XML with Unicode: what am I doing wrong?

[Kevin Dangoor]

This is a followup to a blog post I wrote the other day
http://www.blueskyonmars.com/archive...ementtidy.html

I started out working in the context of elementtidy, but now I am
running into trouble in general Python-XML areas, so I thought I'd toss
the question out here. The code below is fairly self-explanatory. I have
a small HTML snippet that is UTF-8 encoded and is not 7-bit ASCII
compatible. I use Tidy to convert it to XHTML, and this particular setup
returns a unicode instance rather than a string.

import _elementtidy as et
from xml.parsers import expat

data = unicode(open("snippetWithUnicode.html").read(), "utf-8")
html = et.fixup(data)[0]
parser = expat.ParserCreate()
parser.Parse(html)

UnicodeEncodeError: 'ascii' codec can't encode character '\ub5' in
position 542: ordinal not in range(128)

If I set my default encoding to utf8 in sitecustomize.py, it works just
fine. I'm thinking that I can't be the only one trying to pass unicode
to expat... Is there something else I need to do here?

Thanks,
Kevin
Blazing Things

[Diez B. Roggisch]

> I started out working in the context of elementtidy, but now I am

running into trouble in general Python-XML areas, so I thought I'd toss
the question out here. The code below is fairly self-explanatory. I have
a small HTML snippet that is UTF-8 encoded and is not 7-bit ASCII
compatible. I use Tidy to convert it to XHTML, and this particular setup
returns a unicode instance rather than a string.

import _elementtidy as et
from xml.parsers import expat

data = unicode(open("snippetWithUnicode.html").read(), "utf-8")
html = et.fixup(data)[0]
parser = expat.ParserCreate()
parser.Parse(html)

UnicodeEncodeError: 'ascii' codec can't encode character '\ub5' in
position 542: ordinal not in range(128)

If I set my default encoding to utf8 in sitecustomize.py, it works just
fine. I'm thinking that I can't be the only one trying to pass unicode
to expat... Is there something else I need to do here?

you confuse unicode with utf8. Expat can parse the latter - the former is
internal to python. And passing it to something that needs a string will
result in a conversion - which fails because of the ascii encoding.

Do this:

parser.Parse(html.encode('utf-8'))

--
Regards,

Diez B. Roggisch

[Just]

In article <ct*************@news.t-online.com>,
"Diez B. Roggisch" <de*********@web.de> wrote:

I started out working in the context of elementtidy, but now I am
running into trouble in general Python-XML areas, so I thought I'd toss
the question out here. The code below is fairly self-explanatory. I have
a small HTML snippet that is UTF-8 encoded and is not 7-bit ASCII
compatible. I use Tidy to convert it to XHTML, and this particular setup
returns a unicode instance rather than a string.

import _elementtidy as et
from xml.parsers import expat

data = unicode(open("snippetWithUnicode.html").read(), "utf-8")
html = et.fixup(data)[0]
parser = expat.ParserCreate()
parser.Parse(html)

UnicodeEncodeError: 'ascii' codec can't encode character '\ub5' in
position 542: ordinal not in range(128)

If I set my default encoding to utf8 in sitecustomize.py, it works just
fine. I'm thinking that I can't be the only one trying to pass unicode
to expat... Is there something else I need to do here?

you confuse unicode with utf8. Expat can parse the latter - the former is
internal to python. And passing it to something that needs a string will
result in a conversion - which fails because of the ascii encoding.

Do this:

parser.Parse(html.encode('utf-8'))

Possibly preceded by

parser = expat.ParserCreate('utf-8')

...so there's no confusion with the declared encoding, in case that's not
utf-8.

Just