Datasets by geometry

[alrocar]

We have these two entities in the catalog

CREATE TABLE datasets
(
id TEXT  PRIMARY KEY ,
provider_id TEXT,
category_id TEXT,
data_source_id TEXT,
name TEXT,
description TEXT,
country_iso_code3 TEXT,
language_iso_code3 TEXT,
geography_id TEXT,


CREATE TABLE geographies
(
id TEXT PRIMARY KEY,
provider_id TEXT,
name TEXT,
description TEXT,
country_iso_code3 TEXT,
language_iso_code3 TEXT,
version TEXT,
geom_coverage GEOMETRY, -- admin,quadkey,zipcode...

Where geography_id is a FK.

Given a polygon, we want to be able to select the datasets in the catalog with available data for that region. Something like:

datasets = catalog.datasets(polygon)

The format of the polygon is yet to be defined (but it can be a shapely geometry).

We'd need to query the catalog with a query like this:

SELECT d.*
FROM geographies g
LEFT JOIN datasets d
ON g.id = d.geography_id
WHERE ST_Intersects(ST_geometryfromtext(local_polygon_wkt, 4326), geom_coverage)

So basically, convert the polygon to a WKT or WKB and perform the spatial intersection to get the list of datasets.

[alrocar]

cc. @oleurud

[alasarr]

First version working:

It only supports a dataset as input, but I'd like to add support to:

• Pandas dataframe in WKT or WKB
• GeoPandasDataFrame
• Single geom in WKT

[alasarr]

Just added support for WKT and GeoPandasDataFrame 💪