AutoGraph-Obsidian: Automated Knowledge Graph Generation.

Introduction

AutoGraph is a tool that enables rapid, automated knowledge graph generation. AutoGraph does this by mining scientific literature against a search query for keywords. From this data, an Obsidian vault is made where each mined keyword has its own markdown file containing: i) the name of the paper the keyword was scraped from and ii) links to other keywords from that paper. When two papers share a keyword, a link is established between those articles through that term. Over many papers, this allows a network of interactions between articles in a field to be visualized. The purpose of this tool is not only to establish graph-based summaries of topics but also to identify hidden links between divergent fields - largely inspired by the works of Manfred Kochen.

Prerequisites:

Python >= 3.6
Obsidian.md

Quickstart:

pip3 install autograph-obsidian

Usage:

Usage: autograph [OPTIONS] QUERY

  Arguments:

  QUERY The main search string.

Options:
  -l, --limit INTEGER  Number of papers to mine. Default = 500.
  -v, --version        Show version number and exit.
  --help               Show this message and exit.

e.g.

autograph 'Genetic Code Expansion' -l 100

Case Study

Generating the graph with autograph

Viewing the graph with Obsidian.md

Acknowledgements

The mining of scientific literature is handled by the pygetpapers package developed by ContentMine.