import genno
import pandas as pd
import pytest
from message_ix_models.tools.exo_data import prepare_computer
from message_ix_models.tools.iea.eei import IEA_EEI # noqa: F401
from message_ix_models.util import HAS_MESSAGE_DATA
pytestmark = pytest.mark.skipif(
condition=not HAS_MESSAGE_DATA, reason="No fuzzed/random test data for this source."
)
# Infill data for R12 nodes not present in the IEA data
# NB these are hand-picked as of 2022-07-20 so that the ratio of freight activity / GDP
# is roughly consistent across regions
# FIXME replace with actual data
R12_MAP = [
("R12_EEU", "R12_AFR"),
("R12_NAM", "R12_CHN"),
("R12_EEU", "R12_EEU"),
("R12_FSU", "R12_FSU"),
("R12_LAM", "R12_LAM"),
("R12_PAO", "R12_MEA"),
("R12_NAM", "R12_NAM"),
("R12_PAO", "R12_PAO"),
("R12_LAM", "R12_PAS"),
("R12_LAM", "R12_RCPA"),
("R12_WEU", "R12_SAS"),
("R12_WEU", "R12_WEU"),
]
[docs]
class TestIEA_EEI:
@pytest.mark.parametrize(
"source_kw, dimensionality",
(
(
dict(
indicator="Passenger load factor",
# broadcast_map="bc:n-n2",
),
{"Mode/vehicle type", "SECTOR"},
),
),
)
@pytest.mark.parametrize(
"regions, aggregate, N_n",
(
("R12", False, 28),
("R12", True, 5),
),
)
def test_prepare_computer( # pragma: no cover cf. iiasa/message-ix-models#164
self, test_context, source_kw, dimensionality, regions, aggregate, N_n
):
test_context.model.regions = regions
source = "IEA EEI"
source_kw.update(aggregate=aggregate)
c = genno.Computer()
s = pd.Series(1.0, index=pd.MultiIndex.from_tuples(R12_MAP, names=["n", "n2"]))
c.add("bc:n-n2", genno.Quantity(s))
keys = prepare_computer(test_context, c, source, source_kw)
# Keys have informative names
assert "passenger load factor" == keys[0].name
# Preparation of data runs successfully
result = c.get(keys[0])
# assert 1394 == result.size
assert 400 <= result.size
assert {"n", "y"} | dimensionality == set(result.dims)
assert N_n == len(result.coords["n"])