def test_remove_multiple_geodata(mp):
scen = TimeSeries(mp, *test_args, version="new", annotation="fo")
scen.add_geodata(DATA["geo"])
row = DATA["geo"].loc[[False, True, True]]
scen.remove_geodata(row)
scen.commit("adding geodata (references to map layers)")
assert_geodata(scen.get_geodata(), DATA["geo"].loc[[True, False, False]])
def test_remove_multiple_geodata(mp):
scen = TimeSeries(mp, *test_args, version='new', annotation='fo')
scen.add_geodata(DATA['geo'])
row = DATA['geo'].loc[[False, True, True]]
scen.remove_geodata(row)
scen.commit('adding geodata (references to map layers)')
assert_geodata(scen.get_geodata(), DATA['geo'].loc[[True, False, False]])
def test_get_timeseries_iamc(mp):
scen = TimeSeries(mp, *test_args)
obs = scen.timeseries(region='World', variable='Testing', iamc=True)
exp = DATA['timeseries'].pivot_table(index=['region', 'variable', 'unit'],
columns='year')['value'].reset_index()
exp['model'] = 'Douglas Adams'
exp['scenario'] = 'Hitchhiker'
npt.assert_array_equal(exp[IAMC_IDX], obs[IAMC_IDX])
npt.assert_array_almost_equal(exp[2010], obs[2010])
def test_get_timeseries_iamc(mp):
scen = TimeSeries(mp, *test_args)
obs = scen.timeseries(region="World", variable="Testing", iamc=True)
exp = (DATA["timeseries"].pivot_table(
index=["region", "variable",
"unit"], columns="year")["value"].reset_index())
exp["model"] = "Douglas Adams"
exp["scenario"] = "Hitchhiker"
npt.assert_array_equal(exp[IAMC_IDX], obs[IAMC_IDX])
npt.assert_array_almost_equal(exp[2010], obs[2010])
def populate_test_platform(platform):
"""Populate `platform` with data for testing.
Many of the tests in :mod:`ixmp.tests.core` depend on this set of data.
The data consist of:
- 3 versions of the Dantzig cannery/transport Scenario.
- Version 2 is the default.
- All have :obj:`HIST_DF` and :obj:`TS_DF` as time-series data.
- 1 version of a TimeSeries with model name 'Douglas Adams' and scenario
name 'Hitchhiker', containing 2 values.
"""
s1 = make_dantzig(platform, solve=True, quiet=True)
s2 = s1.clone()
s2.set_as_default()
s2.clone()
s4 = TimeSeries(platform, **models["h2g2"], version="new")
s4.add_timeseries(
pd.DataFrame.from_dict(
dict(
region="World",
variable="Testing",
unit="???",
year=[2010, 2020],
value=[23.7, 23.8],
)))
s4.commit("")
s4.set_as_default()
def test_cache_invalidate(self, test_mp):
backend = test_mp._backend
ts = TimeSeries(test_mp, model="foo", scenario="bar", version="new")
backend.cache_invalidate(ts, "par", "baz",
dict(x=["x1", "x2"], y=["y1", "y2"]))
def test_new_timeseries_error(test_mp):
scen = TimeSeries(test_mp, *test_args, version="new", annotation="fo")
df = {"year": [2010, 2020], "value": [23.5, 23.6]}
df = pd.DataFrame.from_dict(df)
df["region"] = "World"
df["variable"] = "Testing"
# column `unit` is missing
pytest.raises(ValueError, scen.add_timeseries, df)
def test_new_timeseries_error(test_mp):
scen = TimeSeries(test_mp, *test_args, version='new', annotation='fo')
df = {'year': [2010, 2020], 'value': [23.5, 23.6]}
df = pd.DataFrame.from_dict(df)
df['region'] = 'World'
df['variable'] = 'Testing'
# column `unit` is missing
pytest.raises(ValueError, scen.add_timeseries, df)
def test_new_timeseries_as_iamc(self, test_mp):
# TODO rescue use of subannual= here
scen = TimeSeries(test_mp,
*models["h2g2"],
version="new",
annotation="fo")
scen.add_timeseries(DATA["timeseries"].pivot_table(values="value",
index=IDX_COLS))
scen.commit("importing a testing timeseries")
# compare returned dataframe - default behaviour set to 'auto'
assert_timeseries(scen)
# test behaviour of 'auto' explicitly
assert_timeseries(scen, subannual="auto")
# test behaviour of 'False' explicitly
assert_timeseries(scen, subannual=False)
# test behaviour of 'True' explicitly
exp = (DATA["timeseries"].pivot_table(values="value",
index=IDX_COLS).reset_index())
exp["model"] = "Douglas Adams"
exp["scenario"] = "Hitchhiker"
exp["subannual"] = "Year"
assert_timeseries(scen,
exp=exp,
cols=COLS_WITH_SUBANNUAL,
subannual=True)
def test_default(mp, ts):
# NB this is required before the is_default method can be used
# FIXME should return False regardless
ts.commit('')
# Temporary TimeSeries is has not been set_as_default
assert not ts.is_default()
ts.set_as_default()
assert ts.is_default()
# NB TimeSeries cannot be cloned, so create a new one with the same
# name
ts2 = TimeSeries(mp, ts.model, ts.scenario, version='new')
ts2.commit('')
ts2.set_as_default()
assert ts2.is_default()
# Original TimeSeries is no longer default
assert not ts.is_default()
def test_new_timeseries_as_iamc(test_mp):
scen = TimeSeries(test_mp, *test_args, version='new', annotation='fo')
scen.add_timeseries(DATA['timeseries'].pivot_table(values='value',
index=IDX_COLS))
scen.commit('importing a testing timeseries')
# compare returned dataframe - default behaviour set to 'auto'
assert_timeseries(scen)
# test behaviour of 'auto' explicitly
assert_timeseries(scen, subannual='auto')
# test behaviour of 'False' explicitly
assert_timeseries(scen, subannual=False)
# test behaviour of 'True' explicitly
exp = (DATA['timeseries'].pivot_table(values='value',
index=IDX_COLS).reset_index())
exp['model'] = 'Douglas Adams'
exp['scenario'] = 'Hitchhiker'
exp['subannual'] = 'Year'
assert_timeseries(scen, exp=exp, cols=COLS_WITH_SUBANNUAL, subannual=True)
def test_init2(self, test_mp):
# Scheme argument
with pytest.raises(TypeError, match="'scheme' argument"):
TimeSeries(test_mp, "m", "s", scheme="scheme")
def test_fetch_empty_geodata(mp):
scen = TimeSeries(mp, *test_args, version='new', annotation='fo')
empty = scen.get_geodata()
assert_geodata(empty, DATA['geo'].loc[[False, False, False]])
def test_new_subannual_timeseries_as_iamc(mp):
mp.add_timeslice('Summer', 'Season', 1.0 / 4)
scen = TimeSeries(mp, *test_args, version='new', annotation='fo')
timeseries = DATA['timeseries'].pivot_table(values='value', index=IDX_COLS)
scen.add_timeseries(timeseries)
scen.commit('adding yearly data')
# add subannual timeseries data
ts_summer = timeseries.copy()
ts_summer['subannual'] = 'Summer'
scen.check_out()
scen.add_timeseries(ts_summer)
scen.commit('adding subannual data')
# generate expected dataframe+
ts_year = timeseries.copy()
ts_year['subannual'] = 'Year'
exp = pd.concat([ts_year, ts_summer]).reset_index()
exp['model'] = 'Douglas Adams'
exp['scenario'] = 'Hitchhiker'
# compare returned dataframe - default behaviour set to 'auto'
assert_timeseries(scen,
exp=exp[COLS_WITH_SUBANNUAL],
cols=COLS_WITH_SUBANNUAL)
# test behaviour of 'auto' explicitly
assert_timeseries(scen,
exp=exp[COLS_WITH_SUBANNUAL],
cols=COLS_WITH_SUBANNUAL,
subannual='auto')
# test behaviour of 'True' explicitly
assert_timeseries(scen,
exp=exp[COLS_WITH_SUBANNUAL],
cols=COLS_WITH_SUBANNUAL,
subannual=True)
# setting False raises an error because subannual data exists
pytest.raises(ValueError, scen.timeseries, subannual=False)
def test_timeseries_edit(mp):
scen = TimeSeries(mp, *test_args)
df = {
'region': ['World'] * 2,
'variable': ['Testing'] * 2,
'unit': ['???', '???'],
'year': [2010, 2020],
'value': [23.7, 23.8]
}
exp = pd.DataFrame.from_dict(df)
obs = scen.timeseries()
npt.assert_array_equal(exp[IDX_COLS], obs[IDX_COLS])
npt.assert_array_almost_equal(exp['value'], obs['value'])
scen.check_out(timeseries_only=True)
df = {
'region': ['World'] * 2,
'variable': ['Testing'] * 2,
'unit': ['???', '???'],
'year': [2010, 2020],
'value': [23.7, 23.8]
}
df = pd.DataFrame.from_dict(df)
scen.add_timeseries(df)
scen.commit('testing of editing timeseries (same years)')
scen.check_out(timeseries_only=True)
df = {
'region': ['World'] * 3,
'variable': ['Testing', 'Testing', 'Testing2'],
'unit': ['???', '???', '???'],
'year': [2020, 2030, 2030],
'value': [24.8, 24.9, 25.1]
}
df = pd.DataFrame.from_dict(df)
scen.add_timeseries(df)
scen.commit('testing of editing timeseries (other years)')
mp.close_db()
mp.open_db()
scen = TimeSeries(mp, *test_args)
obs = scen.timeseries().sort_values(by=['year'])
df = df.append(exp.loc[0]).sort_values(by=['year'])
npt.assert_array_equal(df[IDX_COLS], obs[IDX_COLS])
npt.assert_array_almost_equal(df['value'], obs['value'])
def test_new_timeseries_as_year_value(test_mp):
scen = TimeSeries(test_mp, *test_args, version="new", annotation="fo")
scen.add_timeseries(DATA["timeseries"])
scen.commit("importing a testing timeseries")
assert_timeseries(scen)
def test_new_timeseries_as_year_value(test_mp):
scen = TimeSeries(test_mp, *test_args, version='new', annotation='fo')
scen.add_timeseries(DATA['timeseries'])
scen.commit('importing a testing timeseries')
assert_timeseries(scen)
def test_get_timeseries(mp):
scen = TimeSeries(mp, *test_args)
assert_timeseries(scen)
def prepare_scenario(mp, args_all):
scen = TimeSeries(mp, *args_all, version='new', annotation='nk')
scen.add_timeseries(DATA[2010])
scen.commit('updating timeseries in IAMC format')
scen = TimeSeries(mp, *args_all)
return scen
def test_timeseries_edit(mp):
scen = TimeSeries(mp, *test_args)
df = {
"region": ["World"] * 2,
"variable": ["Testing"] * 2,
"unit": ["???", "???"],
"year": [2010, 2020],
"value": [23.7, 23.8],
}
exp = pd.DataFrame.from_dict(df)
obs = scen.timeseries()
npt.assert_array_equal(exp[IDX_COLS], obs[IDX_COLS])
npt.assert_array_almost_equal(exp["value"], obs["value"])
scen.check_out(timeseries_only=True)
df = {
"region": ["World"] * 2,
"variable": ["Testing"] * 2,
"unit": ["???", "???"],
"year": [2010, 2020],
"value": [23.7, 23.8],
}
df = pd.DataFrame.from_dict(df)
scen.add_timeseries(df)
scen.commit("testing of editing timeseries (same years)")
scen.check_out(timeseries_only=True)
df = {
"region": ["World"] * 3,
"variable": ["Testing", "Testing", "Testing2"],
"unit": ["???", "???", "???"],
"year": [2020, 2030, 2030],
"value": [24.8, 24.9, 25.1],
}
df = pd.DataFrame.from_dict(df)
scen.add_timeseries(df)
scen.commit("testing of editing timeseries (other years)")
mp.close_db()
mp.open_db()
scen = TimeSeries(mp, *test_args)
obs = scen.timeseries().sort_values(by=["year"])
df = df.append(exp.loc[0]).sort_values(by=["year"])
npt.assert_array_equal(df[IDX_COLS], obs[IDX_COLS])
npt.assert_array_almost_equal(df["value"], obs["value"])
def test_fetch_empty_geodata(self, mp):
scen = TimeSeries(mp, *models["h2g2"], version="new", annotation="fo")
empty = scen.get_geodata()
assert_geodata(empty, DATA["geo"].loc[[False, False, False]])
def test_fetch_empty_geodata(mp):
scen = TimeSeries(mp, *test_args, version="new", annotation="fo")
empty = scen.get_geodata()
assert_geodata(empty, DATA["geo"].loc[[False, False, False]])
def test_new_subannual_timeseries_as_iamc(mp):
mp.add_timeslice("Summer", "Season", 1.0 / 4)
scen = TimeSeries(mp, *test_args, version="new", annotation="fo")
timeseries = DATA["timeseries"].pivot_table(values="value", index=IDX_COLS)
scen.add_timeseries(timeseries)
scen.commit("adding yearly data")
# add subannual timeseries data
ts_summer = timeseries.copy()
ts_summer["subannual"] = "Summer"
scen.check_out()
scen.add_timeseries(ts_summer)
scen.commit("adding subannual data")
# generate expected dataframe+
ts_year = timeseries.copy()
ts_year["subannual"] = "Year"
exp = pd.concat([ts_year, ts_summer]).reset_index()
exp["model"] = "Douglas Adams"
exp["scenario"] = "Hitchhiker"
# compare returned dataframe - default behaviour set to 'auto'
assert_timeseries(scen,
exp=exp[COLS_WITH_SUBANNUAL],
cols=COLS_WITH_SUBANNUAL)
# test behaviour of 'auto' explicitly
assert_timeseries(scen,
exp=exp[COLS_WITH_SUBANNUAL],
cols=COLS_WITH_SUBANNUAL,
subannual="auto")
# test behaviour of 'True' explicitly
assert_timeseries(scen,
exp=exp[COLS_WITH_SUBANNUAL],
cols=COLS_WITH_SUBANNUAL,
subannual=True)
# setting False raises an error because subannual data exists
pytest.raises(ValueError, scen.timeseries, subannual=False)
