Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Scholar Reader

The user interface, API, and data processing scripts for an augmented PDF reader application.

This repository hosts code for three subprojects: the user interface, API, and data processing scripts. To learn about each of these projects and how to run the code for each of them, see the file in the relevant directory.

Key directories include:

  • api/: the web API that provides data about entities in papers and bounding boxes for those entities.
  • data-processing/: a set of data processing scripts that extract entities and their bounding boxes from papers.
  • ui/: the user interface for the augmented PDF reader.

Quick Start

In order to run the reader app locally (pointing to the production api) you only need to do

cd ui
npm install 
npm start

See ui/ for more details on how to use the reader.

However, if you want to run the api locally as well you'll have to:

  • Follow the instructions in api/ to run the api locally
  • In another terminal, run the reader app in the /ui directory via npm run dev

Style guidelines

The code in this directory roughly follows the following style guidelines:

  • Python: mypy for typing, black for formatting, pylint for linting, and pytest for testing.
    Configurations for each can be found in the VSCode .vscode/settings.json file for Python subprojects.

  • Node: Typescript, with Jest for tests, and Prettier for formatting. For linting Typescript, see the tslint.json specs in the Typescript subprojects.

Styling also roughly follows the guides for React and CSS for Allen AI engineering (see the wiki), where it makes sense to use these guides.



This project is licensed under the Apache License 2.0. It is aggregate software that calls GhostScript, which is distributed under the Affero GNU Public License (AGPL). The GhostScript binaries and source code are available upon request. You can learn more about the GhostScript AGPL License here.