covid-19-discord-bot/lib/parse_data.py

#!/usr/bin/python

import requests
from bs4 import BeautifulSoup
import json
from os.path import exists
from inspect import getmembers, isroutine
from lib.covid_data_lib import init_database, set_data, get_formatted_data, get_top_n_rows

# San Antonio url
sa_data_url = 'https://www.sanantonio.gov/health/news/alerts/coronavirus'


def format_parse_int(num):
    output = ''
    count = 0
    while not num == 0:
        if(count == 3):
            output += ','
            count = 0
        output += f'{num%10}'
        num = int(num/10)
        count += 1
    return output[::-1]


def import_config(path='config/config.json'):
    if(exists(path)):
        try:
            with open(path) as config_file:
                config_dict = json.load(config_file)
        except Exception as e:
            print(f'There was some issue opening and loading the config.\n{e}')
            exit(1)
    else:
        print('Didn\'t find the config file.')
        exit(1)
    return config_dict


def update_data():
    try:
        print('Creating session.')
        session = init_database(import_config())
    except Exception as e:
        session.rollback()
        print(f'There was an error trying to create a database session:\n{e}')
    data_html = requests.get('https://www.worldometers.info/coronavirus/')
    if(data_html.status_code == '200' or data_html.status_code == 200):
        parsed_html = BeautifulSoup(data_html.text, features='html.parser')
        table = parsed_html.find('table', id='main_table_countries_today')
        for row in table.findAll('tr'):
            if(row and row.findAll('td')):
                if(row.find('a')):
                    set_data(
                        session, row.find('a').text, [r.text for r in row.findAll('td')])
                elif(row.findAll('td')[0] and row.findAll('td')[0].text):
                    set_data(
                        session, row.findAll('td')[0].text.replace(':', ''), [r.text for r in row.findAll('td')])


def format_covid_data(columns, data):
    output = ''
    get_values = [attr for attr in getmembers(data, lambda a:not(
        isroutine(a))) if not(attr[0].startswith('__') and attr[0].endswith('__')) and not attr[0].startswith('_') and not attr[0] == 'metadata']
    output += f'{columns[4]}: {get_values[4][1]}\n'
    output += f'{columns[6]}: {format_parse_int(int(get_values[6][1]))}\n'
    output += f'{columns[1]}: {format_parse_int(int(get_values[1][1]))}\n'
    output += f'{columns[8]}: {format_parse_int(int(get_values[8][1]))}\n'
    output += f'{columns[2]}: {format_parse_int(int(get_values[2][1]))}\n'
    output += f'{columns[9]}: {format_parse_int(int(get_values[9][1]))}\n'
    output += f'{columns[0]}: {format_parse_int(int(get_values[0][1]))}\n'
    output += f'{columns[5]}: {format_parse_int(int(get_values[5][1]))}\n'
    output += f'{columns[7]}: {get_values[7][1]}\n'
    return output


def get_covid_data(selection):
    print('Updating data.')
    columns, all_data_query = get_formatted_data(
        init_database(import_config()), selection)
    output = ''
    for data in all_data_query:
        output += format_covid_data(columns, data)
    return output


def get_top_data(number):
    top_n_rows = get_top_n_rows(init_database(import_config()), number + 1)
    output = ''
    count = 0
    for row in top_n_rows:
        if(not count == 0):
            output += f'# {count}\n{row.selection_original}: {format_parse_int(int(row.total_cases))}'
        if(not count == number):
            output += '\n'
        count += 1
    return output


if(__name__ == '__main__'):
    print(get_covid_data())