )17:58
aelkneri didn't grok the json part to write build_xls, i just looked at the pickel file17:58
aelknerso if yvl and th1a had no objections,  would love to gently hand it back17:59
yvlI'm busy!!! :))))))))))17:59
aelknerok, so i guess that's a no18:00
th1aI think you can handle it aelkner.18:00
yvlIs there much left with it?18:01
aelknerthat's fine, but i'm, not promising that it will take 30 minutes18:01
yvlwell, for me it was easy to fix occupations (and I just committed)18:01
yvlbut what to do next...18:01
* yvl shrugs18:01
aelknerso perhaps it's already done18:01
aelknerbtw, why did you say it takes hours to run?18:02
aelknerclusters and pathways was 30 seconds combines18:02
aelkneroccupations took 30 seconds to download the json before it crashed18:02
aelkneris it courses that took so long?18:02
aelkneroh, probably DutyAreas18:03
replaceafillit gets 6k skills18:03
yvland dury areas18:03
aelknerah, ok18:03
yvlaelkner, btw, if you don't call LOADED.load() before loadAll()18:03
yvlit will load from scratch18:03
yvlas in - reload everything18:03
yvlnot sure if that was the intent18:03
aelknerLOADED.load() assumes the verso-db pickel file exists18:04
aelknerwhich, ofc, it doesn't before the first successful download18:04
aelknerare you saying the script would run not from scratch?18:05
yvlreturns False on failure18:05
yvlit would run from pickle if available18:05
yvlfrom scratch if not18:05
aelkneroh, i see, so that's why you had the load call in __init__18:06
aelkneri guess i got spooked by the 'Failed to load' error message18:06
aelkneri don't know what we want the user to see18:06
yvldirty hacky script :|18:06
aelknerdo they even need the printing of statistics18:06
yvlI don't think so18:07
*** jelkner has quit IRC18:07
aelknerso we could get rid of the printing of stats and the load fail error message18:07
yvlor better yet18:08
yvladd quiet=False to DB.load method18:08
yvland pass LOADED.load(quiet=True) in the script18:08
aelknerand call it with True, right?18:08
aelknerand remove the call to printStatistics but keep the method around in case18:09
aelknerok, will do18:09
yvl(that method kind of shows what we expect the data to be)18:09
aelknerok, so in the meantime, as i will be waiting for long downloads18:10
aelknerth1a, can we start discussing my next task?18:10
aelkneri could work on a chart if there is one already done that i could use as a pattern18:10
th1aWell, that'll be a lot clearer in a day or two.18:11
th1aAre you in Arlington now?18:11
yvlaelkner, just bzr pull before downloading (I just made another commit ;) )18:11
aelkneryvl, thanks18:11
aelkneri'm here in arlington, but i will be able to work18:12
aelkneris it easier for you to wait until wed to bring me into the chart tasks (or pdf)18:13
th1aOr tomorrow at least.18:14
aelknerok, i'll look for you tomorrow then18:14
aelknerthat's it for me18:14
th1aYou should probably start looking into drawing charts in ReportLab though.18:14
th1aThat is, drawing rectangles.18:14
th1aAnd lines...18:14
th1aOK.  That should do it for now.  Let me know if you're sitting around looking for something to do aelkner.18:16
yvl(I thought I missed a bag of gravel somewhere)18:17
aelknerok th1a, will do18:17
* th1a drops the bag of gravel.18:19
yvlIn other news - congratulations to you in US on successful landing Curiosity. I'm off to a pub called "Mars" for a burger - seems appropriate ;)18:19
th1aEnjoy, yvl.18:19
th1aThanks guys.18:19
yvland see you all Wed!18:20
replaceafillthanks everybody18:20
aelknerthanks guys18:24
replaceafillth1a, should we use ticks in the new layout?18:29
replaceafillmaybe ticks in both directions?18:29
th1aI'd say yes.18:30
replaceafillth1a, zyt?19:26
th1ahi replaceafill19:30
replaceafillth1a, welcome to last friday :D19:30
replaceafilldone it the right way ;)19:30
replaceafilli can position rect's now19:30
replaceafilli also moved the filters19:31
replaceafilllet me know what you think (about the filters)19:31
th1aWe need sexy transforms.  ;-)19:31
th1aOtherwise, we're on the right track.19:31
th1aCan't we do sexy transforms?19:32
replaceafilli think we can19:32
replaceafilli just have to dig deeper in the docs :)19:32
th1aDon't worry about it now.19:32
replaceafillanyway, last part is to split the chart and add the non-passing scores19:32
replaceafilli think i know how to do it19:33
th1aYou'll need to reverse the order of the scores.19:33
replaceafillleft - right19:33
th1aPrototyping with grayscale is probably wise.19:33
replaceafilli just used default opacities19:34
replaceafillwill add the formulas later19:34
th1aOh, ok.19:34
replaceafillth1a, i need math help :D20:33
replaceafilli'm using transforms to move the svg container trying to align it20:33
replaceafillbut i can't get it right :(20:33
th1aDon't get hung up on it now.20:34
replaceafillhhmm is that the main point of this type of chart?20:34
replaceafillthe centering?20:34
th1aWhat's that showing me?20:38
replaceafillthat i haven't aligned them right!20:39
replaceafillall the scores20:39
th1aDid you look at the d3 examples for somthing similar?20:39
th1aI was surpised I didn't see one of these already done.20:39
replaceafillyes, but i couldnt find anything either20:40
*** ignas has quit IRC20:40
replaceafilli know it's the transformation math20:40
replaceafilli'll go get a drink and continue in 1020:40
th1areplaceafill:  Ignore Welsh's instructions about the report text.  ;-)20:44
*** dreich has quit IRC20:46
replaceafillth1a, kk20:47
replaceafillth1a, is this what we want?20:47
replaceafill(gimp btw) ;)20:47
replaceafilli'm just trying to "see" it20:48
replaceafillwith the right numbers20:48
*** yvl has quit IRC21:36
*** yvl has joined #schooltool21:37
*** th1a_ has joined #schooltool21:38
*** th1a has quit IRC21:39
*** aelkner has quit IRC21:41
*** aelkner has joined #schooltool22:04

Generated by 2.15.1 by Marius Gedminas - find it at!