dblp-sax-parser

DBLP SAX Parser

• DBLP SAX Parser
  • What is it?
  • Context and Purpose
  • Usage
  • DBLP Methods
  • License
  • References

What is it?

A parsing package using the Simple API for XML (SAX).

There are a total of 10 elements: "article", "inproceedings", "proceedings", "book", "incollection", "phdthesis", "mastersthesis", "www", "person", "data".

Across the elements, these are the feature types available: "address", "author", "booktitle","cdrom", "chapter", "cite", "crossref", "editor", "ee", "isbn", "journal", "month", "note", "number", "pages", "publisher", "publnr", "school", "series", "title", "url", "volume", "year".

Features

• download dblp files from the dblp website directly
• parse throught the dblp xml file into a dataframe, exported with either csv or pickle format.

Future features for consideration

• add more methods to parse data from a specific attribute. E.g. only for years in 2016
• select which elements or features to be included/excluded

Context and Purpose

I created this package when working on a project as part of a course module. The aim of this package is to provide a quick way to parse DBLP elements directly, with the contents exported as a csv file for further preprocessing based on individual's use case.

Installation

pip install dblp-sax-parser

# import package
from dblp_parser import DBLP_Parser as dp

Usage

First step to using this parser is to instantiate the dblp_parser

# Instantiate the dblp class 
dblp = dp()

You can also DBLP_Parser to download the dblp data assets from the dblp website

# download latest data sets from dblp website
dblp.download_latest_dump()

Parsing the xml file

filename = 'dblp.xml'

# execute the parser from the dblp class
parser, handler = dblp.execute_parser(filename=<filename>)

# you can use the handler to convert the handler output to dataframe
handler.to_df()

# the dataframe can be persisted as a pickle file or exported as csv file
handler.to_csv() # export to csv
handler.save() # persist as pickle

DBLP Methods

class DBLP_Parser

• This is the main class to be instantiated when before using the parser

class DBLP_Parser.download_latest_dump

• Begins downloading the latest dblp files from the dblp website. If the url location where files are hosted is changed/incorrect, a separate url can be used instead.
• This downloads the dblp .dtd and .xml.gz files, and decompress the .gz file into .xml.
• dtd_url[str]: url location of the .dtd file to be downloaded from.
• xml_zip_url [str]: url of the .xml.tz file to be downloaded from.
• xml_zip_filename [str]: specify filename of the downloaded .xml.gz file.
• xml_filename [str]: specify filename of the .xml file that is decompressed.

class DBLP_Parser.execute_parser

• This executes the underlying SAX parser, calling the xml.sax.handler.ContentHandler
• filename [str]: path and name of XML file to be parsed. If **download_latest_dump() was used, the file to be parsed will be "dblp.xml".

License

This code is published under the MIT licence.

References

There are two main references that helped contributed to writing this package. Instantiating the outer dblp class to download dblp materials directly came from from angelosalatino. Some component of the SAX parsing logic itself was borrowed from hibernator11.