Previously, there was a problem with sites that redirect an attempt to view a non-existing username to the main site. For example, if you try to go to https://devrant.com/users/dfoxxxxxxxxx (a user name that does not exist), then we get a redirect to the https://devrant.com/ root of the site. But, the "response_url" checking algorithm was only looking for the configured error URL being included in the response. So, these sites always indicated that the username was not found.
Update the "response_url" detection method so that the request does not allow redirects. If we get a 200 response of some type, then the username has been found. However, if we get something like a 302, then we know that the username was not found as we are being redirected.
This whole method seems fragile, but I did exhaustively test all of the supported sites, and they all work. So, this change is clearly an improvement.
Sherlock has a cool logo so this PR moves it to the top of sherlock.py to show it off when potential contributors read our source. It also makes a BANNER an __r'string'__ so we do not need backslashes to escape characters and the ascii art is more what-you-see-is-what-you-get. This also suppresses [a bunch of flake8 complaints](https://travis-ci.com/TheYahya/sherlock/jobs/172028552#L490-L503) about _invalid escape sequences_. Finally this PR runs isort on the imports for readability. Standard Lib imports come first, etc.
Created SherlockBaseTest() class that provides infrastructure for the tests.
TODO: It seems that devRant is not detecting users that certainly exist.
TODO: While running tests, there is a ResourceWarning from an unclosed SSLSocket. This needs work.
Christopher K. Hoadley
Yahya SayadArbabi
Czechball
cclauss
Jus de Patate
HATI
JakeConnors376W
theodosisathanasakis
Tejasvi (Teju) Nareddy
Czechball
SebastianElvis
Master