QfieldCloud and shared local datasets

[S1lk3]

I use QGIS 3.22.12, QField 2.7.6 on android (tested with both phone and tablet) and the QfieldCloud.

I built a project in QGIS with 2 vector layers (in geopackages), 2 tables (in the same geopackage) and 1 raster file. The raster file is 12GB in size (*.mbtiles based on google satellite and created with "Generate XYZ Tiles" (tile format = JPG)).

Since the raster file is to serve as the basemap for several QField projects that will be used offline, I would like to use it as a "shared local dataset" (https://docs.qfield.org/how-to/outside-layers/?h=basemap). Here, however, I do not succeed.

In detail what I have done:

In QGIS

• added 2 vector layers and 2 tables
• added the raster file (*.mbtiles) from folder "C:/Prive/basiskaarten"
• QGIS > Settings > Options > Datasources: localized data paths -> added "C:/Prive/basiskaarten". (printscreen 1)
• Project Properties - QField: "satellite" -> directly access data source + added as basemap (see printscreen 2)
• all layers and the project itself have project system with EPSG: 3857

On my android device:

• Copied the *.mbtiles file from my desktop to the folder "...\Android\data\ch.opengis.qfield\files\QField\basemaps" with the exact same name.

I save my QGIS project as a *.qgs file (also tried with *.qgz but with a *.qgs I can check if my raster file is effectively saved as localized) and create a new QFieldCloud project ("Convert currently open project to cloud project").

On the qfield cloud webpage, I then see the following error message under uploaded project > jobs: "invalid_dataprovider Provider is not valid (provider: gdal, URI:". (printscreen 3) In Qfield itself, I then see the exact same error message when opening the project. The other layers do load.

The error message seems incomplete and I have no idea what I am doing wrong. Anyone a suggestion? Thanks in advance.

[S1lk3]

Perhaps interesting to add: if I export the project with all files via "package for QField", and import it into QField like this (i.e. without the QField Cloud) then it does manage to load the *.mbtiles file. So I think the problem is not linked to the file itself.

[S1lk3]

I now see that it was mentioned here (#2736) a year ago that this has yet to be implemented for qfieldcloud. Can an update be given on this please?

Is there a workaround to get my 12GB satellite image into my QfieldCloud project, by manually copying it to a folder on my QField device? It is not desirable to put this heavy file in the cloud given the high cost. I do want the vector files in the cloud; I want to be able to pull in the changes daily via QFieldCloud.

[dougpiercey]

Followed the same process and having the same issue, but with a 10GB satellite image .tif. Also looking to avoid storing this in the cloud due to cost and efficiency.

[dougpiercey]

Any suggestions on this? Major issue for us.

[suricactus]

Support for localized datasets with QFieldCloud is coming in the next weeks as a new feature. Will update you when implemented.

[dougpiercey]

Thanks for the update!

[dougpiercey]

Hi...just checking in on progress with this. Thanks

[NellySDIS30]

Hi, just checking in to know when this functionnality is going to be implemented...

[max-mapper]

I tried today on Android to follow the instructions here: https://docs.qfield.org/how-to/outside-layers/. I created a project on QGIS desktop with a basemap.mbtiles file that resided in the folder I added into the Localized Data Paths. I confirmed that file was not synced when I synced it to QfieldCloud. Then I manually placed basemap.mbtiles into my devices ch.opengis.qfield/files/QField/basemaps/ folder. But when I open the synced project in QField it my basemap.mbtiles does not show up in the layers panel. Interestingly, when I try QField Unstable from the Play Store (which turns out to be a build from December 2022) the basemap.mbtiles layer shows up in the layer panel but does not render anything. I am pretty sure this page is inaccurate https://docs.qfield.org/how-to/outside-layers/ and this functionality is totally broken (regardless of if you use QFieldCloud or not) unless someone can point me at a working example

[meyerlor]

@maxogden i work with localized data (raster) paths on many projects. It is working perfectly when syncing via cable. Syncing via cloud does not support local datasets so far, that is what is stated here in this thread.
If you create a local data path in QGIS and put your raster in there, you need to remove the raster from the QGIS project, restart QGIS and then add the raster to the project. Save the project, then sync via cable. To verify if QGIS/QFieldSync recognizes the raster data as coming from a localized path, in QFieldSync options it should look like in my screenshot:

[S1lk3]

Still no update on this issue? Also blocking issue here, qfield cloud is useless to us without this feature, since we work in remote areas.

[janekska]

+1

[dpiercey]

Hi...any update on the implementation of this? Is it going to happen? If not, we will have to cancel our account.

[threalslimtobeeee]

We, as a little NGO working in remote rainforest areas, would also be very happy about this.

[threalslimtobeeee]

Another suitable workaround for us would be:
A Project which is flagged to be used as basemap, would also be the basemap of other Projects in Qfield. Not just of standalone datasets. (see #4311 )

[meyerlor]

somthing just got merged :)
opengisch/QFieldCloud#817

[S1lk3]

Any update on this? I just tested and I still get the same error..

[suricactus]

Hey, I am sorry for the regression introduced in late March! Please try again now, we did a small change that should have fixed this issue again.

I would recommend you re-upload your QGIS file (no modifications needed, just slightly move the map canvas in QGIS and reupload the .qgs file) so all the bad state from your project is cleared.

[jguelat]

It's working again, thanks!

[aghiraldi]

Thanks ! It works great now !

[suricactus]

Thanks a lot for confirmations! And again, apologies for the regression. It is great to see people using this functionality in the wild :)

[meyerlor]

great stuff! working fine again :)