def get_ext_other_df_combine_first(self, context=None):
df = context['_self']
(nr, nc) = df.shape
if np.random.choice([0, 1]) == 0:
val = self.df_generator.call(
DfConfig(num_cols=nc, num_rows=nr, nan_prob=0.2))
val.columns = df.columns
val.index = df.index
else:
val = self.df_generator.call(
DfConfig(index_levels=df.index.nlevels,
column_levels=df.columns.nlevels,
col_prefix='i1_',
nan_prob=0.2))
if df.index.nlevels == 1:
val.index = pd.Index(
random.sample(set((list(df.index) + list(val.index))),
len(val.index)))
else:
val.index = pd.MultiIndex.from_tuples(
random.sample(set((list(df.index) + list(val.index))),
len(val.index)))
if df.columns.nlevels == 1:
val.columns = pd.Index(
random.sample(set((list(df.columns) + list(val.columns))),
len(val.columns)))
else:
val.columns = pd.MultiIndex.from_tuples(
random.sample(set((list(df.columns) + list(val.columns))),
len(val.columns)))
return val
def get_ext_other_df_ne_like(self, context=None):
df = context['_self']
(nr, nc) = df.shape
cond = self.df_generator.call(
DfConfig(num_rows=nr, num_cols=nc, value_bags=Bags.bool_bags))
cond.columns = df.columns
cond.index = df.index
new_df = self.df_generator.call(DfConfig(num_rows=nr, num_cols=nc))
new_df.columns = df.columns
new_df.index = df.index
return df.where(cond, new_df)
def get_ext_other_df_align(self, context=None):
df = context['_self']
(nr, nc) = df.shape
val = self.df_generator.call(
DfConfig(num_rows=random.choice([max((nr - 1), 1), nr, (nr + 1)]),
num_cols=random.choice([max((nc - 1), 1), nc, (nc + 1)]),
col_prefix='i1_',
index_levels=df.index.nlevels,
column_levels=df.columns.nlevels,
value_bags=[*Bags.int_bags, *Bags.float_bags]))
if (np.random.choice([0, 1]) == 0) and (len(val.index) == nr):
val.index = df.index
elif df.index.nlevels == 1:
val.index = pd.Index(
random.sample(set((list(df.index) + list(val.index))),
len(val.index)))
else:
val.index = pd.MultiIndex.from_tuples(
random.sample(set((list(df.index) + list(val.index))),
len(val.index)))
if (np.random.choice([0, 1]) == 0) and (len(val.columns) == nc):
val.columns = df.columns
elif df.columns.nlevels == 1:
val.columns = pd.Index(
random.sample(set((list(df.columns) + list(val.columns))),
len(val.columns)))
else:
val.columns = pd.MultiIndex.from_tuples(
random.sample(set((list(df.columns) + list(val.columns))),
len(val.columns)))
return val
def get_ext_right_df_merge(self, context=None):
df = context['_self']
new_df: pd.DataFrame = self.df_generator.call(
DfConfig(col_prefix='i1_'))
dg1 = collections.defaultdict(list)
dg2 = collections.defaultdict(list)
for (k, v) in dict(df.dtypes).items():
dg1[v].append(k)
for (k, v) in dict(new_df.dtypes).items():
dg2[v].append(k)
c = (set(dg1.keys()) & set(dg2.keys()))
for dt in c:
cols1 = list(dg1[dt])
cols2 = list(dg2[dt])
random.shuffle(cols1)
random.shuffle(cols2)
pairs = list(zip(cols1, cols2))
for pair in pairs:
if np.random.choice([0, 1]) == 0:
new_df[pair[1]] = random.sample(
(list(new_df[pair[1]]) + list(df[pair[0]])),
new_df.shape[0])
if (np.random.choice([0, 1])
== 0) and (pair[0] not in new_df.columns):
new_df = new_df.rename({
pair[1]: pair[0],
}, axis=1)
return new_df
def get_ext_cond_df_where_mask(self, context=None):
df, nr, nc = context['_self'], context['num_rows'], context['num_cols']
cond_df = self.df_generator.call(
DfConfig(num_rows=nr,
num_cols=nc,
value_bags=Bags.bool_bags,
index_like_columns_prob=0.0))
cond_df.index = df.index
cond_df.columns = df.columns
return cond_df
def get_ext_other_df_combine(self, context=None):
df = context['_self']
(nr, nc) = df.shape
new_df = self.df_generator.call(
DfConfig(num_rows=nr,
num_cols=nc,
value_bags=[*Bags.int_bags, *Bags.float_bags]))
new_df.columns = df.columns
new_df.index = df.index
return new_df
def get_ext_other_df_reindex_like(self, context=None):
df = context['_self']
val = self.df_generator.call(
DfConfig(index_levels=df.index.nlevels,
col_prefix=random.choice(['', 'i1_']),
value_bags=[*Bags.int_bags, *Bags.float_bags]))
if df.index.nlevels == 1:
val.index = pd.Index(
random.sample(set((list(df.index) + list(val.index))),
len(val.index)))
else:
val.index = pd.MultiIndex.from_tuples(
random.sample(set((list(df.index) + list(val.index))),
len(val.index)))
return val
def get_ext_other_df_add_like(self, context=None):
df = context['_self']
(nr, nc) = df.shape
v_nc = random.choice([nc, nc, nc, nc - 1, nc + 1])
new_df = self.df_generator.call(
DfConfig(num_cols=v_nc,
num_rows=nr,
column_levels=df.columns.nlevels,
index_like_columns_prob=0.0,
col_prefix='i1_',
value_bags=[*Bags.int_bags, *Bags.float_bags]))
new_df.index = df.index
if (np.random.choice([0, 1]) == 0) and (len(new_df.columns) == nc):
new_df.columns = df.columns
elif df.columns.nlevels == 1:
new_df.columns = pd.Index(
random.sample(set((list(df.columns) + list(new_df.columns))),
len(new_df.columns)))
else:
new_df.columns = pd.MultiIndex.from_tuples(
random.sample(set((list(df.columns) + list(new_df.columns))),
len(new_df.columns)))
return new_df
def get_ext_other_df_equals(self, context=None):
return self.df_generator.call(
DfConfig(col_prefix=random.choice(['', 'i1_'])))
def get_ext_self_df_duplicate_removal(self, context=None):
return self.df_generator.call(DfConfig(min_height=3))
def get_ext_self_df_computational(self, context=None):
return self.df_generator.call(DfConfig(nan_prob=0.1))
def get_ext_self_df_reorder_levels(self, context=None):
return self.df_generator.call(
DfConfig(multi_index_prob=0.6, multi_col_index_prob=0.4))
def get_ext_self_df_combine_first(self, context=None):
return self.df_generator.call(DfConfig(nan_prob=0.2))
def get_ext_self_df_int_and_floats(self, context=None):
return self.df_generator.call(
DfConfig(value_bags=[*Bags.int_bags, *Bags.float_bags],
index_like_columns_prob=0.0))
def get_ext_other_df_where_mask(self, context=None):
df, nr, nc = context['_self'], context['num_rows'], context['num_cols']
other_df = self.df_generator.call(DfConfig(num_rows=nr, num_cols=nc))
other_df.index = df.index
other_df.columns = df.columns
return other_df
def get_ext_self_df_dropna_fillna(self, context=None):
return self.df_generator.call(DfConfig(nan_prob=0.3))
def get_ext_self_df_pivot(self, context=None):
return self.df_generator.call(DfConfig(column_levels=1))
def get_ext_self_df_all_any(self, context=None):
return self.df_generator.call(
DfConfig(value_bags=[
*Bags.int_bags, *Bags.string_bags, *Bags.bool_bags,
*Bags.bool_bags
]))
def get_ext_self_df_count(self, context=None):
return self.df_generator.call(DfConfig(nan_prob=0.5))
def get_ext_input_df_isna_notna(self, context=None):
return self.df_generator.call(DfConfig(nan_prob=0.5))