To run our solution:
$ python queries.py
Example output is below. As you can see from the example output, you don't know which search engine is which until the end of the test. This preserves the double-blind nature of our test. Our solution takes random articles from the internet and tries to find those articles based on the key words from the article title. This provides a uniform and unbiased way to test each search engine. Each of the three search engines is queried and then given a score from 0-10 depending on how high up in the search results our target link is.
Here is an example output:
query 1: assessing google smartphone
Search engine 1: 9
Search engine 2: 0
Search engine 3: 10
query 2: utah random clowns hysteria
Search engine 1: 4
Search engine 2: 10
Search engine 3: 10
query 3: google AI justice
Search engine 1: 10
Search engine 2: 0
Search engine 3: 0
query 4: Rahm Emanuel's Speech violence chicago
Search engine 1: 4
Search engine 2: 10
Search engine 3: 0
query 5: police facial recognition tougher oversight
Search engine 1: 10
Search engine 2: 10
Search engine 3: 10
totals:
{'ask': 30, 'bing': 37, 'yahoo': 30}
Search engines:
1: bing
2: yahoo
3: ask
Winner: bing