Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jul 06 2020 18:23

    ghobona on master

    Update implementations.adoc (compare)

  • May 31 2020 09:56
    jeffharrison commented #3
  • May 06 2020 14:45

    ghobona on master

    Add OpenLayers example Merge pull request #9 from tsch… (compare)

  • May 06 2020 14:45
    ghobona closed #9
  • Apr 29 2020 21:06
    tschaub opened #9
  • Apr 28 2020 07:53

    ghobona on master

    VTP2 draft engineering reports … (compare)

  • Apr 25 2020 12:06

    ghobona on master

    Update README.md (compare)

  • Apr 25 2020 11:59

    ghobona on master

    Update README.md (compare)

  • Apr 25 2020 11:54

    ghobona on master

    Update README.md (compare)

  • Apr 25 2020 11:37

    ghobona on master

    Added screenshots from @robinho… (compare)

  • Apr 25 2020 11:27

    ghobona on master

    Added screenshots of Ecere tile… Merge pull request #8 from jers… (compare)

  • Apr 25 2020 11:27
    ghobona closed #8
  • Apr 24 2020 19:31
    jerstlouis synchronize #8
  • Apr 24 2020 19:10
    jerstlouis synchronize #8
  • Apr 24 2020 18:33
    jerstlouis opened #8
  • Apr 24 2020 17:30
    ghobona closed #3
  • Apr 24 2020 17:30
    ghobona commented #3
  • Apr 24 2020 17:28

    ghobona on master

    Demo Tile Matrix Set for EPSG:2… (compare)

  • Apr 24 2020 17:21

    ghobona on master

    Added Hexagon Client (compare)

  • Apr 24 2020 15:51

    ghobona on master

    Update implementation with pyge… Merge branch 'master' into patc… Merge pull request #6 from fran… (compare)

Jerome St-Louis
@jerstlouis
I have a first example of a "virtual collection" executing a coverage -> vector features in distributed tile manner running
(contours at 50m and at 250m from a DTED)
From those collections you can get e.g. Mapbox Vector Tiles: http://maps.ecere.com/geoapi/collections/vtp/Daraa2/Contours/tiles/CDBGlobalGrid/0/57/216.mvt
Jerome St-Louis
@jerstlouis
Another example with the whole world here: http://maps.ecere.com/geoapi/collections/SRTMVTP500
executing a coverage -> vector feature contour generator process*
Jeff Albrecht
@geospatial-jeff
Jeff Albrecht @geospatial-jeff 08:48
Good morning! Sorry I'm a little late to the party. I work for arturo and we've been developing an implementation with STAC similar to @notthatbreezy. I'd love to give a quick demo if there is time to get feedback from the group!
Peter Parslow
@PeterParslow
@ghobona @geospatial-jeff That makes five demos - looking forward to it!
Jeff Albrecht
@geospatial-jeff
thanks you! @PeterParslow
robinhoutmeyers
@robinhoutmeyers
Jerome St-Louis
@jerstlouis
Thank you @robinhoutmeyers also is that netCDF from Japan open data by any chance? :) loved it!
robinhoutmeyers
@robinhoutmeyers
ghobona
@ghobona
Great job everyone! Please remember to create a Pull Request with screenshots of your demos at https://github.com/opengeospatial/OGC-OS-Sprint-04-2020/tree/master/Screenshots
I'm writing a blog post about the sprint early next week.
Jerome St-Louis
@jerstlouis
Thank you very much @robinhoutmeyers ! @ghobona done: PR #8 https://github.com/jerstlouis/OGC-OS-Sprint-04-2020/tree/master/Screenshots .
Tim Schaub
@tschaub
Bummed I couldn't make the sprint! Work got in the way :(. Hoping to catch up on what happened and put together an implementation sometime soon.
Tim Schaub
@tschaub
Has there been any discussion about having separate properties for top and left to avoid axis order issues with topLeftCorner in a tile matrix?
Tim Schaub
@tschaub
Also curious if anybody has suggested migrating away from scaleDenominator (which requires knowing about a magic 0.00028) and instead using resolution (projected units per rendered pixel).
Tim Schaub
@tschaub
map-tiles.gif
Work in progress pull request for OpenLayers using @aaime's GeoServer implementation openlayers/openlayers#10963
Only showing the "boring" polygon style here. Will work on a vector tile implementation after wrapping this up.
@cholmes - I wanted to try out the Planet implementation, but shouldn't the tile URL template use {tileCol}, {tileRow}, etc. (instead of {x}, {y})?
Joan Masó
@joanma747
No discussion on the topLeftCorner separation. No discussion on the scaleDenominator. About the second, nobody likes but it is a necessary pain. With the interest on vector tiles it gains a bit of sense. This change was introduced in WMS 1.1 many many years ago. In my opinion we should avoid rediscussing it.
Jerome St-Louis
@jerstlouis
@joanma747 @tschaub +1 on keeping scaleDenominator as-is, it is meaningful. About the separation of top and left, this also came up again in VTP2. It is true that it introduces this potential for axis order confusion which it would be good to do without. The confusion is particularly great because for geographic projection you could mention either 4326 or crs84 as the CRS for the same TMS (e.g. WorldCRS84Quad). However TileMatrixSet is already approved as a standard. There are other minor issues reported on GitHub which may require a revision. Perhaps this could be re-considered along with them?
@tschaub @cholmes The tile URL template should indeed use {tileMatrix}, {tileRow}, {tileCol}
Chris Holmes
@cholmes
Cool, I missed that in my quick review of Ian's work, I'll ask him about it.
ghobona
@ghobona
@tschaub Thanks for the demo (animated GIF). When you are ready please upload some screenshots to https://github.com/opengeospatial/OGC-OS-Sprint-04-2020/tree/master/Screenshots
Jim Ressler
@jamesresslernga
@notthatbreezy Chris Brown I have a question about your STAC/Open API demo last Friday. If the OGC pilot is focused on vector tiles, why did your demo focus on Planet imagery rather than vector data? The demo seemed to show raster tiles. Were vector tiles only used to display imagery footprints?
Chris Brown
@notthatbreezy

@jamesresslernga I'm not sure about the OGC pilot work - this work wasn't necessarily related to that and I'm not associated with the pilot. I didn't demo anything related to vector tiles since we were still working on it - but on Friday we finished up work to support viewing STAC label datasets/Features as vector tiles in addition to the imagery footprints.

Overall we used the sprint to evaluate and explore how OGC Tiles (both raster and vector) can interact with STAC - the raster part was using the OGC Tiles spec (based on @cholmes quickstart guide) to view raster tile representations of STAC Items/OGC Features.

Jim Ressler
@jamesresslernga
@notthatbreezy Thank you for explaining the addition of vector tiles. Is the Open API spec used in the demo defined in the Franklin github project ?
I mean the Open APIs used for the demo and vector tiles.
Chris Brown
@notthatbreezy

@jamesresslernga no - it's generated dynamically by the server (still working on making our locked down instance public), but here is a sample spec from a running instance -- https://gist.github.com/notthatbreezy/97cc681ec528f93f6d156ebffbeb6666

The return types are not correct, but it gives a sense of how the server is structured at least

Jim Ressler
@jamesresslernga
@notthatbreezy Thank you Chris, that is helpful.
@robinhoutmeyers Robin, I appreciated the demo of the vector and raster tiles demonstrated in the Luciad Client last Friday. The Implementation link on github points to the Luciad Client. You showed examples of making requests in several dimensions. Is the code and API’s used to build the client requests available on github ?
robinhoutmeyers
@robinhoutmeyers
@jamesresslernga Thanks for the feedback! The client is currently not on github; it is a local prototype built on top of the LuciadLightspeed COTS product. However, I was planning to publish the API service online, which includes the JSON description of the dimensions; this contains the required info to generate requests. I'll keep you informed.
Jerome St-Louis
@jerstlouis
@robinhoutmeyers I am very interested in this as well, but as I mentioned last week I would like to consider these capabilities as spanning the Common, Tiles, Coverages and Maps APIs. (with the ability to get either raw values and/or a rendered map based on the supported conformance classes). Also I think both options of getting tiles containing a single instant and tiles containing a time series should be considered. Time-series could be compressed much more efficiently in an encoding that implements small delta from a previous frame.
Jim Ressler
@jamesresslernga
@robinhoutmeyers Thank you Robin.
Ian Turton
@ianturton
Jim Ressler
@jamesresslernga
@ianturton Thank you Ian. Was this client used in one of the demo on April 24?
Ian Turton
@ianturton
@jamesresslernga - No, I didn't manage to get it finished in time for the demos
robinhoutmeyers
@robinhoutmeyers
@jamesresslernga @jerstlouis The prototype map tiles service I used during my demo is online, see
Fully agreed that this should no solely be linked to maps + to add the possibility to supply a set of dimension values
Jim Ressler
@jamesresslernga
@robinhoutmeyers Thank you for the APIs. I see the dimensions for the maptile layers are defined in another link, such as http://demo.luciad.com:8080/MapTiles/collections/relative_humidity/dimensions. For more than a small demo set, wouldn't the temporal dimension become extremely large at 3 minute intervals? Could the dimension be a timespan and interval value?
robinhoutmeyers
@robinhoutmeyers
@jamesresslernga Yes fully agreed! It could indeed be a timespan and interval value, e.g. 2020-04-18T00:00:00Z/2020-04-19T00:00:00Z/P3M according to the ISO 8601 notification.
ghobona
@ghobona
All, a blog post about the recent OGC API - Tiles Sprint has now been published https://www.ogc.org/blog/3233
ghobona
@ghobona
All, registration for the next OGC Code Sprint has been opened. The OGC invites software developers to the May 2021 OGC API Virtual Code Sprint, to be held from May 26th to May 28th, 2021. The Code Sprint will focus on the following draft OGC API specifications: