036_ratios_peg

#!/usr/bin/env python2

import fnmatch, re
import os
import pandas as pd
from dateutil.relativedelta import relativedelta

import modules.config_details as conf_info

DATE_FORMAT = '%Y-%m-%d'
PE_DATE_COLUMN_NAME = 'Date'
EPS_DATE_COLUMN_NAME = 'Date'
csv_column_dateparse = lambda x: pd.datetime.strptime(x, DATE_FORMAT)

def get_ticker_eps(ticker):
  eps_df = pd.read_csv(conf_info.ratios_dir + conf_info.get_eps_filename(ticker), parse_dates=[EPS_DATE_COLUMN_NAME], date_parser=csv_column_dateparse)
  return eps_df

def calculate_peg_series(eps_df, pe_df):
  peg_df = pd.DataFrame(columns=['Date','PEG'])
  for index_pe_df,pe_row in pe_df.iterrows():
    # finding relevant dates for the calculation
    current_pe_date = pe_row[PE_DATE_COLUMN_NAME]
    index_closest_current_eps = eps_df[EPS_DATE_COLUMN_NAME].searchsorted(current_pe_date)
    one_quarter_before_pe_date = current_pe_date + relativedelta(months=-3)
    index_closest_one_quarter_before_eps = eps_df[EPS_DATE_COLUMN_NAME].searchsorted(one_quarter_before_pe_date)

    # getting the data according to the just found indexes
    row_current_eps = eps_df.loc[index_closest_current_eps]
    row_one_quarter_before_eps = eps_df.loc[index_closest_one_quarter_before_eps]
    eps_current = row_current_eps['EPS'].item()
    eps_one_quarter_before = row_one_quarter_before_eps['EPS'].item()

    # calculating the PEG
    eps_growth_rate = (eps_current - eps_one_quarter_before)
    if (eps_growth_rate != 0):
      current_peg = pe_row['PE_median'] / eps_growth_rate
      # storing the PEG in the DataFrame
      peg_df.loc[index_pe_df] = [current_pe_date,current_peg]
    else:
      # TODO something about 0 deltas?
      current_peg = float('NaN')
      print 'No PEG to save as the delta is 0'
    
  return peg_df

def process_pe_ratios(files_dir):
  pe_median_pattern = re.compile(conf_info.pe_median_filename_regex)
  for file_name in os.listdir(files_dir):
    if fnmatch.fnmatch(file_name, conf_info.pe_median_filename_glob):
      current_ticker = pe_median_pattern.search(file_name).group(1)
      print 'Processing P/E file:', file_name, 'for ticker', current_ticker
      current_ticker_eps_df = get_ticker_eps(current_ticker)
      current_ticker_pe_median_df = pd.read_csv(files_dir + file_name, parse_dates=[PE_DATE_COLUMN_NAME], date_parser=csv_column_dateparse)
      
      peg_df = calculate_peg_series(current_ticker_eps_df, current_ticker_pe_median_df)
      
      current_filename_with_path = conf_info.ratios_dir + conf_info.get_peg_filename(current_ticker)
      print 'Saving PEG series to:', current_filename_with_path
      peg_df.to_csv(current_filename_with_path, index=False)


process_pe_ratios(conf_info.ratios_dir)