Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Brad Hards
    @bradh
    I vaguely recall thinking that it seemed like the OGC SWG couldn't agree, and went with both :-)
    I'm sure it was a bit more sophisticated than that though.
    Vicky Vergara
    @cvvergara
    @estebanzimanyi @mahmsakr
    FYI, today in the board meeting we have you in the agenda
    https://wiki.osgeo.org/wiki/Board_Meeting_2020-12-28
    Mahmoud SAKR
    @mahmsakr
    Hi Vicky. Looks like you didn't have holiday !
    I see this 'good looking' motion:
    https://www.loomio.org/d/i6gzQpBn/accept-mobilitydb-as-an-osgeo-community-project
    Is it a voting by majority, or all must accept ?
    Vicky Vergara
    @cvvergara
    no -1
    Brad Hards
    @bradh
    Is there a recommended OS (or OS + external package source)?
    It looks like my Kubuntu 20.04 doesn't have the required lwgeom packages, and I'm unclear if building it separately will conflict with an "internal" version from postgis.
    Is 18.04 more usual? Or some real Debian release?
    Vicky Vergara
    @cvvergara
    @bradh
    Try the following:
    Remove PostGIS that comes with the OS, then instal PostGIS 2.5 , then install PostGIS 3 that comes with the OS again.
    That will leave lwgeom from 2.5 but the postGIS that you will be using its postGIS 3.
    Brad Hards
    @bradh
    @cvvergara thanks. Will give it a try.
    Did PostGIS 3 stop exporting lwgeom?
    Vicky Vergara
    @cvvergara
    yes it did
    Brad Hards
    @bradh
    Grr. Spatialite went to some effort to switch to that.
    Vicky Vergara
    @cvvergara
    :-(
    So, Spatialite uses lwgeom right?, and also MobilityDB I think
    Brad Hards
    @bradh
    I'd have to check if Sandro dropped it again for Spatialite.
    What is the path forward for MobilityDB?
    Vicky Vergara
    @cvvergara
    Well, I am not that deep involved, but personally I would think that a next step would need to make it work with any Postgis version, to avoid problems like the one you are having.
    Brad Hards
    @bradh
    I wonder how hard that would be.
    Vicky Vergara
    @cvvergara
    Pretty challenging, hmmm which gives me an idea
    @mahmsakr @estebanzimanyi Soon the GSoC program will start,
    I already have the following in pgRouting wiki:
    https://github.com/pgRouting/pgrouting/wiki/GSoC-Ideas%3A-2021
    It would be interesting to have a GSoC student to build MobilityDB run without the need of PostGIS dependency.
    Brad Hards
    @bradh
    Looks like spatialite already switched to rttopo.
    Initial looks at mobilitydb source suggests that it might benefit from the same. Very early days though...
    Brad Hards
    @bradh
    There are some comments at https://github.com/MobilityDB/MobilityDB/blob/master/point/src/geography_functions.c#L5-L6 that suggest a PR against postgis.
    However I don't see that anywhere.
    Is it still the plan?
    Vicky Vergara
    @cvvergara
    I would hope so.
    Brad Hards
    @bradh
    Worth a ticket?
    Vicky Vergara
    @cvvergara
    Lets wait till vacation is over, and ask them.
    @all I will be back on Jan-11
    Brad Hards
    @bradh
    @cvvergara thanks. Enjoy your break.
    Most people are in .eu ? So maybe mid-January?
    Vicky Vergara
    @cvvergara
    https://portal.ogc.org/public_ogc/register/20210217_code_sprint.php
    Rajat Shinde
    @omshinde
    Hello Team.
    Sorry for cross posting! Please see the trailing email if MobilityDBProject would be interested in floating some project ideas for Google Summer of Code 2021.
    https://lists.osgeo.org/pipermail/discuss/2021-February/039230.html
    Mahmoud SAKR
    @mahmsakr
    Thanks for the proposal Rajat. In the moment, we find it more manageable to recruit master thesis students from our school, for the development of MobilityDB. When our developer community will increase in the future, we will hopefully have more capacity to mentor more project, and join GSoC.
    Daniel Kastl
    @dkastl
    Let me just note, that also a master thesis student from your university would be able to apply for GSoC.
    Mahmoud SAKR
    @mahmsakr
    Interesting ! let me check how the timeline goes with the academic semester.
    Student projects announced mid May, and code submission mid August. Any experience how to align this with an academic semester Feb -to- June ?
    Daniel Kastl
    @dkastl
    This year GSoC has changed significantly with the goal to require less time for the students than the years before.
    So I don't have experience with this. I just thought there could be synergies.
    Rajat Shinde
    @omshinde
    @mahmsakr thanks. Yes, correctly pointed by @dkastl ..
    It should be a project requiring around 175hrs of work which could be distributed over 10 weeks.
    Mentors and students can coordinate how to distribute the tasks over 2 evaluation periods (1 after 5 weeks and other towards the end). :)
    Vicky Vergara
    @cvvergara
    @estebanzimanyi https://www.youtube.com/watch?v=gja9yTzK77g
    Jannis R
    @derhuerst

    hey everyone! i have a basic questions:

    is it possible to construct a tgeogpointi from a list of (geography + timestamptz) items, where some of the items don't have a timestamptz?

    specifically, I'm trying to model GTFS "runs" (a trip at a specific service day, using a specific shape) using mobilitydb.

    • a shape is a list of points, without temporal reference.
    • stop_times.txt defines "runs", a list of of stopovers, with temporal and geospatial reference.

    i can match the stopovers (geography + timestamptz) with some of the shape points (geography), but i want to be able to express those shape points in between that can't be matched.

    or to put it another way: i have a geospatial + temporal trajectory, where I only know the temporal value for some of the points.

    Vicky Vergara
    @cvvergara
    From what I understand, in simple terms, linestring:
    You have the timestamp of the startpoint and endpoint, and you want to add times to every node in the linesting,
    I do not know if there is a function that does that automatically (I am also learning mobilitydb), but I would solve it by
    getting the % of each node of the line from the startpoint, So you have the time difference, and match all those % calculated. Then you have all the corresponding times by adding to the start time.
    Jannis R
    @derhuerst

    i don't only have start and end, but some intermediate points as well, that include timestamps. but the problem remains the same:

    ideally, i would like mobilitydb to interpolate between them on the fly. do i understand correctly that there's no such feature?

    Vicky Vergara
    @cvvergara
    I dont know if there is that feature.
    they have a mailing list maybe you can ask there:
    https://lists.osgeo.org/mailman/listinfo/mobilitydb-users
    https://lists.osgeo.org/mailman/listinfo/mobilitydb-dev
    (I am also learning)