Prototype for demonstrating the collection of link and rel in the case of hypermedia APIs.

gethyperel.py

#!/usr/bin/env python2.7
# encoding: utf-8

import sys
import string
from pprint import pprint
from urlparse import urljoin
from lxml.html import html5parser
import argparse

# CONFIG
HTMLNS = "http://www.w3.org/1999/xhtml"


def getURI(uri):
    """Given a URI return the parsed document object"""
    doc = html5parser.parse(uri).getroot()
    return doc


def getarel(doc, uri):
    """return the value of rel for a"""
    hyperdict = []
    # grab all 'a' elements from the document
    alist = doc.xpath('//h:a[@rel]', namespaces={'h': HTMLNS})
    for a in alist:
        rels = a.xpath('@rel')
        # some 'rel' attributes might have a space separated list of values
        for rel in string.split(rels[0]):
            # we add to a list the couple 'rel' and 'href'
            hyperdict.append(dict([
                ('rel', rel),
                ('href', urljoin(uri, a.xpath('@href')[0]))
                ]),
            )
    return hyperdict


def main():
    parser = argparse.ArgumentParser(description="parse a Web site and search for rel values on a")
    parser.add_argument('uri', help='URI to be processed', action='store', nargs=1)
    args = parser.parse_args()

    uri = args.uri[0]
    document = getURI(uri)
    relations = getarel(document, uri)
    pprint(relations)

if __name__ == "__main__":
    sys.exit(main())