def select_series(input_df, kunag=500057582, matnr=103029):
""" selects a series corresponding to the given kunag and matnr """
output_ts = pd.DataFrame()
output_ts = input_df[(input_df["kunag"] == kunag)
& (input_df["matnr"] == matnr)]
output_ts = remove_negative(output_ts)
return output_ts
def frequency(input_df):
"""
returns number of datapoints in last year
:param input_df:
:return:
"""
input_df = transformation.remove_negative(input_df)
latest_date = input_df.index[-1]
latest_year = latest_date.year - 1
latest_month = latest_date.month
latest_day = latest_date.day
last_year_date = datetime(latest_year, latest_month, latest_day)
output_df = input_df[last_year_date:]
freq = output_df.shape[0]
return freq
def normalized_frequency(input_df):
""" returns number of datapoints in last year with normalization
input: a ts with a particular kunag and matnr
output: an integer
"""
input_df = remove_negative(input_df)
input_df["parse_date"] = pd.to_datetime(input_df["date"], format="%Y%m%d")
input_df = input_df.sort_values("parse_date")
if input_df.shape[0] <= 5:
return 0
latest_date = input_df["parse_date"].iloc[-1]
first_date = input_df["parse_date"].iloc[0]
len_in_days = (latest_date - first_date).days
if len_in_days <= 365:
return -1
if len_in_days <= 547:
return -2
latest_year = latest_date.year-1
latest_month = latest_date.month
latest_day = latest_date.day
last_year_date = datetime(latest_year, latest_month, latest_day)
freq = input_df[(input_df["parse_date"] >= last_year_date) & (input_df["parse_date"] <= latest_date)].shape[0]
return freq
if __name__ == "__main__":
import matplotlib.pyplot as plt
import transformation
import data_transformation
file_address = "/home/aman/Desktop/CSO_drug/data/raw_data_drug_store_sample_30_2018-12-10.tsv"
dateparse = lambda dates: pd.datetime.strptime(dates, '%Y%m%d')
data = pd.read_csv(file_address,
sep="\t",
parse_dates=['date'],
index_col='date',
date_parser=dateparse)
data = data.sort_index()
data = transformation.remove_negative(data)
ts = transformation.select_series(data, kunag=600142082, matnr=145105)
ts = ts.reset_index()
ts = data_transformation.get_weekly_aggregate(ts)
ts.Timestamp = pd.to_datetime(ts.dt_week, format='%Y-%m-%d')
ts.index = ts.Timestamp
plt.figure(figsize=(12, 8))
plt.plot(ts["quantity"])
plt.show()
plt.savefig("abc.png")
# print(ts)
# print("done")
norm_freq_last_year_all_series(data)
combination_with_freq = pd.read_csv(
"/home/aman/Desktop/CSO_drug/file_generated/freq_last_year_all_series.csv"
)