def test_covariates_from_event_matrix_with_simple_addition(self):
base_df = pd.DataFrame(
[[1, 0, 5, 1], [2, 0, 4, 1], [3, 0, 8, 1], [4, 0, 4, 1]],
columns=["id", "start", "stop", "e"])
event_df = pd.DataFrame([[1, 1], [2, 2], [3, 3], [4, None]],
columns=["id", "poison"])
cv = utils.covariates_from_event_matrix(event_df, "id")
ldf = utils.add_covariate_to_timeline(base_df,
cv,
"id",
"duration",
"e",
cumulative_sum=True)
assert pd.notnull(ldf).all().all()
expected = pd.DataFrame(
[
(0.0, 0.0, 1.0, 1, False),
(1.0, 1.0, 5.0, 1, True),
(0.0, 0.0, 2.0, 2, False),
(2.0, 1.0, 4.0, 2, True),
(0.0, 0.0, 3.0, 3, False),
(3.0, 1.0, 8.0, 3, True),
(0.0, 0.0, 4.0, 4, True),
],
columns=["start", "cumsum_poison", "stop", "id", "e"],
)
assert_frame_equal(expected, ldf, check_dtype=False, check_like=True)
def test_covariates_from_event_matrix(self):
base_df = pd.DataFrame([
[1, 0, 5, 1],
[2, 0, 4, 1],
[3, 0, 8, 1],
[4, 0, 4, 1]
], columns=['id', 'start', 'stop', 'e'])
event_df = pd.DataFrame([
[1, 1, None, 2 ],
[2, None, 5, None],
[3, 3, 3, 7 ]
], columns=['id', 'promotion', 'movement', 'raise'])
cv = utils.covariates_from_event_matrix(event_df, 'id')
ldf = utils.add_covariate_to_timeline(base_df, cv, 'id', 'duration', 'e', cumulative_sum=True)
expected = pd.DataFrame.from_records([
{'cumsum_movement': 0.0, 'cumsum_promotion': 0.0, 'cumsum_raise': 0.0, 'e': 0.0, 'id': 1.0, 'start': 0.0, 'stop': 1.0},
{'cumsum_movement': 0.0, 'cumsum_promotion': 1.0, 'cumsum_raise': 0.0, 'e': 0.0, 'id': 1.0, 'start': 1.0, 'stop': 2.0},
{'cumsum_movement': 0.0, 'cumsum_promotion': 1.0, 'cumsum_raise': 1.0, 'e': 1.0, 'id': 1.0, 'start': 2.0, 'stop': 5.0},
{'cumsum_movement': 0.0, 'cumsum_promotion': 0.0, 'cumsum_raise': 0.0, 'e': 1.0, 'id': 2.0, 'start': 0.0, 'stop': 4.0},
{'cumsum_movement': 0.0, 'cumsum_promotion': 0.0, 'cumsum_raise': 0.0, 'e': 0.0, 'id': 3.0, 'start': 0.0, 'stop': 3.0},
{'cumsum_movement': 1.0, 'cumsum_promotion': 1.0, 'cumsum_raise': 0.0, 'e': 0.0, 'id': 3.0, 'start': 3.0, 'stop': 7.0},
{'cumsum_movement': 1.0, 'cumsum_promotion': 1.0, 'cumsum_raise': 1.0, 'e': 1.0, 'id': 3.0, 'start': 7.0, 'stop': 8.0},
{'cumsum_movement': None, 'cumsum_promotion': None, 'cumsum_raise': None, 'e': 1.0, 'id': 4.0, 'start': 0.0, 'stop': 4.0}
])
assert_frame_equal(expected, ldf, check_dtype=False, check_like=True)
def test_covariates_from_event_matrix(self):
df = pd.DataFrame([[1, 1, None, 2], [2, None, 5, None], [3, 3, 3, 7]],
columns=['id', 'promotion', 'movement', 'raise'])
ldf = pd.DataFrame(
[[1, 0, 5, 1], [2, 0, 4, 1], [3, 0, 8, 1], [4, 0, 4, 1]],
columns=['id', 'start', 'stop', 'e'])
cv = utils.covariates_from_event_matrix(df, 'id')
ldf = utils.add_covariate_to_timeline(ldf,
cv,
'id',
'duration',
'e',
cumulative_sum=True)
assert ldf.loc[ldf['id'] == 1]['cumsum_movement'].tolist() == [0, 0, 0]
assert ldf.loc[ldf['id'] == 1]['cumsum_promotion'].tolist() == [
0, 1, 1
]
assert ldf.loc[ldf['id'] == 1]['cumsum_raise'].tolist() == [0, 0, 1]
assert ldf.loc[ldf['id'] == 1]['start'].tolist() == [0, 1., 2.]
def test_covariates_from_event_matrix(self):
base_df = pd.DataFrame(
[[1, 0, 5, 1], [2, 0, 4, 1], [3, 0, 8, 1], [4, 0, 4, 1]], columns=["id", "start", "stop", "e"]
)
event_df = pd.DataFrame(
[[1, 1, None, 2], [2, None, 5, None], [3, 3, 3, 7]], columns=["id", "promotion", "movement", "raise"]
)
cv = utils.covariates_from_event_matrix(event_df, "id")
ldf = utils.add_covariate_to_timeline(base_df, cv, "id", "duration", "e", cumulative_sum=True)
expected = pd.DataFrame.from_records(
[
{
"cumsum_movement": 0.0,
"cumsum_promotion": 0.0,
"cumsum_raise": 0.0,
"e": 0.0,
"id": 1.0,
"start": 0.0,
"stop": 1.0,
},
{
"cumsum_movement": 0.0,
"cumsum_promotion": 1.0,
"cumsum_raise": 0.0,
"e": 0.0,
"id": 1.0,
"start": 1.0,
"stop": 2.0,
},
{
"cumsum_movement": 0.0,
"cumsum_promotion": 1.0,
"cumsum_raise": 1.0,
"e": 1.0,
"id": 1.0,
"start": 2.0,
"stop": 5.0,
},
{
"cumsum_movement": 0.0,
"cumsum_promotion": 0.0,
"cumsum_raise": 0.0,
"e": 1.0,
"id": 2.0,
"start": 0.0,
"stop": 4.0,
},
{
"cumsum_movement": 0.0,
"cumsum_promotion": 0.0,
"cumsum_raise": 0.0,
"e": 0.0,
"id": 3.0,
"start": 0.0,
"stop": 3.0,
},
{
"cumsum_movement": 1.0,
"cumsum_promotion": 1.0,
"cumsum_raise": 0.0,
"e": 0.0,
"id": 3.0,
"start": 3.0,
"stop": 7.0,
},
{
"cumsum_movement": 1.0,
"cumsum_promotion": 1.0,
"cumsum_raise": 1.0,
"e": 1.0,
"id": 3.0,
"start": 7.0,
"stop": 8.0,
},
{
"cumsum_movement": None,
"cumsum_promotion": None,
"cumsum_raise": None,
"e": 1.0,
"id": 4.0,
"start": 0.0,
"stop": 4.0,
},
]
)
assert_frame_equal(expected, ldf, check_dtype=False, check_like=True)
def test_error_is_raised_if_columns_are_missing_in_seed_df(self, seed_df, cv1):
del seed_df["start"]
with pytest.raises(IndexError):
utils.add_covariate_to_timeline(seed_df, cv1, "id", "t", "E")
def test_error_is_raised_if_columns_are_missing_in_seed_df(
self, seed_df, cv1):
del seed_df['start']
with pytest.raises(IndexError):
utils.add_covariate_to_timeline(seed_df, cv1, 'id', 't', 'E')
def test_error_is_raised_if_columns_are_missing_in_seed_df(self, seed_df, cv1):
del seed_df['start']
with pytest.raises(IndexError):
utils.add_covariate_to_timeline(seed_df, cv1, 'id', 't', 'E')
{"id": 4, "z": 0, "time": 5},
{"id": 7, "z": 1, "time": 5},
{"id": 8, "z": 0, "time": 5},
{"id": 5, "z": 0, "time": 5},
{"id": 9, "z": 1, "time": 5},
{"id": 10, "z": 1, "time": 5},
{"id": 4, "z": 0, "time": 6},
{"id": 7, "z": 1, "time": 6},
{"id": 8, "z": 0, "time": 6},
{"id": 5, "z": 1, "time": 6},
{"id": 9, "z": 1, "time": 6},
{"id": 10, "z": 1, "time": 6},
{"id": 7, "z": 1, "time": 7},
{"id": 8, "z": 0, "time": 7},
{"id": 5, "z": 1, "time": 7},
{"id": 9, "z": 1, "time": 7},
{"id": 10, "z": 1, "time": 7},
{"id": 8, "z": 0, "time": 8},
{"id": 5, "z": 1, "time": 8},
{"id": 9, "z": 1, "time": 8},
{"id": 10, "z": 1, "time": 8},
{"id": 9, "z": 1, "time": 9},
{"id": 10, "z": 1, "time": 9},
])
dfcv = add_covariate_to_timeline(df, cv, "id", "time", "event", add_enum=False)
"time": 8
},
{
"id": 5,
"z": 1,
"time": 8
},
{
"id": 9,
"z": 1,
"time": 8
},
{
"id": 10,
"z": 1,
"time": 8
},
{
"id": 9,
"z": 1,
"time": 9
},
{
"id": 10,
"z": 1,
"time": 9
},
])
dfcv = add_covariate_to_timeline(df, cv, "id", "time", "event", add_enum=False)
# Prepare DataFrame for lifelines analysis
#------------------------------------------------------------------------------
lf = to_long_format(df, 'age_at_exit')
# Piped version (possibly faster?)
# lf = df.pipe(to_long_format, 'age_at_exit')\
# .pipe(add_covariate_to_timeline(fund_cv_amt,
# 'id', 'time_to_funding', 'success',
# cumulative_sum=False)\
# NOTE WARNING THIS LINE IS SUPER F*****G SLOW.
# Add raised_amount_usd as time-varying covariate
fund_cv_amt = fund_cv[['id', 'raised_amount_usd', 'time_to_funding']]
lf = add_covariate_to_timeline(lf,
fund_cv_amt,
'id',
'time_to_funding',
'success',
cumulative_sum=False)
# Add cumulative funding as covariate
# lf = add_covariate_to_timeline(lf, fund_cv_amt,
# 'id', 'time_to_funding', 'success',
# cumulative_sum=True)
# Add funding round type as covariate
# WRITE TO PICKLE FILE!!!
# lf.to_pickle('../data/survival_input.pkl')
#==============================================================================
#==============================================================================
