return engine.interpolate_wot_on_v_grid(wot, n2vs)
results = {
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"])
]
display(
cmpr.compare((cc1, cc2),
equivalent_props + equivalent_series,
dataset_names,
describe=True))
# %%
## Repeat, to compare while coding.
# display(cmpr.compare((cc1, cc2), equivalent_props + equivalent_series, dataset_names, describe=True))
# %% [markdown]
# AccDB vehicles: 42, 46, 52, 53 & 90 have broken `v_max`, 48 has broken `wot(ASM)`.
# %%
known_bads = set(wio.veh_names([42, 46, 48, 52, 53, 90]))
display(
cmpr.compare((cc1, cc2), equivalent_props + equivalent_series,
dataset_names).set_properties(subset=(known_bads, idx[:]),
color="red"))
# %%
## Repeat, to compare while coding.
# display(cmpr.compare((cc1, cc2), equivalent_props + equivalent_series, dataset_names).set_properties(subset=(known_bads, idx[:]), color='red'))
# %% [markdown]
# ## Compare a vehicle from AccDB <-->PyAlgo *interactively*
# **TODO:** collect and hide all this comparison GUI code below into a python module.
# %%
case_loaded = [None, None]
wot = wot[["Pwot", "ASM"]]
wot["n"] = wot.index
return engine.interpolate_wot_on_v_grid(wot, n2vs)
all_cases = vehdb.all_vehnums(h5_accdb)
results = {wio.veh_name(case): interpolate_veh(case) for case in all_cases}
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(all_cases)).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_n_mins_smoke():
base = engine.calc_fixed_n_min_drives({}, 500, 4000)
results = [base]
for values in itt.product([None, 1], [None, 2], [None, 3], [None, 4],
[None, 5]):
mdl = dict(
zip(
def test_v_max(h5_accdb):
from . import conftest
veh_samples = None
# DEBUG: to reduce clutter in the console.
# veh_samples = 12
# DEBUG: to study buggy cars.
# veh_samples = [76] # diff det_by_nlim
# veh_samples = [3, 21, 22, 104, ] # diff gear
# veh_samples = [38] # diff vmax order higher 1st
# veh_samples = [31] # [23]
def make_v_maxes(vehnum):
props, wot, n2vs = vehdb.load_vehicle_accdb(h5_accdb, vehnum)
wot = wot.rename({"Pwot": "p"}, axis=1)
wot["n"] = wot.index
gwots = engine.interpolate_wot_on_v_grid(wot, n2vs)
gwots = engine.calc_p_avail_in_gwots(gwots, SM=0.1)
gwots["p_resist"] = vehicle.calc_road_load_power(
gwots.index, props.f0, props.f1, props.f2)
rec = vmax.calc_v_max(gwots)
return (props["v_max"], rec.v_max, props["gear_v_max"], rec.g_vmax,
rec.wot)
def _package_wots_df(gear_wot_dfs):
assert gear_wot_dfs
## Merge all index values into the index of the 1st DF,
# or else, themerged-df contains n-gear dupes in each index-value.
#
# first_df, *rest_dfs = gear_wot_dfs.values()
# full_index = np.unique(np.hstack(df.index for df in gear_wot_dfs))
# first_df = first_df.reindex(full_index)
wots_df = pd.concat(
# [first_df] + rest_dfs,
gear_wot_dfs.values(),
axis=1,
# join="inner",
keys=gear_names(gear_wot_dfs.keys()),
names=["item", "gear"],
verify_integrity=True,
)
return wots_df
veh_nums = vehdb.all_vehnums(h5_accdb)
if not isinstance(veh_samples, (list, tuple)):
veh_samples = random.sample(veh_nums,
veh_samples) if veh_samples else veh_nums
recs = [make_v_maxes(vehnum) for vehnum in veh_samples]
vehres = pd.DataFrame(
recs,
columns="vmax_accdb vmax_python gmax_accdb gmax_python wot".split(
),
index=veh_names(veh_samples),
).astype({
"gmax_accdb": "Int64",
"gmax_python": "Int64"
})
wots_df = pd.concat(vehres["wot"].values,
keys=veh_names(veh_samples),
names=["vehicle"])
vehres = vehres.drop("wot", axis=1)
vehres["vmax_diff"] = (vehres["vmax_python"] - vehres["vmax_accdb"]).abs()
vehres["gmax_diff"] = (vehres["gmax_python"] - vehres["gmax_accdb"]).abs()
with pd.option_context(
"display.max_rows",
130,
"display.max_columns",
20,
"display.width",
120,
# "display.precision",
# 4,
# "display.chop_threshold",
# 1e-8,
"display.float_format",
"{:0.2f}".format,
):
print(
f"++ nones: {vehres.vmax_python.sum()} (out of {len(veh_samples)})"
f"\n++++\n{vehres}"
# f"\n++++\n{wots_df.sample(80, axis=0)}"
)
with pd.option_context(
"display.max_columns",
20,
"display.width",
120,
"display.float_format",
"{:0.4f}".format,
):
print(f"\n++++\n{vehres.describe().T}")
vehres = vehres.dropna(axis=1)
# npt.assert_array_equal(vmaxes["vmax_python"], vmaxes["vmax_accdb"])
aggregate_tol = 1e-4 # The digits copied from terminal.
assert (vehres["vmax_diff"].describe() -
[125.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000]
< aggregate_tol).all()
assert (vehres["gmax_diff"].describe() -
[125.0000, 0.1040, 0.3552, 0.0000, 0.0000, 0.0000, 0.0000, 2.0000]
< aggregate_tol).all()
assert (vehres["vmax_diff"] == 0).sum() == 125 and (vehres["gmax_diff"]
== 0).sum() == 125