BeautifulSoup on calibre

[thiago.eec]

Hi, everyone.

How can I get the attribute of a tag using BeautifulSoup?

I was trying this:

from calibre.ebooks.BeautifulSoup import BeautifulStoneSoup

snippet = "<section xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" epub:type="epigraph">"

soup = BeautifulSoup(snippet)
tag = soup.contents[0]

if 'epub:type' in tag.attrs:
   epub_type = tag['epub:type']

But it doesn't work. The condition is being evaluted as false.

How should I look for the attribute?

[kovidgoyal]

dont use beautiful soup, use lxml.

from calibre.ebooks.oeb.polish.parsing import parse
root = parse(binary_data)

[thiago.eec]

Originally Posted by kovidgoyal

dont use beautiful soup, use lxml.

Thanks, Kovid.

I tried selecting an attribute, but still doesn't work:

from calibre.ebooks.oeb.polish.parsing import parse

snippet = "<section xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" epub:type="epigraph">"
root = parse(snippet)

if 'epub:type' in root.attrib:
    epub_type = root.attrib['epub:type']

I want epub_type to read the 'epigraph' attribute.

[thiago.eec]

To better explain:

I am reading from a json file, a snippet like this:

Json:

"html": "<section xmlns=\"http://www.w3.org/1999/xhtml\" xmlns:epub=\"http://www.idpf.org/2007/ops\" epub:type=\"cover\">"

The value for 'html' varies, so the element is not always 'section'.

Now, on the main script, I want to check if the html snippet has an 'epub:type' attribute. If it does, I want to save it to the 'epub_type' variable.

[Doitsu]

@thiago.eec Note that Calibre comes with BeautifulSoup 3.0.5. (The current version is 4.4.)

For BeautifulSoup 3.0.5 you'll have to slightly change your code:

from calibre.ebooks.BeautifulSoup import BeautifulSoup
snippet = '"<section xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" epub:type="epigraph">"'
soup = BeautifulSoup(snippet)
if soup.section.has_key('epub:type'):
    epub_type = soup.section['epub:type']

[thiago.eec]

Thanks, @Doitsu

This worked!