ftd_analyzer

import pandas as pd
import numpy as np
from pandas import DataFrame
from datetime import datetime
import os

class Analyze:
    ONE_MILLION = 1_000_000
    ONE_BILLION = 1_000_000_000

    def __init__(self):
        self.ftd_data = None
        self.file_dir = 'files'
        self.tickers_floats = None


    def read_data(self):
        concat_data = pd.DataFrame()
        for name in os.listdir('files/'):
            if name != "tickers_floats.csv":
                data = pd.read_csv(f"{self.file_dir}/{name}", delimiter='|')
                # last two columns are summary data
                concat_data = pd.concat([concat_data, data[0:-2]])
        self.ftd_data = concat_data

    def transform_data(self):
        self.ftd_data = self.ftd_data.rename(columns={
            'QUANTITY (FAILS)': 'Quantity'
        })
        # # Commented stuff was for running web scraping.
        # TODO: Clean up
        self.ftd_data = self.ftd_data.merge(self.read_float(), on='SYMBOL', how='left')
        self.ftd_data['Percentage of Float'] = self.ftd_data.apply(lambda x: x['Quantity'] / x['float'], axis=1)

        self.ftd_data = self.ftd_data[self.ftd_data['PRICE'] != "."]
        self.ftd_data['Total Amount'] = self.ftd_data.apply(lambda x: x['Quantity'] * float(x['PRICE']), axis=1)

        # custom filters
        self.ftd_data = self.ftd_data[self.ftd_data['float'] != -1.0]
        self.ftd_data = self.ftd_data.sort_values(by='Percentage of Float', ascending=False)



    def read_float(self):
        return Analyze.transform_float(pd.read_csv('files/tickers_floats.csv'))


    @staticmethod
    def transform_float(df: DataFrame):
        df = df[(df['float'] != 'UNKNOWN') | (df['float'].notna())]
        df['float'] = df['float'].replace(np.nan, "-1")
        def transform_money(amt):
            # million
            if 'M' in amt:
                return float(amt[:-1])*Analyze.ONE_MILLION
            # billion
            elif 'B' in amt:
                return float(amt[:-1])*Analyze.ONE_BILLION
            return -1

        df['float'] = df['float'].apply(lambda x: transform_money(x))
        return df



    def write_data(self):
        self.tickers_floats.to_csv('files/tickers_floats.csv', index=False)


    @staticmethod
    def do_print(title, content):
        print(f"{title.ljust(40)[:40]}: {content}")


    def run(self):
        # self.read_data()
        # self.transform_data()
        # the above two commands read in and process SEC split data
        # below command just reads in file of the result of the above
        self.ftd_data = pd.read_csv('aggregate_ftds.csv')
        # self.write_data()
        Analyze.do_print("Number of tickers (that also have price)", self.ftd_data['SYMBOL'].unique().shape[0])
        gme = self.ftd_data[self.ftd_data['SYMBOL'] == 'GME']
        Analyze.do_print("Avg % of float, all tickers", self.ftd_data['Percentage of Float'].mean())
        Analyze.do_print("Avg % of float, GME", gme['Percentage of Float'].mean())
        # divide two above for ratio of gme to all
        Analyze.do_print("Avg Net Amount FTD, all tickers", self.ftd_data['Total Amount'].mean())
        Analyze.do_print("Avg Net Amount FTD, GME", gme['Total Amount'].mean())

        print(gme.sort_values(by='Total Amount', ascending=False))

        Analyze.do_print("Number of dates", self.ftd_data['SETTLEMENT DATE'].unique().shape[0])
        Analyze.do_print("GME # of appearances", gme['SETTLEMENT DATE'].unique().shape[0])

if __name__ == "__main__":
    runner = Analyze()
    runner.run()