Fix thumbnail to default aspect ratios

[lucernae]

fix for #497 #517

Now, after we got meters unit, we can properly calculate new extent based on GeoNode default aspect ratios: [200, 150] pixels. We used 400, 300 for clarity, because we displayed the basemap. It helps for the label resolutions to become readable.

There are some weird bug though.

Previously, when we request to WMS in QGIS Server, we provided bbox in the format: [ymin,xmin,ymax,xmax] in EPSG:4326. However, in EPSG:3857, I had to use [xmin,ymin,xmax,ymax] for it to display the correct extent (took me a while to realize). I don't know if this is expected from QGIS Server, cc @gubuntu ?

[lucernae]

Waiting to fix the thumbnail unittests.

[lucernae]

Preview:

[codecov[bot]]

Codecov Report

Merging #522 into 2.8.x-qgis_server will increase coverage by 0.04%. The diff coverage is 93.47%.

@@                  Coverage Diff                  @@
##           2.8.x-qgis_server     #522      +/-   ##
=====================================================
+ Coverage              41.39%   41.44%   +0.04%     
=====================================================
  Files                    412      412              
  Lines                  29597    29625      +28     
  Branches                3746     3748       +2     
=====================================================
+ Hits                   12252    12278      +26     
+ Misses                 16590    16587       -3     
- Partials                 755      760       +5

Impacted Files	Coverage Δ
geonode/qgis_server/tests/test_views.py	99.28% <87.5%> (-0.47%)	:arrow_down:
geonode/qgis_server/helpers.py	69.15% <96.66%> (+0.82%)	:arrow_up:
geonode/base/models.py	74.92% <0%> (-0.29%)	:arrow_down:
geonode/qgis_server/tasks/update.py	77.1% <0%> (+3.61%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 66f33d8...5b5f8c7. Read the comment docs.

[gubuntu]

LGTM

@lucernae the swapped coordinate order for geographic CRS is a 'feature' of WMS 1.3.0

[lucernae]

@lucernae the swapped coordinate order for geographic CRS is a 'feature' of WMS 1.3.0

Oh no, this is true... https://gis.stackexchange.com/questions/23347/getmap-wms-1-1-0-vs-1-3-0

Now we used EPSG:3857 directly in the request, so it should be safe now.