Mailman 3 March 2011 - lxml - The Python XML Toolkit

Re: [lxml] Error with validating xml file against schema
by Stefan Behnel March 27, 2011

March 27, 2011

Wu, Hua, 25.03.2011 13:22: > I am trying using lxml to validate xml files against opengis gml 3.1.1 > schema but I got an error.I googled for solutions, I only found > similar errors out there but no solutions.Could anyone help me with > this error? I use lxml 2.3 and python 2.7. > > > code snippet: > > from lxml import etree > > schema = etree.parse("http://www.isotc211.org/2005/gml/gml.xsd") > xmlschema = etree.XMLSchema(schema) > > I got: > Error : local list type: A type, derived by list or union, must have > the simple ur-type definition as base type, not > '{http://www.opengis.net/gml}doubleList'., line 1 > > > Others run xmllint got similar error > http://gis.stackexchange.com/questions/6721/why-doesnt-the-gml-3-1-1-s > chema-validate-with-xmllint > > Someone thought this could be a bug, does this make sense? > http://www.mail-archive.com/xml@gnome.org/msg07477.html Yes, likely a bug in libxml2. XML Schema support is known to have its rough edges. Could you check if there's a ticket for this on the libxml2 bug tracker? https://bugzilla.gnome.org/buglist.cgi?product=libxml2 Stefan

1 0

Error with validating xml file against schema
by Wu, Hua March 25, 2011

March 25, 2011

> Hi, > > I am trying using lxml to validate xml files against opengis gml 3.1.1 > schema but I got an error.I googled for solutions, I only found > similar errors out there but no solutions.Could anyone help me with > this error? I use lxml 2.3 and python 2.7. > > > code snippet: > > from lxml import etree > > schema = etree.parse("http://www.isotc211.org/2005/gml/gml.xsd") > xmlschema = etree.XMLSchema(schema) > > I got: > Error : local list type: A type, derived by list or union, must have > the simple ur-type definition as base type, not > '{http://www.opengis.net/gml}doubleList'., line 1 > > > Others run xmllint got similar error > http://gis.stackexchange.com/questions/6721/why-doesnt-the-gml-3-1-1-s > chema-validate-with-xmllint > > Someone thought this could be a bug, does this make sense? > http://www.mail-archive.com/xml@gnome.org/msg07477.html > Regards Hua

1 0

Error loging of lxml xonsume all disk space
by Андрей Никишаев March 24, 2011

March 24, 2011

Hello, I have problem with lxml. Fore some reason sometimes it consumes all server memory when parsing data. I using it in web-spider(daemonized) so it's not so cool. So i set ulimit -m 600000 and ulimit -v 500000 for spider process. After this lxml start write data to error log, till there is free space on disk. So i get this error: "Exception MemoryError: MemoryError() in 'lxml.etree._BaseErrorLog._receive' ignored" One more trouble that i can't handle this exception in my code. To make this stop i have monkeypatched lxml.etree._BaseErrorLog but it's not helped. Can someone tell me how i can handle this exception, and how i can make lxml to not write at error log bypassing the python logging module? Thanks for help. Best regards, Andrey Nikishaev Blog http://creotiv.in.ua/ Skype creotiv.in.ua

2 1

Iterparse error
by Roger Hoover March 23, 2011

March 23, 2011

Hi, I compiled lxml version 2.3 on Centos 5 and get this error trying to parse anything with iterparse. Any ideas what the problem could be? File "lxml_test.py", line 6, in <module> context = etree.iterparse("foo.xml", events=('end,'), tag='event') File "iterparse.pxi", line 421, in lxml.etree.iterparse.__init__ (src/lxml/lxml.etree.c:94766) File "parser.pxi", line 766, in lxml.etree._BaseParser._getPushParserContext (src/lxml/lxml.etree.c:73704) File "iterparse.pxi", line 450, in lxml.etree.iterparse._createContext (src/lxml/lxml.etree.c:94972) File "iterparse.pxi", line 123, in lxml.etree._IterparseContext._setEventFilter (src/lxml/lxml.etree.c:92127) File "iterparse.pxi", line 30, in lxml.etree._buildIterparseEventFilter (src/lxml/lxml.etree.c:91456) ValueError: invalid event name 'e' Thanks, Roger

2 3

whitespace and the HTMLParser
by Tim Arnold March 22, 2011

March 22, 2011

Hi, I don't understand how whitespace is handled and I want to make sure the following is the desired behavior. The first tree= statement uses the HTMLParser which drops the space between the <img/> and the <a/> tags. If I do not use the HTMLParser, the whitespace is kept. Am I doing something wrong? Obviously I want to keep the whitespace and the files I'm working with are XHTML. Should I be using the normal XML parser in that case? thanks, --Tim Arnold from lxml import etree s = ''' <p>Suppose you have a <img src="images/g0088.png" alt="" /> <a href="g278.htm">matrix</a> as seen in Figure 3. </p> ''' parser = etree.HTMLParser() tree = etree.fromstring(s, parser=parser)#space is dropped #tree = etree.fromstring(s)# space is kept print etree.tostring(tree)

2 3

specifying catalog location
by Greg Baker March 21, 2011

March 21, 2011

I'm trying to do a validating parse of HTML and XHTML content. I have gotten this far easily enough: parser = etree.XMLParser(dtd_validation=True, no_network=True) etree.fromstring(content, parser=parser) But this may be running on a machine where I don't have root access, so populating /etc/xml and /etc/sgml isn't an option. Is there some way to specify the location of the catalog so I can store the DTDs in a local directory? (Or is there some other way entirely of doing this that I haven't thought of?) -- Greg Baker, Senior Lecturer School of Computing Science Simon Fraser University Burnaby, BC, V5A 1S6 E-mail: ggbaker(a)cs.sfu.ca

2 2

Re: [lxml] no xml_declaration for unicode?
by info＠damiro.net March 18, 2011

March 18, 2011

Hi Holger, please don't get me wrong. I don't want to mix 'utf-8' or 'unicode' or something else. Now I've understood what happend and everything is working fine. I think, the missing distinction between byte-buffers and string-types in py <2.6 leads to misapprehension in my case (see also pep-0358). The method called 'tostring', so I expect to get a object which I can handle as string and not a byte-buffer. At this point of view, it makes no sense to get a py 2.x 'str' type which really acts as bytebuffer of unicode characters. Imho it's more clear, if there two methods. One for string- and one for byte-buffer style returns. Take this as a hint. Apart from that, lxml is really great and makes fun to work with. Daniel

1 0

no xml_declaration for unicode?
by info＠damiro.net March 17, 2011

March 17, 2011

Hi, there is a strange behavior, if im trying to use the 'xml_declaration' parameter on unicode encoding: >>> from lxml import etree >>> root = etree.Element("root") >>> res = etree.tostring(root, encoding="unicode", xml_declaration=True) raises: Traceback (most recent call last): File "<pyshell#1>", line 4, in <module> res = etree.tostring(root, encoding="unicode", xml_declaration=True) File "lxml.etree.pyx", line 2838, in lxml.etree.tostring (src/lxml/lxml.etree.c:53452) ValueError: Serialisation to unicode must not request an XML declaration I don't understand why it is not allowed to control the xml declaration if i'm using unicode. My environment is lxml 2.3.0 on CPython 2.7.1 (WinXP 32bit) Thanks in advance, Daniel

3 5

Google Summer of Code 2011
by Arc Riley March 8, 2011

March 8, 2011

Hey guys Have you considered participating in Google's Summer of Code? We (PSF) are taking applications for Python-based smaller Python-based projects to participate under our umbrella as we've done the last five years. The gist is you'd potentially get a student or two being paid by Google to work on lxml through the Summer. They propose concrete projects to work on which they're expected to complete (by judgement of their mentors) in order to get paid the full amount. The application is here: https://spreadsheets.google.com/viewform?formkey=dHh3WFNGYzkyWWE0ZjM1eFFoUU…

2 1

Re: [lxml] lxml pxd files (+ .h files) in binary distributions
by Stefan Behnel March 8, 2011

March 8, 2011

Giovanni Bajo, 08.03.2011 14:12: > Hello Stefan, Hi! > we are trying to use lxml through its cython API. This is made more > difficult than necessary because binary packages for Windows do not contain > lxml pxd files necessary to interface with it That's a bug. The public API currently only includes etreepublic.pxd and tree.pxd (ok, and python.pxd, but that's just a left-over), so that's all that should be needed, plus "etree_defs.h", which defines certain C level things, some of which are exported in etreepublic.pxd as well. You will need those three to use lxml.etree from Cython code. > nor libxml .h files they depend upon. That shouldn't matter all that much. Is this supposed to run only on Windows, locally for you, or do you want to support different (user) platforms? I.e. are you going to ship your binaries to users and expect this to work on their side? > Finding the correct libxml .h files requires knowledge of which libxml > version was used to compile the binary package, to avoid random crashes in > case the ABI is different between the version the binary package was > compiled against and the one that the developer downloaded from the website. The ABI shouldn't normally differ, libxml2 is pretty good in that. There may be a dependency on the C compiler, but that should normally be the VC version that fits the CPython version. > To simplify this, my proposal would be to modify lxml's setup.py to include > both the pxd files and the required libxml .h files in the bdist target; > the files would then be automatically bundled within a subdirectory called > "include" within the lxml package (eg: site-packages/lxml/include), so that > the correct versions will be available to developers willing to play with > them. > > Does this sound a good plan? You should be able to build against other libxml2 versions, but it depends on what you want to do with it. Working with the in-memory tree shouldn't be a problem at all as Daniel takes great care to keep the ABI stable. Using things like libxml2's parser directly may cause problems as you will be using a separate library instance with its own separate setup if lxml is built statically. Could you provide some details of what you are going to do on the other side of the interface? Stefan

2 6