/paper.miner

Gatech CSE6242 Spring2020 Project:rocket:

Primary LanguagePython

PaperMiner

Authors: Zan Huang, Kaiwen Luo, Jiaji Liu, Kai Li, Maiqi Ding

Overview

PaperMiner is a novel computer science literature search platform for college students. It combines literature search, academic data presentation, interactive visualization, and open-source API for your research.

Data

In this project, sources of research paper includes:
ACL Home Association for Computational Linguistics(ACL)
Computer Vision Foundation(CVF)
Journal of Machine Learning Research(JMLR)
Neural Information Processing Systems Conference(NIPS)

Environment

Backend

arrow(Release v0.15.5)
beautifulsoup4(Version 4.9.0)
bibtexparser(Version 1.0.1)
pysocks(Version 1.7.1)
requests(Release v2.23.0)
nltk(Version 3.5)
six(Version 1.14.0)
Flask(Version 1.1.2)
pandas(Version 1.0.3)

etc. pls check requirements.txt for details and refer to this page if met gensim related issues.

Pls download the word2vec model and other required files from our dropbox folder, put all files inside abstract_search folder on dropbox into local backend folder for abstract based paper/researcher/conference recommendation.

Frontend

We used Node.js for web development, mainly for frontend part of the project, pls have Node.js installed then npm install gulp, finally cd frontend and gulp node, the website would be avaiable on http://localhost:3000.

Usage

Pls install the dependencies first(sorry for the workload), then start the backend/frontend service according to instructions in READMEs. Then test APIs by sending GET requests to port 5000 and enjoy our search/trends/explore services by visiting the website http://localhost:3000 using your browser with internet access.

  1. pls search your interested paper by inputing keywords in our search page.
  2. pls enjoy the data visualizations presented in our trends page for know statistics about conferences and researchers.
  3. pls get paper/researcher/conference recommendations by inputing your paper abstract in our explore page.