A web app to find new words given a relationship between two words.
The vector representation of 50,000 words are loaded when the web app is started. Then whenever a word relation is submitted such as: relation_word1, relation_word2 and target_word, the word of the closest vector to the vector calculation relation_word1 - relation_word2 + target_word is given to the user.
Highly encouraged to create a python environment first.
Clone the repository:
$ git clone https://github.com/xemeds/vocabulus.git
Move into the cloned folder and install the required libraries:
$ cd vocabulus
$ pip install -r requirements.txt
After that run with:
$ python run.py
Visit the below URL to view the flask app:
127.0.0.1:5000
If you do not have a dedicated server, I highly recommend using Linode, Heroku or PythonAnywhere to host your application.
Before deploying, make sure to set the following environment variables:
$ export SECRET_KEY=
$ export DEBUG=
If not they will default to the following values:
SECRET_KEY=SECRET_KEY
DEBUG=true
The vec2word wordlist: CS50AI/Lecture 6/vectors
This project is under the MIT license.