DBLP Dataset Parser

It is a python parser for DBLP dataset, the XML format dumped file can be downloaded here from DBLP Homepage.

This parser requires dtd file, so make sure you have both dblp-XXX.xml (dataset) and dblp-XXX.dtd files. Note that you also should guarantee that both xml and dtd files are in the same directory, and the name of dtd file shoud same as the name given in the <!DOCTYPE> tag of the xml file. Such information can be easily accessed through head dblp-XXX.xml command. As shown below

<?xml version="1.0" encoding="ISO-8859-1"?>
<!DOCTYPE dblp SYSTEM "dblp-2017-08-29.dtd">
<dblp>
<phdthesis mdate="2016-05-04" key="phd/dk/Heine2010">
<author>Carmen Heine</author>
<title>Modell zur Produktion von Online-Hilfen.</title>
...

A sample to use the parser:

def main():
    dblp_path = 'dataset/dblp.xml'
    save_path = 'article.json'
    try:
        context_iter(dblp_path)
        log_msg("LOG: Successfully loaded \"{}\".".format(dblp_path))
    except IOError:
        log_msg("ERROR: Failed to load file \"{}\". Please check your XML and DTD files.".format(dblp_path))
        exit()
    parse_article(dblp_path, save_path, save_to_csv=False)  # default save as json format

Some extracted results:

Count the number of all different type of publications:

Count the number of all different attributes among all publications:

Count the number of five different features of articles:

Distribution of published year of articles: