| 02:36:12 | CIA-70: | wonder * r10684 /trunk/qgis/src/gui/qgsmapcanvas.h: fixed documentation for zoomByFactor( ) |
| 02:36:15 | sigq: | tracfeed: Changeset [10684]: fixed documentation for zoomByFactor( ) <https://trac.osgeo.org/qgis/changeset/10684> |
| 02:55:04 | sigq: | tracfeed: Ticket #1661 ( bug created ): Importing a void shapefile created with ftools crashes Qgis <https://trac.osgeo.org/qgis/ticket/1661> |
| 04:18:18 | mib_p8f6oc: | Hi! |
| 04:21:37 | dassouki: | hi |
| 04:27:31 | sigq: | tracfeed: Ticket #1662 ( enhancement created ): Keyboard shortcuts for "Edit" menu <https://trac.osgeo.org/qgis/ticket/1662> |
| 07:27:57 | cgs_bob: | hello all. I'm trying to get qgis to work with grass 6.5 ( I also have grass 7.0 ), but qgis crashes with an "Incompatible library version for module" error. I compiled qgis with the correct prefix ( /usr/local/grass-6.5.svn ), but it seems to be picking up the 7.0.svn.so version of grass. anyone got suggestions? |
| 07:41:14 | cgs_bob: | I'll go ahead and file a ticket on this and check the logs later : ) |
| 08:28:23 | racicot: | anyone know of a timeline ( or if anyone is/will be working on ) support for data source paths to be relative to the location of a project file when they are stored in the project file? |
| 09:10:53 | dassouki: | is there a way to use the distance matrix ( ftools ) as a cost matrix where cost can be any field in the layer; furthermore, can i do the distance matrix based on the linear ( road ) layer |
| 09:53:39 | : | <_-Slash-_> yop |
| 10:15:37 | wildintellect: | dassouki, a little coding and v.net could get you a distance matrix via road length - I'm not sure what you mean by your 1st question |
| 10:16:01 | wildintellect: | racicot, I've seen the topic come up but don't recall if theres even a ticket for it |
| 10:16:52 | dassouki: | well ftools using distance as a measure between zones |
| 10:17:01 | racicot: | wildintellect: there is https://trac.osgeo.org/qgis/ticket/1211 but I was not sure if others were actively looking at it. |
| 10:17:03 | sigq: | Title: #1211 ( Relative paths within QGIS project file ) - Quantum GIS - Trac ( at trac.osgeo.org ) |
| 10:17:03 | dassouki: | ftools uses distance to measure the distance between points |
| 10:17:15 | dassouki: | i was wondering if there is a way to use a cost function instead |
| 10:17:19 | dassouki: | based on field paramters |
| 10:17:27 | racicot: | I have a need ( or potential client ) who needs to be able to ship a project file with relative paths to data... |
| 10:17:33 | dassouki: | think of it as a penalty |
| 10:17:45 | wildintellect: | dassouki, what you want is called least cost path |
| 10:17:56 | wildintellect: | its a raster based operation |
| 10:18:33 | dassouki: | well my table has the "length", "cost" and "penalty" built into it |
| 10:18:39 | wildintellect: | you convert your penalty layer to raster and then use least cost path to find the cheapest path across from A to B |
| 10:19:13 | dassouki: | ooh k |
| 10:19:29 | wildintellect: | well are you still talking about a network? |
| 10:19:40 | wildintellect: | or direct over land distance |
| 10:20:20 | wildintellect: | racicot, I'm only aware of a thread this week highlighting the need for this feature |
| 10:20:39 | wildintellect: | although I bet I asked about it before |
| 10:22:08 | racicot: | I am trying to hack a side solution for now... since this would have to work with current deployments of qgis in this case. I can modify the working directory via python and then relative paths will work in the project file, but hooking into a signal that happens when a project file is loaded *before* the layers are loaded is proving to be challenging |
| 10:27:25 | wildintellect: | racicot, what about a plugin that does the magic on the qgis file and then passes it to the open project routine |
| 10:28:14 | jachym: | ping doktoreas |
| 10:32:19 | racicot: | well, what I was hoping is that I could have someone install the plugin and then it would automatically switch the working directory to the newly opened project directory automagically from then on... no user intervention needed besides installing it once and then forgetting about it... |
| 10:32:30 | racicot: | no messing with the project file itself. |
| 10:34:41 | wildintellect: | so you mean on SIGNAL of open project your plugin would sit and wait to catch that change with a SLOT to a function that changes the qgis working directory |
| 10:35:14 | racicot: | yep... but it would have to be before the layers are loaded... |
| 10:36:39 | racicot: | http://doc.qgis.org/head/classQgsProject.html#8b66ad18c789739bbd0ab9c16ab9dbe6 happens *after* the project is read in... to late to get the paths correct |
| 10:36:40 | sigq: | Title: Quantum GIS API Documentation: QgsProject Class Reference ( at doc.qgis.org ) |
| 10:37:33 | racicot: | might not be possible... but was thinking about it this morning as a possible work around. oh well |
| 10:43:05 | wildintellect: | racicot, do you know where the UI file is for the general user interface |
| 10:43:22 | dassouki: | wildintellect: if u have a few minutes, what i'm trying to do is the all-or-nothing method provided in teh following pdf http://www.eng.wayne.edu/legacy/forms/4/Traffic_Assignment.pdf |
| 10:44:57 | wildintellect: | dassouki, the method I sent you the other day is what you want then v.net accepts a cost variable - I didn't use it in my work but that doesn't mean you can't |
| 10:46:00 | dassouki: | this is a bit different wildintellect as it evaulated teh cost between all links, and let's say one link is at capacity, the next best link is chosen. the code you provided teh other day was great in connecting the centroids to the network |
| 10:50:55 | wildintellect: | what do you mean by capacity |
| 10:52:40 | racicot: | wildintellect: the UI is built dynamically off a pretty much "clean slate" for the general UI. There is no designer file that you can use to swizzle things in the general interface... |
| 10:54:33 | dassouki: | wildintellect: road capacity. a rule of thumb is that a road can handle 1,800 vehicles per hour per lane per direction. so let's say my input data says there aer 10,000 vechiles going from zone a to zone b at given time ti, and let's say the linka,b capcity is 5,000 then 5,000 will be put ot on link a,b and the other 5 will be diverted to other routes |
| 10:55:26 | wildintellect: | but if you simply go in order how do you decide which of your 5005 routes you just calculated to shify |
| 10:56:22 | wildintellect: | racicot, ah it would be easier in the new version 1.1 or trunk since there is a Ui file for the project opening now... |
| 10:56:45 | dassouki: | wildintellect: that's where the cost function comes in |
| 10:57:10 | dassouki: | so cost a,b = $5 cost a->c, c-> $7 with capacity 3,000 |
| 10:57:32 | dassouki: | and cost a->e, e->b cost $15 capacit 40,000 |
| 10:57:33 | racicot: | wildintellect: i want to be able to hook into a signal since most of the time it will be someone double clicking a project file or passing the project file as a command line argument in a startup script. |
| 10:58:00 | dassouki: | it'll chose the a-c, c-b and then dump the rest 2,000 on a-e,e-b |
| 10:58:25 | wildintellect: | racicot, sorry best Idea I have is make them click a button you make |
| 10:58:46 | wildintellect: | dassouki, I see 2 things here, 1 is a cost weighted network distance matrix |
| 10:58:51 | racicot: | yep... might have to fall back to that... if that is OK with the person who wants this |
| 10:59:17 | racicot: | thanks for bouncing the idea around wildintellect |
| 10:59:49 | wildintellect: | dassouki, 2 is an optimization of possible routes and capacities to pick between the choices |
| 11:01:21 | dassouki: | wildintellect: right now ihave a corss system between qgis and openoffice.calc . the final step is actually figuring out the costs of the link paths |
| 11:01:48 | wildintellect: | ah well that we can do |
| 11:02:01 | wildintellect: | gimme a sec to show you some psuedo code |
| 11:05:34 | dassouki: | thanks |
| 11:15:49 | wildintellect: | hmm, I see the dilemna now - you need to map all possible routes between A and B |
| 11:18:59 | dassouki: | wildintellect: another way to think of it, plot an extra map between a and b if and only if the cheapest route is filled to capacity |
| 11:19:40 | wildintellect: | how much are you willing to write code |
| 11:19:49 | wildintellect: | I don't think this can be solved out of the box |
| 11:20:00 | dassouki: | wildintellect: a third way of thinking of it, is that routes c to d might be c to a, a to b, b to d. |
| 11:20:03 | jachym: | dassouki: otto? |
| 11:20:04 | wildintellect: | well maybe with TransCad but that costs $ |
| 11:20:53 | wildintellect: | dassouki, that would be a traveling salesman |
| 11:21:11 | dassouki: | wildintellect: well whatever 3 4 hours of coding will get me |
| 11:21:11 | dassouki: | for today at least |
| 11:21:24 | dassouki: | wildintellect: well the altorithms and equations were given in teh pdf, it's a model we use in the industry, and in a way it's frowned upon if we use a tsp or a nearest insertion |
| 11:21:32 | dassouki: | i'd prefer to use a nearest insertion |
| 11:22:29 | wildintellect: | I think we might want to move your new network to postgis or spatialite |
| 11:22:48 | dassouki: | wildintellect: my network is on postgis already |
| 11:22:59 | wildintellect: | ah |
| 11:23:49 | wildintellect: | have you seen http://pgrouting.postlbs.org/wiki/pgRoutingDocs |
| 11:23:50 | sigq: | Title: pgRoutingDocs - pgRouting - Trac ( at pgrouting.postlbs.org ) |
| 11:24:42 | dassouki: | yes i have |
| 11:25:48 | dassouki: | wildintellect: this is a simpler pseudo code for what i'm trying to do http://www.webs1.uidaho.edu/niatt_labmanual/Chapters/traveldemandforecasting/theoryandconcepts/TripAssignmentAnalysis.htm |
| 11:25:49 | sigq: | Title: Trip Assignment Analysis ( at www.webs1.uidaho.edu ) |
| 11:28:07 | wildintellect: | well I see 2 issues, 1 it's a lot of work to calculate every possible route from A to B, 2 because you need to optimize for capacity every possible combination of paths needs to be assigned and test |
| 11:29:08 | wildintellect: | I mean you could calculate shortest path for all of the options, and if by chance you don't exceed any capacities you're in luck |
| 11:29:31 | dassouki: | wildintellect: well from teh field data i know that at least 4 bridges are goign to exceed |
| 11:30:05 | wildintellect: | if you do exceed capacity then one by one for all routes the break capacity you'll need to drop that segment from the network and run again |
| 11:30:29 | wildintellect: | is this a class assignment? or research? |
| 11:30:34 | dassouki: | neither |
| 11:30:51 | wildintellect: | work or play? |
| 11:30:55 | dassouki: | the software we use to do this costs us $10,000 per year, to basically do that task |
| 11:31:14 | dassouki: | i already switched all our layers to postgis |
| 11:31:19 | dassouki: | and we do most of the stuff on qgis now |
| 11:31:37 | dassouki: | the two things that i'm working on right now, is finishign up trip assignment and the other is an equation editor |
| 11:31:40 | dassouki: | and then i'll be set |
| 11:31:44 | dassouki: | we are researchers yes |
| 11:31:53 | wildintellect: | the short answer is it can be done but you'll have write the optimization algorithms |
| 11:32:01 | wildintellect: | I suggest pgRouting |
| 11:32:49 | wildintellect: | basically when you hit capacity on a model run, exclude that segment with SQL so you can assess all the possible routes |
| 11:33:05 | dassouki: | well i got all summer to do it |
| 11:33:18 | dassouki: | lots of pythong and postgis learning to go through |
| 11:33:24 | dassouki: | python* |
| 11:33:44 | wildintellect: | it's a nested loop that call's itself with the over capacity segment dropped from the inner loop |
| 11:35:22 | dassouki: | after all that is done, i need to implement some form of a way to identify road directions ( one vs. two direction ) ( left turn only, left+Straight, so on .., ) as well |
| 11:35:43 | dassouki: | wildintellect: thanks a lot of the help, it is greatly appreciated |
| 11:36:03 | wildintellect: | pgrouting can do that |
| 11:36:27 | dassouki: | i'll look deeper into pgrouting today |
| 11:36:30 | sigq: | tracfeed: Ticket #1663 ( bug created ): QGIS uses wrong GRASS library. <https://trac.osgeo.org/qgis/ticket/1663> |
| 11:36:31 | wildintellect: | it has some way of letting you write your own optimization algorithms too( I've heard ) |
| 11:37:22 | wildintellect: | dassouki, http://pastebin.com/m3a765788 |
| 11:37:23 | sigq: | Title: pastebin - collaborative debugging tool ( at pastebin.com ) |
| 11:37:56 | dassouki: | ah thanks, i suspect i'm gonna have to resort to numpy for matrix analysis as well |
| 11:43:09 | wildintellect: | dassouki, haven't read much about but also look at graphserver - python routing stuff |
| 11:45:01 | dassouki: | ah that seems awesome multimodal stuff is ideal :D thank sa bundle i'll look into that |
| 11:45:09 | dassouki: | anyways time to lock the office down |
| 11:45:10 | dassouki: | have a good day |
| 11:48:54 | wildintellect: | cgs_bob, that's odd when I run ldd it doesn't tell me anything about grass - maybe that's my issue |
| 11:50:20 | cgs_bob: | wildintellect: talking about my bug report? : ) |
| 11:50:28 | wildintellect: | yes |
| 11:50:44 | wildintellect: | wondering about my issues related to grass as well |
| 11:51:38 | cgs_bob: | what's your issue? |
| 11:52:13 | wildintellect: | oh I compiled several versions of qgis yesterday - grass rasters won't load in any of them and the module list never shows up |
| 11:52:52 | wildintellect: | gotta be related to gdal-grass |
| 11:53:16 | cgs_bob: | are you using the gdal-grass plugin? |
| 11:53:44 | wildintellect: | I think it's installed |
| 11:53:58 | wildintellect: | my repo based install of qgis works fine |
| 11:58:04 | cgs_bob: | so you installed it from source? btw, I found out in #gdal that you can make a new gdal-grass tar ball, by going into gdal's frmts/grass directory and do a make dist |
| 11:58:32 | wildintellect: | ah thanks I'll have to try that |
| 11:59:27 | wildintellect: | well I think the source thing is where all the confusion starts - qgis from source yes everything else no |
| 11:59:43 | cgs_bob: | you're welcome |
| 11:59:54 | wildintellect: | although I did try a grass from source, and I bet that failed cause I didn't have gdal-grass built for it |
| 12:01:07 | lynx_r: | hi, you can try this script to compile gda-grass plug-in http://dpaste.com/39404/ |
| 12:01:15 | sigq: | Title: dpaste: #39404 ( at dpaste.com ) |
| 12:01:55 | cgs_bob: | I've built gdal and qgis from source. I've had little problems until I got the prob in ticket #1663 |
| 12:04:03 | cgs_bob: | I got to run to lunch...bbiab |
| 12:05:58 | sigq: | tracfeed: Ticket #1664 ( bug created ): fTools geoprocessing changes field format <https://trac.osgeo.org/qgis/ticket/1664> |
| 21:03:31 | sigq: | tracfeed: Ticket #1278 ( bug closed ): Layer transparence in tif format fails <https://trac.osgeo.org/qgis/ticket/1278#comment:12> |
| 22:53:26 | CIA-70: | cengel * r10685 /docs/trunk/english_us/user_guide/ ( 2 files in 2 dirs ): |
| 22:53:26 | CIA-70: | added subsections |
| 22:53:26 | CIA-70: | edited changes for postgis sll, improved handling of postgis and pixels as |
| 22:53:26 | CIA-70: | map units in chap 5 |