I have been inspired to create a bot that calculates which answer is the best answer because I'm tired of losing. Plus, all the other public bots have been taken down (hopefully mine next).
What it does
The application takes an image from HQ Trivia with the question and the three possible answers visible, and outputs the option that it most thinks is most likely the answer.
How I built it
Given an image from HQ Trivia, it extracts the question and three possible answers as Strings from the image using Google's Vision Platform. Then, I use Google's CustomSearch API to search the question in a Google search, in which the top links (defined by user) from the search result are retrieved. The application visits each link and parses through the HTML to find sequence of character matches using rollover hashing. Every time a match is found, it increments a counter, and then the counters are printed out to show which possible answer came up with the most results.
Challenges I ran into
- Compatibility issues with my IDE and JDK version, as well as my JDK with my Grade version.
- Deploying Google Visual Platform
- Implementing Roll-over hashing
- Bringing all three parts of project together (roll-over hashing with query searching with Optical Character Recognition
Accomplishments that I'm proud of
- Using Google's platforms
- Using Google's API
- Parsing JSON data
What I learned
- 3rd party platforms
- 3rd party APIs
- ArrayList.toString() actually includes the brackets and commas
- I'm not that flexible from Sunrise Yoga
What's next for HQ Apptitude
- Grab image straight from the live show
- Deploying it to a website for public