How to work with unicode in Python

Question

I am trying to clean all of the HTML out of a string so the final output is a text file. I have some some research on the various 'converters' and am starting to lean towards creating my own dictionary for the entities and symbols and running a replace on the string. I am considering this because I want to automate the process and there is a lot of variability in the quality of the underlying html. To begin comparing the speed of my solution and one of the alternatives for example pyparsing I decided to test replace of \xa0 using the string method replace. I get a

UnicodeDecodeError: 'ascii' codec can't decode byte 0xa0 in position 0: ordinal not in range(128)

The actual line of code was

s=unicodestring.replace('\xa0','')

Anyway-I decided that I needed to preface it with an r so I ran this line of code:

s=unicodestring.replace(r'\xa0','')

It runs without error but I when I look at a slice of s I see that the \xaO is still there

Why would you prefix '\xa0' with an r? That makes it a raw string - that is, it literally contains backslash, x, a, 0. Without the r, it contained a single character with hex code a0, which I think is what you wanted. — David Z, Apr 15 '09 at 18:26
Because I was trying to guess why I got the error and I know that sometimes to force the \ to be read you have to make it a string literal and also the \xa0 is what actually exists in my source. what is hex code a0? — PyNEwbie, Apr 15 '09 at 18:44

score 25 · Accepted Answer · answered Apr 15 '09 at 18:22

25

may be you should be doing

s=unicodestring.replace(u'\xa0',u'')

answered Apr 15 '09 at 18:22

z33m

5,993
1
31
42

So how did you know to do this since I have not seen this in any example? Thanks – PyNEwbie Apr 15 '09 at 18:30

score 6 · Answer 2 · answered Apr 15 '09 at 20:33

s=unicodestring.replace('\xa0','')

..is trying to create the unicode character \xa0, which is not valid in an ASCII sctring (the default string type in Python until version 3.x)

The reason r'\xa0' did not error is because in a raw string, escape sequences have no effect. Rather than trying to encode \xa0 into the unicode character, it saw the string as a "literal backslash", "literal x" and so on..

The following are the same:

>>> r'\xa0'
'\\xa0'
>>> '\\xa0'
'\\xa0'

This is something resolved in Python v3, as the default string type is unicode, so you can just do..

>>> '\xa0'
'\xa0'

I am trying to clean all of the HTML out of a string so the final output is a text file

I would strongly recommend BeautifulSoup for this. Writing an HTML cleaning tool is difficult (given how horrible most HTML is), and BeautifulSoup does a great job at both parsing HTML, and dealing with Unicode..

>>> from BeautifulSoup import BeautifulSoup
>>> soup = BeautifulSoup("<html><body><h1>Hi</h1></body></html>")
>>> print soup.prettify()
<html>
 <body>
  <h1>
   Hi
  </h1>
 </body>
</html>

I appreciate this answer. I have used BS to extract data from tables and it is very useful. However, it seems to me that to remove the html using BS I have to know what is present. Am I wrong about that? — PyNEwbie, Apr 15 '09 at 23:11
I'm not sure what you mean? You can remove HTML via countless ways, from the first table in a div, to by-class-or-id etc.. — dbr, Apr 16 '09 at 14:01

score 3 · Answer 3 · answered Apr 15 '09 at 18:17

3

Look at the codecs standard library, specifically the encode and decode methods provided in the Codec base class.

There's also a good article here that puts it all together.

answered Apr 15 '09 at 18:17

Wayne Koorts

10,861
13
46
72

Thanks-great article you are right it does put a lot together. – PyNEwbie Apr 15 '09 at 18:32

score 2 · Answer 4 · edited Jun 30 '14 at 11:44

2

Instead of this, it's better to use standard python features.

For example:

string = unicode('Hello, \xa0World', 'utf-8', 'replace')

or

string = unicode('Hello, \xa0World', 'utf-8', 'ignore')

where replace will replace \xa0 to \\xa0.

But if \xa0 is really not meaningful for you and you want to remove it then use ignore.

edited Jun 30 '14 at 11:44

André Dion

21,269
7
56
60

answered Sep 13 '12 at 13:19

Tejas Tank

1,100
2
16
28

score 1 · Answer 5 · answered Apr 15 '09 at 18:18

1

Just a note regarding HTML cleaning. It is very very hard, since

<
body
>

Is a valid way to write HTML. Just an fyi.

answered Apr 15 '09 at 18:18

Ólafur Waage

68,817
22
142
198

score 0 · Answer 6 · answered Apr 15 '09 at 18:18

0

You can convert it to unicode in this way:

print u'Hello, \xa0World'  # print Hello,  World

answered Apr 15 '09 at 18:18

Jason Coon

17,601
10
42
50

How to work with unicode in Python

6 Answers6

Linked