DBLP Dataset Parser

The repo is forked from IsaacChanghau/DBLPParser with major codebase re-designed and bugs fixed.

This script provides a simple way to convert the XML datafile provided by DBLP Computer Science Bibliography to a user-friendly JSON format.

The script was tested on DBLP screenshot published on 2019-04-29 which has 6,850,920 documents in total.

Installation

pip install lxml

Usage

1. Download dblp.xml.gz and dblp.dtd from DBLP Computer Science Bibliography.
2. Decompress dblp.xml.gz.
3. Run the below script. Make sure that dblp.xml and dblp.dtd are in the same directory.

python main.py --dblp [path_to_dblp.xml] --output [output.json]

Each line of the generated document is a JSON record. An example is shown as below.

{"author": ["Carmen Heine"], "title": "Modell zur Produktion von Online-Hilfen.", "year": "2010", "school": ["Aarhus University"], "pages": ["1-315"], "isbn": ["978-3-86596-263-8"], "ee": ["http://d-nb.info/996064095"], "genre": "phdthesis"}