Mailman 3 April 2005 - lxml - The Python XML Toolkit

[lxml-dev] off-topic: libxml on OS X
by Marc-Antoine Parent 14 Apr '05

14 Apr '05

BTW, if you need a fresh version of libxml binaries for 10.3, I maintain a generally up-to-date installer of libxml2, libxsl and python bindings at http://maparent.webhop.net/tinderbox/libxml.pkg.tgz Marc-Antoine Parent

1 0

[lxml-dev] etree.parse() works, and it's quick.
by Wade Leftwich 13 Apr '05

13 Apr '05

I have tested etree as a dropin replacement for ElementTree, in an application that parses and transforms xml representations of magazine articles. It seems to work fine on the 300,000 articles in my database: that is to say, no exceptions, **no segfaults**, and 20 or so records pulled at random appear to be correct. With no desire to start a big testing argument going, I can also say that etree.parse() is comparable to cElementTree.parse() in terms of speed. Reading 100,000 xml documents from a MySql db, with the xml moderately complicated and a median length of 4800 bytes, each parser contributes less time to the program than the overhead of running the query. ### In [119]: timeparse(noop, c) query time 78.665211916 sec loop time 1.53149199486 sec 100000 rows In [120]: timeparse(cElementTree.parse, c) query time 105.771811008 sec loop time 39.1069951057 sec 100000 rows In [121]: timeparse(etree.parse, c) query time 78.0830960274 sec loop time 56.8308110237 sec 100000 rows In [122]: timeparse(cElementTree.parse, c) query time 137.834351063 sec loop time 40.2058398724 sec 100000 rows In [123]: In [123]: timeparse(etree.parse, c) query time 101.882040977 sec loop time 34.8395290375 sec 100000 rows # Nice! In [118]: def timeparse(parse, cursor): .....: global query .....: qt0 = time.time() .....: cursor.execute(query) .....: print "query time", time.time() - qt0, "sec" .....: t0 = time.time() .....: i = 0 .....: for row in cursoriter(cursor): .....: f = StringIO(row[0]) .....: doc = parse(f) .....: i += 1 .....: print "loop time", time.time() - t0, "sec" .....: print i, "rows" .....: def noop(*args, **kwargs): """nada""" In [133]: def cursoriter(cursor): .....: while 1: .....: row = cursor.fetchone() .....: if row is None: .....: break .....: yield row .....:

2 3

[lxml-dev] default namespace for xpath
by Tim Parkin 12 Apr '05

12 Apr '05

Hello, You'll have to excuse me if this is a bit of a newbie question but I'm trying to parse an xhtml document with a default namespace and was wondering if there is a simple way of setting this when applying an xpath query? I think that I'm going to have to wait for xpath2 but for the sake of 'ease of use' is there a way of working that will allow me to use '//body' on a document that uses a default namespace xmlns="http://www.w3.org/1999/xhtml"? (or any other default namespace for that matter). At the moment I think that I need to choose a namespace prefix for the 'default' namespace .. eg .xpath('//html:body', {'html': 'http://www.w3.org/1999/xhtml'} ) I would like to be able to do something similar to the xpath extensions where .xpath('//body', {None: 'http://www.w3.org/1999/xhtml'} ) Many thanks for what I think will become a very popular package Tim -- __ / \ Tim Parkin, Pollenation Internet Ltd \__/ http://www.pollenation.net

2 1

[lxml-dev] lxml 0.5.1 released!
by Martijn Faassen 10 Apr '05

10 Apr '05

Hi there, Some critical unicode-related errors were discovered in lxml 0.5, so I've done a 0.5.1 release which fixes them. Enjoy! The release can be found on: http://codespeak.net/lxml Regards, Martijn

1 0

[lxml-dev] etree.Element() segfault
by Wade Leftwich 10 Apr '05

10 Apr '05

Hi, Using lxml 0.5 with Python 2.3.5 and libxml 2.6.16 (Debian Unstable packages), I get a segfault by passing a unicode string containing a high-bit character as an attribute value to the Element() constructor. This does not happen when passing the same value to Element.set(): ###### In [9]:gooduni = u'"straight quotes"' In [10]:baduni = u'\u201ccurly quotes\u201d' In [11]:a = etree.Element('a') In [12]:a.set('goodattr', gooduni) In [13]:etree.dump(a) <a goodattr=""straight quotes""/> In [14]:b = etree.Element('b', {'goodattr':gooduni}) In [15]:etree.dump(b) <b goodattr=""straight quotes""/> In [16]:c = etree.Element('c') In [17]:c.set('badattr', baduni) In [18]:etree.dump(c) <c badattr="“curly quotes”"/> In [19]:d = etree.Element('d', {'badattr':baduni}) Exception exceptions.UnicodeEncodeError: <exceptions.UnicodeEncodeError instance at 0xb7bec28c> in 'etree._createElement' ignored Segmentation fault ### Wade Leftwich Ithaca, NY

2 2

[lxml-dev] lxml 0.5 released
by Martijn Faassen 09 Apr '05

09 Apr '05

Hi there, I've just released lxml version 0.5. A release at last! You can find more information about it, and download, here: http://codespeak.net/lxml/ Regards, Martijn

2 1