Emissions GQL, load and cache from etsource

[noracato]

Introduces a new Qernel::Emissions class, containing hashed keys for getting and setting emissions from the new emissions.csv, for each dataset in ETSource.

It mimicks the behaviour of Qernel::Area in the sense of area_attributes. However, emissions can be queried by a sector scope.

Examples

# Get 1990 emissions
EMISSIONS(households, co2, 1990) => 2506777
# Get start year emissions
EMISSIONS('industry.metal', ghg)     => 20778881

# Set value for ghg for a sector (not possible for 1990)
UPDATE(EMISSION(households), ghg, VALUE )
# Set value for co2 for a sector with a subsector (not possible for 1990)
UPDATE(EMISSION('industry.metal'), co2, VALUE )

Still to do:

• specs are missing for Qernel::Emissions

[louispt1]

Based on the altered csv structure, I've updated the branch slightly - I will explain further online @noracato but basically the input csvs were split into two separate csvs and the column structure was altered slightly.

There is a slight functional difference in that you need to now specify energetic/non_energetic as 'type':

# Get 1990 emissions
EMISSIONS(households, energetic, co2, 1990) => 2506777
# Get start year emissions
EMISSIONS('industry.metal', energetic, ghg)     => 20778881

# Set value for ghg for a sector (not possible for 1990)
UPDATE(EMISSION(households), energetic, ghg, VALUE )
# Set value for co2 for a sector with a subsector (not possible for 1990)
UPDATE(EMISSION('industry.metal'), energetic, co2, VALUE )

I also added a lot of spec - there are a couple of TODOs where I wasn't sure whether returning an error or nil was better.

The fixtures have been updated to match the expected input from etsource, and I also tested using those fixtures in etsource and sandboxing results, which worked well. Ideally the fixtures would end up with a few more values and we can adjust the expectations in the spec at that point.

[noracato]

Let's discuss this as well.

[louispt1]

Let me know when you have time 👍