improve logging and add option to disable proxy

• PR234: Adds command line argument -dp or --disableproxy to disable to use of proxy when querying.
• PR261: Improve logging; there is no ts_logger file, logger is initiated in main.py and query.py, loglevel is set via CLI.

Fixed query.py and some small updates

Fixed

• PR304: Fixed query.py by adding 'X-Requested-With': 'XMLHttpRequest' to header value.
• PR253: Fixed Docker build

Added

• PR313: Added example to README (section 2.3.1).
• PR277: Support emojis by adding the alt text of images to the tweet text.

additional tweet attributes

Add new Tweet attributes:

• links, hashtags
• image urls, video_url,
• whether or not it is a reply
• tweet-ID of parent tweet in case of reply,
• list of usernames to who is replied

Deleted some Tweet attributes:

• Tweet.retweet_id
• Tweet.retweeter_username
• Tweet.retweet_userid

additional user info and tweet fields

This version includes some additional fields in the output:

• is_veriried to indicate whether an user has verified status
• is_retweet to indicate whether an tweet is an retweet
• retweeter_username if it is an retweet
• retweeter_userid if it is an retweet
• retweet_id if it is an retweet
• epoch timestamp

In addition it is using billiard for multiprocessing, which makes it possible that twitterscraper is used in Celery.
fake-useragent is used in order to generate random useragent headers.

Add possibility to save to CSV format (and some fixes)

• Users can now save the tweets to a CSV-format, by using the command line arguments "-c" or "--csv"

• The default value of begindate is set to 2006-03-21. The previous value (2017-01-01)
  was chosen arbitrarily and leaded to questions why not all tweets were retrieved.

• By using linspace() instead of range() to divide the number of days into
  the number of parallel processes, edge cases ( p = 1 ) now also work fine.

First release on GitHub.

This is the first release of twitterscraper on github.