Add to_geopandas() method

[andersy005]

Towards #36

[jhamman]

pinging @scottyhq and @matthewhanson for review/comment.

[matthewhanson]

This looks good to me

[jhamman]

can we also assert that the new df has a few features that we expect (e.g. number or rows, index values, etc.)?

[andersy005]

Sounds good. I added a few more tests to address this

[scottyhq]

Thanks @andersy005 for moving this forward. I think it's a helpful start, but we shouldn't close issue 36 yet because it would be great to have some sort of querying and indexing directly on the catalog as described by Joe
#36 (comment)

Some other things to be aware of are changes to the STAC JSON that can affect this functionality. We should add tests that reflect the STAC version being used and maybe in the near future just enforce using STAC>0.9. It works currently because we're using @matthewhanson's satstac library which saves a self-contained STAC catalog as geojson "FeatureCollection" (which can be read by geopandas) - radiantearth/stac-spec#691

gf = gpd.GeoDataFrame.from_features(cat._stac_obj.geojson(), crs={'init' :'epsg:4326'})

Finally, we'll want to read CRS from the STAC catalog if it exists, and if not assign EPSG:4326 which is the usual. https://github.com/radiantearth/stac-spec/tree/master/extensions/projection

How geopandas handles CRS might change in 0.7 geopandas/geopandas#1003

[andersy005]

Finally, we'll want to read CRS from the STAC catalog if it exists, and if not assign EPSG:4326 which is the usual.

If I understand this correctly, extracting the CRS from the STAC catalog for catalogs that use the commons extension should be straightforward since the CRS is likely to be among the common metadata fields. What should we do in those cases where the commons extension is not used i.e. the CRS field is included on every Item? Do we just pick one item from the collection and assume that its CRS field is common for the rest of the items?

[andersy005]

I think it's a helpful start, but we shouldn't close issue 36 yet because it would be great to have some sort of querying and indexing directly on the catalog as described by Joe
#36 (comment)

Happy to work on this as part of this PR

[jhamman]

Thanks @andersy005!