2012.10.25 Weekly Check In

• 13:31 <demory> hello everyone
• 13:31 <demory> shall we get started?
• 13:32 -!- janinge [[email protected]] has quit [Read error: Connection reset by peer]
• 13:33 <novalis_dt> sure
• 13:33 <demory> ok
• 13:33 -!- janinge [[email protected]] has joined #opentripplanner
• 13:33 <demory> my update: mainly working through various graph-specific issues with the North American API endpoint. Also wrapping up the Analyst IDEA grant app
• 13:33 <novalis_dt> I've been working on misc OTP bugs. I wanted to get working on that bike bug from Portland, but when I went to compare routes, TriMet's test instances were down. I haven't checked if they're back up yet, since other bugs have come in
• 13:35 -!- janinge [[email protected]] has quit [Read error: Connection reset by peer]
• 13:35 -!- janinge [~[email protected]] has joined #opentripplanner
• 13:35 <novalis_dt> Matt Conway, who isn't here right now, also merged his cars branch
• 13:36 <novalis_dt> Which is a great start.
• 13:36 <novalis_dt> I made some updates related to that.
• 13:36 <eshon> kpw & angle: will there be a basic geocoded for manila gtfs? http://nominatim.openstreetmap.org/search.php?q=central+terminal%2C+manila&viewbox=120.88%2C14.66%2C121.09%2C14.53
• 13:36 <eshon> aogle
• 13:36 <eshon> geocoder*
• 13:38 <novalis_dt> abyrd, what have you been up to?
• 13:38 <kpw_> eshon, for the gtfs?
• 13:39 <eshon> yeah for the UI
• 13:39 <kpw_> in the editor? no i don't think there's a role for a geocoder as i understand it
• 13:40 <abyrd> making the way analyst gets results from spts more modular
• 13:40 <eshon> i'm worried about their comment that ppl aren't used to maps there… ok metro manila is pretty big, so many pieced together cities
• 13:41 <demory> novalis_dt, regarding your raptor question earlier: so, all of the graphs on the endpoint were built w/ raptor enabled, but i forgot the part in app-context on the main OTP server
• 13:41 <demory> so i guess that means raptor is not currently enabled
• 13:41 <demory> but that's an easy fix
• 13:41 <novalis_dt> demory, OK, thanks for letting me know
• 13:42 <demory> it does require rebooting tomcat though. abyrd, will 0.9 automatically load all of the graphs in /var/otp/graphs? or do I have to do that manually?
• 13:43 <abyrd> so we don't have to modify the source code to change between different result types, or to switch between removing initial wait and ignoring it
• 13:44 <abyrd> demory, you can give it a list of graphs to auto-register
• 13:44 <abyrd> in the spring configuration
• 13:45 -!- jcassee [~[email protected]] has joined #opentripplanner
• 13:46 <demory> abyrd, ah, yes. (should've checked docs before asking)
• 13:47 <demory> kpw, new NYC graph is building now
• 13:47 <demory> (w/ tranfers fix enabled)
• 13:47 <abyrd> demory, do you need auto-load-everything functionality?
• 13:48 <demory> abyrd, that would be helpful for times like this
• 13:48 <demory> where you want to restart an entire instance with 50 or so graphs loaded
• 13:48 <demory> i'll probably end up writing a script to generate the config based on the directory contents
• 13:49 <abyrd> so there's no other component of the system that knows which graphs have been placed on that particular server?
• 13:50 <abyrd> It is of course possible to just scan the directory names, but we are using a general resource name instead of a file name as the base path
• 13:50 <demory> well, there is actually, but on a different instance
• 13:50 <abyrd> so it's messy to handle both cases
• 13:51 <abyrd> though the only reason we're using a resource instead of a filesystem path is that someone wanted to run OTP on a hosting service that required everything to be inside the WAR
• 13:51 <abyrd> which is not a typical way to work with OTP, I'd say
• 13:52 <demory> i might as well write a "restart tomcat" action into the Deployer workflow, which passes over the list of graphs -- probably will come in handy again
• 13:52 <abyrd> I could just make a separate simple resource graph service that will load graphs from inside a WAR
• 13:52 <demory> so don't worry about changing anything in OTP for this
• 13:53 <kpw_> demory, great! thanks
• 13:55 <demory> FrankP, anything on your end?
• 13:55 <FrankP> Hey David...yeah, last week, I see there was mention of a mobile ui...not immediately interested in that (although it's where I'd like to take the trimet map ui eventually). I'm working to replace our current text trip planner, and all the ancillary tools around that.
• 13:56 <FrankP> I've started to put stuff here. http://opentransittools.com
• 13:57 <FrankP> System requirement on this end is python / pyramid. Will work with OTP and gtfsdb.
• 13:57 <demory> FrankP, right. let's chat about the UI. we're planning to resume work on the Leaflet based UI more intensely next month, hopefully having something ready for a 1.0 release
• 13:58 <FrankP> Goal is May 2013 to be able to turn off current system for front-facing web stuff. Other backend stuff needed.
• 13:58 <demory> we should talk about how we can share common elements between the OpenLayers webapp and Leaflet
• 13:59 <demory> e.g. the bike triangle, the elevation renderer
• 13:59 <demory> perhaps some of the narrative generation code
• 14:00 <demory> at some point soon i will probably create an "opentripplanner-webapp-leaflet" module and merge in the cibi.me/bikeplanner.org code
• 14:01 <demory> mayeb we also have an "opentripplanner-webapp-common" module for those shared JS bits
• 14:01 <demory> does that sound ok?
• 14:02 <FrankP> Yeah, that sounds good David. Might be able to use mvn to package up the common stuff...for TriMet map, I simply used the compressed .js output from mvn. I could include two...
• 14:03 <FrankP> One piece of the current .js code that I need for the text planner effort is the translations (En/Es) ... we'd talked years back about Portable Objects, and I'll work on that in my project. But maybe the .js code might want to move to .po objects too.
• 14:04 <novalis_dt> That would be awesome
• 14:04 <FrankP> https://github.com/openplans/OpenTripPlanner/issues/261
• 14:05 <demory> ok. we haven't really tackled localization at all in the Leaflet UI
• 14:06 <demory> but I'll assume the PO approach going forward
• 14:09 <demory> alright, anything else for the check-in?
• 14:09 <novalis_dt> nope
• 14:09 <FrankP> Also looking forward to a refactor of the entire .js codebase. backbone.js is something all the kids are talking about these days, so maybe a good rethink is in order. (That said, I'm back to template generated html till May...with maybe just a bit of jQuery sprinkled in)