def empirical_distribution(contest):
df = ccd.summary(contest)
n = df.shape[0]
dist = np.zeros((n, 3)) # [unfunny, somewhat funny, funny]
for i in range(n):
dist[i] = get_single_dist(df.iloc[i])
# tmp = np.zeros((100,3))
# tmp[:50] = dist[:50]
# tmp[50:] = dist[-50:]
return dist
def test_same_dataframe(filename: str):
if "590" in filename:
pytest.xfail("TODO")
fname = filename.split("/")[-1]
df1 = ccd.summary(fname)
df2 = prd.process(str(raw_dashboards / fname))
assert (df1.columns == df2.columns).all()
for col in df1.columns:
if "float" in df1[col].dtype.name:
assert np.allclose(df1[col], df2[col])
else:
assert (df1[col] == df2[col]).all()
def test_all_summaries(contest):
start = time()
df = ccd.summary(contest)
print("{:0.1f}ms".format(1000 * (time() - start)))
assert isinstance(df, pd.DataFrame)
def df(request):
filename = str(request.param)
fname = filename.split("/")[-1]
return ccd.summary(fname)
def get_means(contest):
means = ccd.summary(contest)['score'].to_numpy()
# tmp = np.zeros(100)
# tmp[:50] = means[:50]
# tmp[50:] = means[-50:]
return means