Weekly check in 2012.01.05 - GeoSmartCity-CIP/gsc-opentripplanner GitHub Wiki
13:31 <demory> Hello everyone. Shall we do a quick check-in?
13:32 -!- mele [d819d289@gateway/web/freenode/ip.216.25.210.137] has joined #opentripplanner
13:32 <andrewbyrd> Hello
13:32 <demory> Hi
13:33 <mele> good morning!
13:33 <andrewbyrd> Good evening ;)
13:33 <demory> good afternoon!
13:33 <demory> we are a far flung bunch aren't we?
13:34 <demory> well why don't we get started
13:34 <grant_h> mornin'
13:34 <demory> I'll go first. Two main things I'm focusing on this week:
13:34 <demory> OTPsetup is now successfully building and deploying graphs; have done some successful testing w/ a larger system (Chicago). Currently working on a more permanent AWS architecture with elastic IPs, etc.
13:35 -!- pj_houser_ [617375b0@gateway/web/freenode/ip.97.115.117.176] has joined #opentripplanner
13:35 <demory> (Obviously we'll need a better way to address the machines, working w/ Evan at OP on that now)
13:36 <demory> The other thing I'm working on is the web-based visualizer (using Leaflet); should have something basic to show at TranspoCamp DC. Currently focused on visualizing graph internals (vertices & edges) and perhaps the graph-builder annotations as well.
13:36 <demory> that's it for me
13:37 <andrewbyrd> Just talked to Brian Ferris this afternoon about OTP/OBA integration and the multiple traverse result issue.
13:38 <andrewbyrd> We agreed that the need for multiple results is due to specific implementation decisions (refreshing the graph structure itself rather than objects within in - the trip pattern tables).
13:38 <andrewbyrd> There is no production use of OBA+OTP, and realtime OBA depends on an specific old release of OTP anyway.
13:39 <andrewbyrd> So the plan is to go ahead with encapsulating edge list modification, allowing dynamic generation of edge lists for extensions, but probably not even using it in future realtime-informed trip planning.
13:40 <andrewbyrd> Also delved into various solutions to the point location problem (which voronoi region does an arbitrary point fall in)
13:41 <andrewbyrd> Probably not worth the time investment right now to do that (allow arbitrary resampling on the fly for analytics) but can be done if needed
13:42 <demory> andrewbyrd is any of the analytics work to the point where I could incorporate it into the visualizer/front-end stuff I'm doing? say if I wanted to show basic travel-time isochrone generation?
13:43 <andrewbyrd> so for now just implementing travel time rasters as a WMS service via Jersey, cacheing pixel/road relationships in the assumption that you're not going to hit it with various coordinate systems in rapid succession
13:45 <andrewbyrd> demory: it does work now, just not as a standard web service. I could maybe get it stable enough for a demo this weekend if that's what you're thinking.
13:45 <andrewbyrd> but it depends on my branch of OTP which I didn't want to merge in until I talked to Brian about the OBA stuff
13:46 <demory> yeah, if there were something simple we could show in DC that would be great. Though I can always use the old isochrone generator I wrote if needed
13:46 <demory> just want people to see the direction we're going w/ this stuff
13:47 <andrewbyrd> The front end I'm using for testing is just your old isochrone generator with an image overlay layer instead of vector
13:47 <andrewbyrd> and when you move the origin point it updates the travel time overlay immediately
13:47 <andrewbyrd> well ~1sec lag
13:48 <andrewbyrd> oh DC is the 21st, yes we can have something ready.
13:49 <demory> yeah, we have a little over a week
13:49 <demory> even if we just show your modified version of the openlayers front-end that would be fine
13:50 <demory> though eventually i'd like to recreate that in leaflet
13:50 <demory> i think that's the direction I'm going w/ the visualizer
13:50 <andrewbyrd> If the WMS approach works out, you should just be able to throw on an openlayers WMS layer.
13:51 <demory> right
13:51 <demory> well let's follow up on this later
13:52 <andrewbyrd> sure
13:52 <demory> great. TriMet folks anything on your end?
13:53 <grant_h> I just have a minor issue based on some feedback
13:53 <grant_h> A few users have been confused by the collapsed legs of the itinerary
13:54 <grant_h> If we added a plus/minus sign when it collapsed/expanded I think that would help make it more intuitive
13:55 <demory> sounds simple enough. did you ticket it?
13:55 <demory> i can work on that but prob not until after the 21st
13:56 <grant_h> I'll write one up today
13:56 <demory> great
13:56 <demory> anything else?
13:57 <mele> Nope, things are looking great with OSMSplit
13:58 <pj_houser_> and i have nothing to add either
13:58 <demory> awesome, i'll pass that along to David (he's out this week so I don't think he's actually here)
14:00 <mele> Great, yeah it's having all sorts of other benefits by just making it easier for us to access current OSM data
14:01 <demory> yeah, i'm excited to hear how that's working out. keep us posted!
14:01 <andrewbyrd> demory, I can give you a call if you want to discuss analytics
14:02 <demory> sure, when is best for you?
14:02 <andrewbyrd> now is fine if you're free
14:03 <demory> yeah i'm free now. do you want to use the OP conference line?
14:03 <andrewbyrd> if anyone else wants to participate sure, otherwise I'll call you directly
14:04 <demory> anyone else interested in analytics?
14:04 <demory> if not I think we're done here -- have a good week everyone