dcyphr

dcyphr is a web application built in Flask/SQL/HTML/JS. It is a crowd-sourced platform informed by NLP to make academic research more accessible. Learn more

• Register and login
• Request articles to be distilled
• Browse knowledge base through Search or Explore
• Interact with articles by liking, endorsing, sharing, commenting, raising issues.

Installation

Clone this repositiory git clone https://github.com/dcyphr/dcyphr.git. Navigate to the repository on your local machine.

Set up a python3 virtual environment.

python3 -m venv myenv
source myenv/bin/activate

Install requirements pip install -r requirements.txt

Run the setup script

source ./setup.sh

You should have the Flask app running on your local environment!

Getting started

Each route is defined by an @app.route and routes that require the user to be logged in have the decorator @login_required

db.execute is used to run an SQL query on the database and uses string interpolation through setting a variable in the string by preceding it with a colon and then defining the variable as a second, third, etc. argument in the db.execute method.

TODO

• Train NLP on PLOS dataset
• Integrate NLP feature for automatic distillation
  1. Article URL -> distillation
  2. Article DOI -> distillation
  3. Article PDF -> distillation
  4. Ability to request human pruning or mark as correct
• Generate wrapper for NLP to periodically retrain on new data
• Integrate Stripe payment and create premium features
• React overhaul