Weekly check in 2012.04.05

• 13:34 <demory> hey everyone, should we go ahead and do our check in?
• 13:35 <novalis_dt> sure
• 13:35 <novalis_dt> I've been working on some area routing code, as well as misc bug fixes
• 13:36 <demory> novalist_dt what's the status of #670?
• 13:37 <demory> ^ novalis_dt sorry
• 13:37 <novalis_dt> I did fix what I'm pretty sure was a genuine bug
• 13:37 <novalis_dt> in the NED interpolation
• 13:38 <novalis_dt> however, that bug could not have been directly responsible for this crash
• 13:38 <novalis_dt> Andrew is investigating the actual issue.
• 13:38 <demory> ok
• 13:38 <novalis_dt> The bug fix I made did mask the bug.
• 13:38 <demory> so the graphs do build succesfully now?
• 13:39 <novalis_dt> Seattle does
• 13:39 <andrewbyrd> I've built the SF one several times with NED and it did not fail
• 13:40 <andrewbyrd> just switched to edge sets instead of edge lists which should help prevent or detect causes of orphan edges
• 13:40 <demory> ok. so are you hoping to fix the actual issue before cutting the 0.5.4 release?
• 13:40 <andrewbyrd> plus now edgelists are threadsafe
• 13:41 <novalis_dt> Cool.
• 13:41 <andrewbyrd> I would hope so, but need to rerun that build now that I'm using identity equality for edges and see what happens
• 13:41 <qlex> on a sidenote, is elevation data only available for the US or is there any alternatives for Europe? andrewbyrd: maybe you know any ?
• 13:42 <novalis_dt> There's SRTM, globally, but it is low-resolution
• 13:42 <andrewbyrd> qlex, I don't know of a DEM for Europe since I haven't worked much with elevation data outside the US
• 13:42 <andrewbyrd> looks like the EEA has one
• 13:42 <demory> andrewbyrd, ok. a new release is somewhat less urgent now that i'm using the task-order workaround w/ 0.5.3 on deployer
• 13:43 <grant_h> qlex: NASA has free 30 meter DEMs for the whole planet
• 13:43 <qlex> grant_h: usable with otp or require some kind of loader ?
• 13:43 <novalis_dt> That is probably too low resolution to use
• 13:43 <kpw> grant, qlex, not sure if 30m is sufficient for transit, but it would be great to know if there are other higher res dems available outside of the us
• 13:44 <kpw> my sense is that data is hard to come by via publically available sources
• 13:44 <kpw> but it would be a great research project if someone is up to it
• 13:45 <kpw> for refence we're using a 10m dem in otp--so that or better would be the goal
• 13:46 <demory> My update: dealing mostly w/ Deployer workflow/debugging issues earlier this week. Currently focused on the new OTP community website which we're getting ready to relaunch.
• 13:46 <demory> Also, have been setting up some regional, multi-feed deployments for major US cities to serve as a showcase for Deployer... A few are ready for use now:
• 13:46 <demory> http://nyc.deployer.opentripplanner.org/
• 13:46 <demory> http://sfbay.deployer.opentripplanner.org/
• 13:46 <demory> http://dc.deployer.opentripplanner.org/
• 13:46 <demory> One for Chicago should be up later this afternoon.
• 13:47 <novalis_dt> Cool!
• 13:48 <kpw> nice work!
• 13:49 <demory> yeah, the ny and sf ones especially are both large/complex builds and OTP seems to be holding up very well
• 13:49 <novalis_dt> TriMet, what's new with you?
• 13:49 <kpw> used your DC demo at a talk last night. the bike trangle drew a huge "oooh!" from the crowd... :)
• 13:49 <andrewbyrd> I fixed a few routing/itinerary building bugs that were causing truncated walk legs, and causing unnecessary branching in MOA*
• 13:49 <mele> I was doing some wiki page editing while OSM was down, now back to accessibility edits
• 13:49 <grant_h> qlex: It might actually be higher than 30 m, it's worth checking out, as far as compatiblilty with otp, I don't know what that takes but one that I used recently for the Netherlands came in .tif format, here's the link: http://reverb.echo.nasa.gov/reverb/#utf8=%E2%9C%93&spatial_map=satellite&spatial_type=rectangle
• 13:50 <mele> I added a large section on OSM tagging to http://wiki.openstreetmap.org/wiki/OpenTripPlanner
• 13:50 <demory> great!
• 13:50 <mele> and we're in the first stages of setting up an automated data testing suite
• 13:51 <mele> but that's pretty much just for us :)
• 13:51 <grant_h> The redaction from the license change to OSM have started too
• 13:51 <andrewbyrd> now just working on this edge list coherency issue. also added a bounding box check to UnifiedGridCoverage to reduce the exception overhead.
• 13:51 <grant_h> Shouldn't effect us too much though
• 13:53 <kpw> oh! what's up with the data testing suite?
• 13:53 <novalis_dt> oh, that reminds me that next time I'm in chinatown I should replace the redacted street there. I wonder if there's a good Android app to help me do that
• 13:53 <FrankP> Was off last week...trying to get in the swing of things ... initial release of https://github.com/fpurcell/SOLRGeoCoder (just my current SOLR config files ... hard part is populating SOLR)
• 13:54 <mele> we're just pulling together some trips in a CSV format that Frank can feed into OTP automatically, and hopefully we can get a visual compare going as well as notify us if it doesn't find strings that it's expecting
• 13:54 <novalis_dt> FrankP, do you have scripts for populating solr?
• 13:54 <FrankP> BTW Dave -- has OBA been fixed recently? Just wondering the status of the DST bug - https://github.com/openplans/OpenTripPlanner/issues/634
• 13:54 <novalis_dt> FrankP, nope
• 13:54 <novalis_dt> FrankP, we're still waiting on a release
• 13:55 <novalis_dt> As soon as that release happens, we'll fix things. And in the unlikely event that it doesn't happen by (say) September, we'll fork it
• 13:55 <FrankP> thanks (been holding a gtfs for DST to test)...yes scripts, but the ping our database, so of limited use...also in Groovy (making them more limited) ... some are checked into current github repo
• 13:56 <FrankP> As time permits, wouldj like to rewrite the loader with GTFSdb and python to make them more universal.
• 13:57 <FrankP> Posted them this week because Andrea Anime was asking.
• 13:57 <novalis_dt> We should definitely investigate these!
• 13:59 <grant_h> novalis_dt: w/r/t osm apps this is the only one I know of that allows mapping of ways: http://wiki.openstreetmap.org/wiki/Vespucci , I know there are a couple students are working to improve this one for GSoC
• 14:00 <FrankP> demory,looking at the NYC deployment ... very cool. One thing to possibly add to config.js is turn "on" Agency, so that appears in the itinerary.
• 14:00 <novalis_dt> thanks, I'll look into that
• 14:00 <demory> FrankP, good call. I'll do that
• 14:01 <demory> in fact, perhaps Deployer should do that automatically for multi-feed builds
• 14:01 <FrankP> strange that it's not there, since showAgencyInfo is true by default ... maybe no Agency url in those feeds?
• 14:02 <novalis_dt> For NYC, we should probably have a warning about weekends, too.
• 14:02 <demory> hmm.. well let me check the actual settings on those
• 14:03 <novalis_dt> FrankP, is the intersection code in the solr geocoder compatible with the tables produced by OSMsplit?
• 14:04 <FrankP> (looking at the code, you don't need url ... but do need Agency Name)
• 14:04 <FrankP> novalis_dt, yeah I think so...was going to make intersection a perm table -- not temporary
• 14:04 <FrankP> (and use those intersections to populate SOLR)
• 14:05 <novalis_dt> Sounds like a plan.
• 14:05 <jopie> FrankP: you're the GUI goto-guy right?
• 14:05 <demory> FrankP, yeah, that must be the issue, since showAgencyInfo is set to true
• 14:05 <FrankP> jopie I do mostly GUI stuff on the OTP project
• 14:06 <FrankP> (webapp specifically)
• 14:07 <demory> novalis_dt what is the NYC weekend warning?
• 14:07 <jopie> Ok, is there a reason ExtJS locales aren't loaded in the default index.html? There's not a lot of strings in there, but some OTP strings don't seem to work unless the locale is loaded. Can patch, just wondering if there was a reason
• 14:07 <novalis_dt> demory, "MTA does construction on weekends but doesn't tell us the schedule in advance; trip planning results are likely to be wrong"
• 14:07 <mattwigway> Incidentally, does the SF Bay instance not have Oakland? I'm trying to plan a trip to my former office but can't.
• 14:08 <demory> mattwigway, it should include all of BART and AC Transit.. what else covers Oakland?
• 14:08 <demory> sorry I don't know the area that well
• 14:09 <mattwigway> No, that should be it... let me confirm it's not my starting address.
• 14:10 <mattwigway> OK, never mind... it's something about my starting address... I'll investigate further later. Does the map not include Caltrain? It suggested I bike from the South Bay to San Francisco to catch BART.
• 14:11 <FrankP> jopie ... English bias, I guess. Actually, I didn't realize that Ext didn't load strings automatically based on the user-agent or some other http headers. When I test different languages in OTP (but with my default en-US browser setting), I don't see any problems. Let me know what to add to index.htm, and I'll happily do it.
• 14:11 -!- Skinkie [~Skinkie@2001:470:1f15:9ee:21d:60ff:feaf:2e9c] has joined #opentripplanner
• 14:12 <demory> mattwigway, that's odd.. it does include CalTrain. i can plan trips on it
• 14:12 <demory> http://sfbay.deployer.opentripplanner.org/#/submit&fromPlace=37.478356,-122.200008&toPlace=37.786283,-122.428530&mode=TRANSIT,WALK&min=QUICK&maxWalkDistance=840&time=2:09 pm&date=04/05/2012&arr=Depart&itinID=1&wheelchair=false
• 14:12 <mattwigway> http://sfbay.deployer.opentripplanner.org/#/submit&fromPlace=37.574607,-122.342774&toPlace=37.821665,-122.277017&mode=TRANSIT,BICYCLE&min=TRIANGLE&triangleTimeFactor=0&triangleSlopeFactor=0&triangleSafetyFactor=0.999999999&maxWalkDistance=4828&time=7:02%20am&date=04/05/2012&arr=Depart&itinID=1&wheelchair=false
• 14:13 <novalis_dt> Is the caltrain running at that time?
• 14:13 <mattwigway> It must be a bike thing... this one will only let me take the ferry, but bikes are definitely allowed on Caltrain and AC Transbay at rush hour.
• 14:13 <novalis_dt> We probably don't have defaultBikeAllowed set
• 14:13 <mattwigway> Yes, Caltrain is running. At that time, BART doesn't allow bikes, but Caltrain always allows bikes, as does AC.
• 14:13 <novalis_dt> And they don't specify (using the unofficial GTFS extension)
• 14:14 <demory> that's a bike+transit trip btw -- if you switch it to transit only, it has you on the train
• 14:14 <mattwigway> OK, explains it.
• 14:14 <novalis_dt> demory, should we have a defaultBikesAllowed checkbox in OTPSetup?
• 14:15 <demory> that would be a feed-specific option?
• 14:15 <novalis_dt> Yes
• 14:16 <mattwigway> Yeah, it's telling me I can't take AC Transbay either.
• 14:16 <mattwigway> With a bike that is.
• 14:16 <demory> ok, i'll ticket that
• 14:16 <andrewbyrd> demory, novalis_dt: req42 builds fine for me here now, no warnings about repeated edge inserts but noticed something else
• 14:16 <mattwigway> AFAIK, BART is the only Bay Area agency with any bike restrictions.
• 14:17 <andrewbyrd> with NED enabled I get many "A state's weight is being incremented by NaN while traversing edge"
• 14:17 <kpw> demory, we should probably think about generalized ways of handling thes flags. we've got a similar need for nyc transfers, correcT?
• 14:17 <novalis_dt> andrewbyrd, I saw that one recently too during graph build
• 14:17 <andrewbyrd> (which is not within NEDBuilder itself, but the local stops finder searches)
• 14:17 <novalis_dt> Which is very odd.
• 14:17 <andrewbyrd> turning off the NED builder they go away
• 14:18 <novalis_dt> andrewbyrd, is this with my recent changes?
• 14:18 <demory> kpw, yes, i can see this sort of thing coming up repeatedly
• 14:19 <demory> i'll write the ticket as a generalized approach
• 14:19 <andrewbyrd> yes, I am running with some commits on top of origin/master
• 14:19 <novalis_dt> andrewbyrd, Do you want to investigate or should I?
• 14:20 <andrewbyrd> just pushed everything. I've got a build running, already looking at it
• 14:21 <novalis_dt> ok
• 14:21 <novalis_dt> I'll keep on area routing
• 14:21 <mattwigway> novalis, are you taking the visibility graph approach>
• 14:22 <novalis_dt> mattwigway, yes, I'm porting visilibity (sic).
• 14:24 <novalis_dt> OK, anyone have anything else?
• 14:25 <demory> nothing here
• 14:25 <mattwigway> I probably won't be able to come to the meetings again until early July :( I'll try to stay in the loop though.
• 14:25 <mattwigway> I'm just finishing up my Analyst poster this week. Andrew agreed to review it, if anyone else would like to glance at it that'd be great.
• 14:26 <novalis_dt> Glad to take a peek
• 14:26 <novalis_dt> And we'll look for you on the mailing list or at other times.
• 14:26 <demory> mattwigway I'm happy to look at that too
• 14:26 <andrewbyrd> mattwigway, yes send it along whenever you're ready
• 14:26 <mattwigway> I'll send you guys an email with a Google Docs link; it has a lot of maps so it looks like it'll be too big to email.
• 14:27 <mattwigway> Thanks! I'll make an effort to read meeting transcripts too.
• 14:29 <FrankP> take care all...good luck at school mattwigway
• 14:29 <mattwigway> Thanks FrankP!
• 14:29 <qlex> mattwigway: include me too
• 14:29 <qlex> [email protected]
• 14:29 <qlex> thanx
• 14:30 <mattwigway> Will do, thanks
• 14:30 <demory> alright have a good week everyone!