The get_gsp_id_to_shape(get_gsp_id_to_shape, sheffield_solar_region_path) function produces a table which has the wrong polygons associated with each GSP ID. This has a fairly big effect on the GSP data pipeline since we use the polygons to select the location for satellite and NWP cropped images.
This happens when the function is used with our default arguments:
This is ultimately caused by the dataframe loaded from the sheffield_solar_region_path containing multiple entries for each GSP name (GSPs). The duplicate entries from the table are shown below.
The function includes logic to combine rows where the RegionID is the same. However, each row has a different region ID even when the GSPs column is the same. This may be a leftover from when the GSPs changed, perhaps it worked then.
This should be easily solved by combining the rows with the same GSPs entry.
Describe the bug
The
get_gsp_id_to_shape(get_gsp_id_to_shape, sheffield_solar_region_path)
function produces a table which has the wrong polygons associated with each GSP ID. This has a fairly big effect on the GSP data pipeline since we use the polygons to select the location for satellite and NWP cropped images.This happens when the function is used with our default arguments:
This is ultimately caused by the dataframe loaded from the
sheffield_solar_region_path
containing multiple entries for each GSP name (GSPs
). The duplicate entries from the table are shown below.The function includes logic to combine rows where the
RegionID
is the same. However, each row has a different region ID even when theGSPs
column is the same. This may be a leftover from when the GSPs changed, perhaps it worked then.This should be easily solved by combining the rows with the same
GSPs
entry.