def filter_traces_intersecting(df, dt1, dt2, parameters=None):
"""
Filter traces intersecting the given interval
Parameters
----------
df
Pandas dataframe
dt1
Lower bound to the interval (possibly expressed as string, but automatically converted)
dt2
Upper bound to the interval (possibly expressed as string, but automatically converted)
parameters
Possible parameters of the algorithm, including:
Parameters.TIMESTAMP_KEY -> Attribute to use as timestamp
Parameters.CASE_ID_KEY -> Column that contains the timestamp
Returns
----------
df
Filtered dataframe
"""
if parameters is None:
parameters = {}
timestamp_key = exec_utils.get_param_value(Parameters.TIMESTAMP_KEY,
parameters,
DEFAULT_TIMESTAMP_KEY)
case_id_glue = exec_utils.get_param_value(Parameters.CASE_ID_KEY,
parameters, CASE_CONCEPT_NAME)
dt1 = get_dt_from_string(dt1)
dt2 = get_dt_from_string(dt2)
needs_conversion = check_pandas_ge_024()
if needs_conversion:
dt1 = dt1.replace(tzinfo=pytz.utc)
dt2 = dt2.replace(tzinfo=pytz.utc)
dt1 = pd.to_datetime(dt1, utc=True)
dt2 = pd.to_datetime(dt2, utc=True)
grouped_df = df[[case_id_glue, timestamp_key]].groupby(df[case_id_glue])
first = grouped_df.first()
last = grouped_df.last()
last.columns = [str(col) + '_2' for col in last.columns]
stacked = pd.concat([first, last], axis=1)
stacked1 = stacked[stacked[timestamp_key] > dt1]
stacked1 = stacked1[stacked1[timestamp_key] < dt2]
stacked2 = stacked[stacked[timestamp_key + "_2"] > dt1]
stacked2 = stacked2[stacked2[timestamp_key + "_2"] < dt2]
stacked3 = stacked[stacked[timestamp_key] < dt1]
stacked3 = stacked3[stacked3[timestamp_key + "_2"] > dt2]
stacked = pd.concat([stacked1, stacked2, stacked3], axis=0)
i1 = df.set_index(case_id_glue).index
i2 = stacked.set_index(case_id_glue).index
return df[i1.isin(i2)]
def convert_timestamp_columns_in_df(df,
timest_format=None,
timest_columns=None):
"""
Convert all dataframe columns in a dataframe
Parameters
-----------
df
Dataframe
timest_format
(If provided) Format of the timestamp columns in the CSV file
timest_columns
Columns of the CSV that shall be converted into timestamp
Returns
------------
df
Dataframe with timestamp columns converted
"""
needs_conversion = check_pandas_ge_024()
for col in df.columns:
if timest_columns is None or col in timest_columns:
if df[col].dtype == 'object':
try:
if timest_format is None:
if needs_conversion:
df[col] = pd.to_datetime(df[col], utc=True)
else:
df[col] = pd.to_datetime(df[col])
else:
if needs_conversion:
df[col] = pd.to_datetime(df[col],
utc=True,
format=timest_format)
else:
df[col] = pd.to_datetime(df[col])
except ValueError:
# print("exception converting column: "+str(col))
pass
return df
def apply_events(df, dt1, dt2, parameters=None):
"""
Get a new log containing all the events contained in the given interval
Parameters
----------
df
Pandas dataframe
dt1
Lower bound to the interval (possibly expressed as string, but automatically converted)
dt2
Upper bound to the interval (possibly expressed as string, but automatically converted)
parameters
Possible parameters of the algorithm, including:
Parameters.TIMESTAMP_KEY -> Attribute to use as timestamp
Returns
----------
df
Filtered dataframe
"""
if parameters is None:
parameters = {}
timestamp_key = exec_utils.get_param_value(Parameters.TIMESTAMP_KEY,
parameters,
DEFAULT_TIMESTAMP_KEY)
dt1 = get_dt_from_string(dt1)
dt2 = get_dt_from_string(dt2)
needs_conversion = check_pandas_ge_024()
if needs_conversion:
dt1 = dt1.replace(tzinfo=pytz.utc)
dt2 = dt2.replace(tzinfo=pytz.utc)
dt1 = pd.to_datetime(dt1, utc=True)
dt2 = pd.to_datetime(dt2, utc=True)
df = df[df[timestamp_key] > dt1]
df = df[df[timestamp_key] < dt2]
return df
def get_dt_from_string(dt: Union[datetime, str]) -> datetime:
"""
If the date is expressed as string, do the conversion to a datetime.datetime object
Parameters
-----------
dt
Date (string or datetime.datetime)
Returns
-----------
dt
Datetime object
"""
if type(dt) is str:
dt = datetime.strptime(dt, "%Y-%m-%d %H:%M:%S")
needs_conversion = check_pandas_ge_024()
if needs_conversion:
dt = dt.replace(tzinfo=pytz.utc)
dt = pd.to_datetime(dt, utc=True)
return dt