FacebookTemplate.py

Go to the documentation of this file.

import re
import awswrangler as wr
from pandas import to_datetime
 
class FacebookTemplate:
    def preprocessing(self, filename, features, session, rel_col):
        """Loads file and do some fixes to fit it to our standard and make it able to be processed
 
        Args:
            filename (str): s3 full path of filename
            features (dict): contains delimiter, skip_rows and encoding required for current file
            session (boto3 obj): AWS client connection
            rel_col (str): column name used to identify upc/release_id
        Returns: df (pandas dataframe)
        """
        delimiter = features["delimiter"]
        skip_rows = features["skip_rows"]
        encoding = features["encoding"]
        df = wr.s3.read_csv(filename, sep=delimiter, skiprows=skip_rows, encoding=encoding, low_memory=False, boto3_session=session)
        df["event_count"].fillna(0, inplace=True)
        df["event_count"] = df["event_count"].astype("int")
        df["start_date"] = to_datetime(df["start_date"], format='%Y/%d/%m')
        # df["Units"].fillna(0, inplace=True)
        return df
 
    def date(self, filename):
        """Sets date column given the filename (it contains the date)
           For example GYROstream_Pty_Ltd_Facebook-AL-Production_Usage-Report_202203.csv
        Args:
            filename (str): current filename
        Returns: date_str (str)
        """
        # GYROstream_Pty_Ltd_Facebook-AL-Production_Usage-Report_202203.csv
        date = re.findall(r'Report_\d{6}', filename)[0]
        date = date.replace("Report_", "")
        date_str = date[:4]+"-"+date[4:]+"-01"
        # df["date_from_file"] = date_str
        return date_str