def load_accdb_and_python_datasets(veh_nums=None):
p1, c1 = vehdb.merge_db_vehicle_subgroups(
inp_h5fname, "prop", "cycle", veh_nums=veh_nums
)
p2, c2 = vehdb.merge_db_vehicle_subgroups(
out_h5fname, "oprop", "cycle", veh_nums=veh_nums
)
## Originally fetched as 2-levels (veh, item) MultiIndex Series.
p1 = p1.unstack()
p2 = p2.unstack()
## By the spec, V rounded to 2-digits,
# But exporting MSAccess --> Excel outputs garbage decimals!
#
v_cols = "v v_orig v_cap v_downscale".split()
c1[v_cols] = vround(c1[v_cols])
## accdb does not offer `n_max`.(?)
p1["n_max"] = nround1(p1["n_max1 n_max2 n_max3".split()].max(axis=1))
return p1, c1, p2, c2
def test_make_v_grid(start, end, nsamples):
grid = engine._make_v_grid(start, end)
assert len(grid) == nsamples
npt.assert_allclose(grid, invariants.vround(grid))
wio.veh_name(case): interpolate_veh(case)
for case in vehnums_to_run
}
df = pd.concat(
results.values(),
axis=0,
keys=results.keys(),
names=["vehicle", "v"],
verify_integrity=True,
)
assert df.index.names == ["vehicle", "v"]
assert (df.index.levels[0] == wio.veh_names(vehnums_to_run)).all()
assert df.columns.names == ["item", "gear"]
assert not (set("n Pwot ASM".split()) - set(df.columns.levels[0]))
npt.assert_allclose(df.index.levels[1],
invariants.vround(df.index.levels[1]))
def test_attach_p_avail_in_gwots_smoketest(h5_accdb):
gwots = pd.DataFrame({("p", "g1"): [], ("ASM", "g1"): []})
engine.attach_p_avail_in_gwots(gwots, f_safety_margin=0.1)
def test_n_max_pipeline():
pipe = pipelines.n_max_pipeline()
props = goodvehicle.goodVehicle()
wot = engine.preproc_wot(props, props["wot"])
sol = pipe.compute(
{
**dtz.keyfilter(
lambda k: k in ("n_rated", "p_rated", "n2v_ratios"), props),
def test_dsc_pipelines(wltc_class):
aug = wio.make_autograph()
ops = aug.wrap_funcs([
*pipelines.downscale_pipeline().ops,
*pipelines.compensate_capped_pipeline().ops,
*pipelines.v_distances_pipeline().ops,
# fake dsc & cap
operation(None, "FAKE.V_dsc", "wltc_class_data/V_cycle", "V_dsc"),
])
pipe = compose(..., *ops)
inp = {
"wltc_data": datamodel.get_wltc_data(),
"wltc_class": wltc_class,
"v_cap": 60, # class1 max(V) is 60.4
}
sol = pipe.compute(inp)
assert len(sol["compensate_phases_t_extra"]) == len(
sol["class_phase_boundaries"])
exp_t_missing = ([0, 1, 0], [0, 10, 43, 192], [0, 8, 81,
203], [0, 8, 81, 203])
assert sol["compensate_phases_t_extra"].tolist(
) == exp_t_missing[wltc_class]
exp_compensated_phases = [
[(0, 589), (590, 1023), (1023, 1612)],
[(0, 589), (599, 1032), (1075, 1530), (1722, 2045)],
[(0, 589), (597, 1030), (1111, 1566), (1769, 2092)],
[(0, 589), (597, 1030), (1111, 1566), (1769, 2092)],
]
compensated_phases = sol["compensated_phase_boundaries"]
assert compensated_phases == exp_compensated_phases[wltc_class]
V_compensated = sol["V_compensated"]
assert compensated_phases[-1][-1] == V_compensated.index[-1]
assert (sol["V_dsc"].sum() - V_compensated.sum()) < inp["v_cap"]
exp_compensated_distances = [
[11988.4, 29146.9, 41135.3],
[11162.2, 28209.7, 51666.1, 70103.2],
[11140.3, 28110.5, 50564.1, 69069.5],
[11140.3, 28241.0, 50779.8, 69285.2],
]
compensated_distances = sol["compensated_distances"]
assert (inv.vround(compensated_distances["cumsum"]) ==
exp_compensated_distances[wltc_class]).all()
## No CAPPING
inp = {
"wltc_data": datamodel.get_wltc_data(),
"wltc_class": wltc_class,
"v_cap": 0,
}
sol = pipe.compute(inp)
assert len(sol["compensate_phases_t_extra"]) == len(
sol["class_phase_boundaries"])
regular_phases = sol["class_phase_boundaries"]
V_compensated = sol["V_compensated"]
assert regular_phases[-1][-1] == V_compensated.index[-1]
assert sol["V_dsc"].sum() == sol["V_compensated"].sum()