SandroLuck

SandroLuck

import random
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns

get_point = lambda: {'Type': random.choice(['Liker', 'Subscriber', 'Like & Sub']),
                     'Age': random.randint(20, 40),
                     'Gender': random.choice(['M', 'W', 'T']),
                     'MemberSince': random.randint(2010, 2020)}
df = pd.DataFrame([get_point() for i in range(1000)])

SandroLuck

# Our ML things
import pytorch_lightning as pl
import torch
from torch.utils.data import DataLoader, Dataset

from captum.attr import IntegratedGradients
from pytorch_lightning import seed_everything
from pytorch_lightning import Trainer
# Visualization
import pandas as pd

SandroLuck

# To read Csvs
import pandas as pd
# To plot Csvs
import seaborn as sns
# Library that helps us deal with Windows/Linux/OSX
import os
# The standard python plotting library that people use in python
# Seaborn is build on top of this library
import matplotlib.pyplot as plt
import matplotlib.animation as animation

SandroLuck

import urllib.request
from time import sleep
import pandas as pd
import bar_chart_race as bcr
from tqdm import tqdm
import re
import cProfile

find_count = re.compile(r'([0-9,]+) (available|commit)')
START_YEAR = 2000

SandroLuck

def search_git_get_count(terms, file_name='freq.csv', language="Python"):
    """
    Collects the amount of function in terms calls for all year from 2000 to 2020
    :param terms: array of terms that we want to aggreate e.g. ["print", "len"]
    :param language: Language we want to search for e.g. "Python"
    :return: Filename of the Dataframe including the results
    """
    function_calls_by_date = []
    print(
        f"Starting to gather the data this will take approx. {(len(terms) * (END_YEAR - START_YEAR + 1)) // 10} minutes"

SandroLuck

#!/usr/bin/env python3

# The following keys you need to get from Twitter: https://developer.twitter.com/en/apps
# NOTE: It is not good practice to store API-KEYS like this (We only do it for simplicity)
API_KEY = "The key you get from twitter"
API_SECRET = "The key you get from twitter"
ACCESS_TOKEN = "The key you get from twitter"
ACCESS_TOKEN_SECRET = "The key you get from twitter"

import tweepy

SandroLuck

  def __init__(self):
      # Authenticate to Twitter
      auth = tweepy.OAuthHandler(API_KEY, API_SECRET)
      auth.set_access_token(ACCESS_TOKEN, ACCESS_TOKEN_SECRET)

      self.api = tweepy.API(auth)
      self.str_to_time = lambda x: datetime.strptime(x, '%a %b %d %H:%M:%S %z %Y')
      self.selection_function = lambda x: int(x['favorite_count']) + int(x['retweet_count'])
      self.last_tweet_time = datetime.now(timezone.utc) - timedelta(minutes=10)
      # run it once now and then every 10 minutes after that

SandroLuck

    def run_bot(self):
        OUR_SEARCH_TERMS = ['Python Programming', 'Numpy', 'Pandas Python', 'Pip Python','Python Software Foundation']
        try:
            found_tweets = []
            for term in OUR_SEARCH_TERMS:
                found_tweets += self.api.search(term, lang='en', result_type='recent', count=1000)
            # remove the unneeded things
            found_tweets = [t._json for t in found_tweets]
            # make sure not to old
            found_tweets = [t for t in found_tweets if self.str_to_time(t['created_at']) > self.last_tweet_time]

SandroLuck

if __name__ == '__main__':
    seed_everything(42)
    device = 'cpu'
    early_stop_callback = EarlyStopping(monitor='val_loss', min_delta=0.00, patience=5, verbose=True, mode='auto')
    model = OurModel().to(device)
    trainer = Trainer(max_epochs=100, min_epochs=1, auto_lr_find=False, auto_scale_batch_size=False,
                      progress_bar_refresh_rate=10, early_stop_callback=early_stop_callback)
    trainer.fit(model)
    save(model.state_dict(), 'Location of our saved model')