- Automated online Treasure hunt
Word-Hunter is a question suggester for Online Treasure Hunt games (i.e .. this automatically prepares clues using online searches extensively) . All you'll have to do is to give it a Topic of Interest , and it'll search for different words relating to the topic and automatically set a question based on the frequency of words and other factors .
.... and here is what your question would look like :
MVC
anddevelopment
are word clues- It gives us an idea that the answer could be a MVC framework
- The visual clue is the picture of a
Flask
- At this point , the visual clue doesn't make much sense
- On hovering over the picture , the title
Python
pops up- Alas , all of these words ,
MVC , development , Flask , Python
seem to make sense !- Hence the final answer has to be a python MVC framework , which is
Django
- Hence the final answer has to be a python MVC framework , which is
- Alas , all of these words ,
- OS : All Linux distributions
- Core language : python 2.7.x
- Google custom search API (For both images and words)
Please note that , using the google custom search API , you can make only 100 calls per day
and currently , 8 calls are made per execution
This can easily be implemented on any python server (To host similar competitions) .
You can also extract out some kind of a suggestion in the form of words , in case you're setting questions for an online treasure hunt competition .
$ git clone https://github.com/shravan97/WordHunter
$ cd WordHunter
$ python wordhunt.py
- Include a loading bar , when the fetched URLs are being scrapped and when frequency is being calculated
- Better way to filter out the massive amount of noise that comes along with search words
- Add pictures and gifs of a sample execution
- Better word suggester (Both interms of filtering and its frequency)
- Search term used for image search needs to be automatically found
- Implement multiprocessing in parts which are slower to run
- Program will have to terminate smoothly on keyboard interrupts and other hindrances
- Instead on taking a topic everytime as input , try taking in as a bunch initially and print questions with a random topic of interest
- Give weightage to the page on which the search result appears on querying
Got any cool idea that could be implemented here ? Go ahead and give a pull request . Or You may as well put it up as an issue here . Please read this document here for more details about contributing
[Please don't forget update the contributors' list below when you're giving a pull request 😄 ]