Reading long articles are a hassle; they are boring, and most of the information is actually unnecessary. This can make researching or just reading the news painstakingly long and off-putting. Even worse, referencing while strictly adhering to the MLA or the APA referencing style can also be a thorn in your side.

What it does

Our telegram bot has 3 main functions

  • Summarise
  • Cross Reference
  • Reference Generator

Our telegram bot summarises any article or text significantly, ensuring a no-frills reading experience. references**. In case you do not know the meaning of a word in the article, you can ask the bot to define it for you.

You can also cross-reference multiple articles and see how relevant/similar they are to one another.

Additionally, it converts the URLs into MLA/APA references that you can slot right into your paper or report. All this to make sure that you can put your energy into things that really matter.

How I built it

The telegram bot was built using Python. Various libraries such as nltk and gensim were also utilized. Pycharm was the IDE of choice.

Challenges I ran into

  • Working around the character and flooding limits in Telegram
  • Cross-reference function very difficult to implement
  • Various overflow errors
  • Server-side issues
  • Choosing an appropriate threshold in cross-referencing

Accomplishments that I'm proud of

  • Successfully deployed a Telegram bot that was able to fetch the contents of web pages and provide a summary that was around 20 percent of the original length
  • Successfully implemented functions to generate the MLA and APA references for a list of URLs
  • Successfully implemented a function to cross-reference across multiple URLs
  • Successfully implemented a define word function

What I learned

  • Applying external python librarires
  • Making a Telegram Bot
  • Adding command words to a Telegram Bot
  • Deploying a Telegram Bot

What's next for Project Assistant Bot

  • A bias detector for articles
  • Fake news detector

Built With

Share this project: