Program for answering trivia questions from the popular mobile game HQ with Bing Web Search API, MediaWiki API, and NLTK in Python 3.
If you don't already have pip3 installed, you'll need to run the following command in your terminal before proceeding:
$ sudo apt-get install python3-pip
You'll need to install the following python modules:
$ pip3 install nltk
$ pip3 install beautifulsoup4
$ pip3 install requests
$ pip3 install markdown
$ pip3 install html5lib
$ pip3 install urllib3
The following command runs the Bing Search Model and Wiki Search Model on sample questions located in /src/input_files and outputs the results:
$ python3 test.py
Bing Search Model
The Bing Search Model uses Microsoft Azure's Bing Web Search API to retrieve a search term's top 5 results, and then counts how many times each of the answer choices appear across the 5 pages.
Wikipedia Answer Model
The Wikipedia Answer Model uses MediaWiki's Web API to retrieve JSON pages of Wikipedia articles for each of the answer choices, and then counts the number of times the key words co-occur within a 10-15 word window.
Question 1
Q: Which of these is a standard cheerleading jump?
A: ['Herkie', 'Flap', 'Striker']BING SEARCH MODEL
Bing Guess: Herkie
Guess is correct!
Bing Search Model took 2.7611379623413086 secondsWIKI SEARCH MODEL
Wiki Guess: Herkie
Guess is correct!
Wiki Search Model took 0.8776249885559082 seconds
Question 2
Q: Which of these companies is NOT owned by Williams-Sonoma, Inc.?
A: ['Pottery Barn', 'West Elm', 'Crate & Barrel']BING SEARCH MODEL
Bing Guess: Crate & Barrel
Guess is correct!
Bing Search Model took 4.79827880859375 secondsWIKI SEARCH MODEL
Wiki Guess: Crate & Barrel
Guess is correct!
Wiki Search Model took 1.7192120552062988 seconds
Question 3
Q: Which city’s busiest airport is NOT named after an elected politician?
A: ['Oklahoma City', 'Houston', 'Las Vegas']BING SEARCH MODEL
Bing Guess: Oklahoma City
Guess is correct!
Bing Search Model took 4.7103049755096436 secondsWIKI SEARCH MODEL
Wiki Guess: Oklahoma City
Guess is correct!
Wiki Search Model took 1.8565549850463867 seconds
Question 4
Q:Which of these mammals averages the largest litter?
A: ['Naked mole rat', 'Burmese cat', 'Jackrabbit']BING SEARCH MODEL
Bing Guess: Naked mole rat
Guess is correct!
Bing Search Model took 8.854448318481445 secondsWIKI SEARCH MODEL
Wiki Guess: Naked mole rat
Guess is correct!
Named Entity Recognition:
Question = "Which of these agencies began as the North West Police Agency?"
NER Output = {"North West Police Agency" - organization}.
Open Information Extraction:
Question = "In which college class would you be most likely to study a syzygy?"
OpenIE Output = {"you" /subj, "study" /relation, "syzygy" /obj}
Grammatical Dependencies:
Dependencies = -> likely/JJ (root)
-> class/NN (nmod)
-> In/IN (case)
-> which/WDT (det)
-> college/NN (compound)
-> would/MD (aux)
-> you/PRP (nsubj)
-> be/VB (cop)
-> most/RBS (advmod)
-> study/VB (xcomp)
-> to/TO (mark)
-> syzygy/NN (dobj)
-> a/DT (det)
-> ?/. (punct)
Syntactic Parses:
Parse tree = (ROOT (SBARQ (WHPP (IN In) (WHNP (WDT which) (NN college) (NN class))) (SQ (MD would) (NP (PRP you)) (VP (VB be) (ADJP (RBS most) (JJ likely) (S (VP (TO to) (VP (VB study) (NP (DT a) (NN syzygy)))))))) (. ?)))
I'd like to thank Professor Tim Hunter for guiding me with the project and providing key insights necessary to complete the project as part of UCLA's Student Research Program 199.