trAvis - MANAGER

Edit File: unicode-frustrations.html

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
 <head>
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 
 <title>Overcoming frustration: Correctly using unicode in python2 &mdash; kitchen 1.1.1 documentation</title>
 
 <link rel="stylesheet" href="_static/default.css" type="text/css" />
 <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
 
 <script type="text/javascript">
 var DOCUMENTATION_OPTIONS = {
 URL_ROOT: '',
 VERSION: '1.1.1',
 COLLAPSE_INDEX: false,
 FILE_SUFFIX: '.html',
 HAS_SOURCE: true
 };
 </script>
 <script type="text/javascript" src="_static/jquery.js"></script>
 <script type="text/javascript" src="_static/underscore.js"></script>
 <script type="text/javascript" src="_static/doctools.js"></script>
 <link rel="search" type="application/opensearchdescription+xml"
 title="Search within kitchen 1.1.1 documentation"
 href="_static/opensearch.xml"/>
 <link rel="top" title="kitchen 1.1.1 documentation" href="index.html" />
 <link rel="up" title="Using kitchen to write good code" href="tutorial.html" />
 <link rel="next" title="Designing Unicode Aware APIs" href="designing-unicode-apis.html" />
 <link rel="prev" title="Using kitchen to write good code" href="tutorial.html" /> 
 </head>
 <body>
 <div class="related">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="genindex.html" title="General Index"
 accesskey="I">index</a></li>
 <li class="right" >
 <a href="py-modindex.html" title="Python Module Index"
 >modules</a> |</li>
 <li class="right" >
 <a href="designing-unicode-apis.html" title="Designing Unicode Aware APIs"
 accesskey="N">next</a> |</li>
 <li class="right" >
 <a href="tutorial.html" title="Using kitchen to write good code"
 accesskey="P">previous</a> |</li>
 <li><a href="index.html">kitchen 1.1.1 documentation</a> &raquo;</li>
 <li><a href="tutorial.html" accesskey="U">Using kitchen to write good code</a> &raquo;</li> 
 </ul>
 </div>

<div class="document">
 <div class="documentwrapper">
 <div class="bodywrapper">
 <div class="body">
 
 <div class="section" id="overcoming-frustration-correctly-using-unicode-in-python2">
<h1>Overcoming frustration: Correctly using unicode in python2<a class="headerlink" href="#overcoming-frustration-correctly-using-unicode-in-python2" title="Permalink to this headline">¶</a></h1>
In python-2.x, there&#8217;s two types that deal with text.
<ol class="arabic simple">
<li><tt class="xref py py-class docutils literal">str</tt> is for strings of bytes. These are very similar in nature to
how strings are handled in C.</li>
<li><tt class="xref py py-class docutils literal">unicode</tt> is for strings of unicode <a class="reference internal" href="glossary.html#term-code-points">code points</a>.</li>
</ol>
<div class="admonition note">
Note
Just what the dickens is &#8220;Unicode&#8221;?
One mistake that people encountering this issue for the first time make is
confusing the <tt class="xref py py-class docutils literal">unicode</tt> type and the encodings of unicode stored in
the <tt class="xref py py-class docutils literal">str</tt> type. In python, the <tt class="xref py py-class docutils literal">unicode</tt> type stores an
abstract sequence of <a class="reference internal" href="glossary.html#term-code-points">code points</a>. Each <a class="reference internal" href="glossary.html#term-code-point">code point</a>
represents a <a class="reference internal" href="glossary.html#term-grapheme">grapheme</a>. By contrast, byte <tt class="xref py py-class docutils literal">str</tt> stores
a sequence of bytes which can then be mapped to a sequence of <a class="reference internal" href="glossary.html#term-code-points">code
points</a>. Each unicode encoding (<a class="reference internal" href="glossary.html#term-utf-8">UTF-8</a>, UTF-7, UTF-16, UTF-32,
etc) maps different sequences of bytes to the unicode <a class="reference internal" href="glossary.html#term-code-points">code points</a>.
What does that mean to you as a programmer? When you&#8217;re dealing with text
manipulations (finding the number of characters in a string or cutting
a string on word boundaries) you should be dealing with <tt class="xref py py-class docutils literal">unicode</tt>
strings as they abstract characters in a manner that&#8217;s appropriate for
thinking of them as a sequence of letters that you will see on a page.
When dealing with I/O, reading to and from the disk, printing to
a terminal, sending something over a network link, etc, you should be dealing
with byte <tt class="xref py py-class docutils literal">str</tt> as those devices are going to need to deal with
concrete implementations of what bytes represent your abstract characters.
</div>
In the python2 world many APIs use these two classes interchangably but there
are several important APIs where only one or the other will do the right
thing. When you give the wrong type of string to an API that wants the other
type, you may end up with an exception being raised (<tt class="xref py py-exc docutils literal">UnicodeDecodeError</tt>
or <tt class="xref py py-exc docutils literal">UnicodeEncodeError</tt>). However, these exceptions aren&#8217;t always raised
because python implicitly converts between types... sometimes.
<div class="section" id="frustration-1-inconsistent-errors">
<h2>Frustration #1: Inconsistent Errors<a class="headerlink" href="#frustration-1-inconsistent-errors" title="Permalink to this headline">¶</a></h2>
Although converting when possible seems like the right thing to do, it&#8217;s
actually the first source of frustration. A programmer can test out their
program with a string like: <tt class="docutils literal">The quick brown fox jumped over the lazy dog</tt>
and not encounter any issues. But when they release their software into the
wild, someone enters the string: <tt class="docutils literal">I sat down for coffee at the café</tt> and
suddenly an exception is thrown. The reason? The mechanism that converts
between the two types is only able to deal with <a class="reference internal" href="glossary.html#term-ascii">ASCII</a> characters.
Once you throw non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> characters into your strings, you have to
start dealing with the conversion manually.
So, if I manually convert everything to either byte <tt class="xref py py-class docutils literal">str</tt> or
<tt class="xref py py-class docutils literal">unicode</tt> strings, will I be okay? The answer is.... sometimes.
</div>
<div class="section" id="frustration-2-inconsistent-apis">
<h2>Frustration #2: Inconsistent APIs<a class="headerlink" href="#frustration-2-inconsistent-apis" title="Permalink to this headline">¶</a></h2>
The problem you run into when converting everything to byte <tt class="xref py py-class docutils literal">str</tt> or
<tt class="xref py py-class docutils literal">unicode</tt> strings is that you&#8217;ll be using someone else&#8217;s API quite
often (this includes the APIs in the <a class="reference external" href="http://docs.python.org/library">python standard library</a>) and find that the API will only
accept byte <tt class="xref py py-class docutils literal">str</tt> or only accept <tt class="xref py py-class docutils literal">unicode</tt> strings. Or worse,
that the code will accept either when you&#8217;re dealing with strings that consist
solely of <a class="reference internal" href="glossary.html#term-ascii">ASCII</a> but throw an error when you give it a string that&#8217;s
got non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> characters. When you encounter these APIs you first
need to identify which type will work better and then you have to convert your
values to the correct type for that code. Thus the programmer that wants to
proactively fix all unicode errors in their code needs to do two things:
<ol class="arabic simple">
<li>You must keep track of what type your sequences of text are. Does
<tt class="docutils literal">my_sentence</tt> contain <tt class="xref py py-class docutils literal">unicode</tt> or <tt class="xref py py-class docutils literal">str</tt>? If you don&#8217;t
know that then you&#8217;re going to be in for a world of hurt.</li>
<li>Anytime you call a function you need to evaluate whether that function will
do the right thing with <tt class="xref py py-class docutils literal">str</tt> or <tt class="xref py py-class docutils literal">unicode</tt> values. Sending
the wrong value here will lead to a <tt class="xref py py-exc docutils literal">UnicodeError</tt> being thrown when
the string contains non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> characters.</li>
</ol>
<div class="admonition note">
Note
There is one mitigating factor here. The python community has been
standardizing on using <tt class="xref py py-class docutils literal">unicode</tt> in all its APIs. Although there
are some APIs that you need to send byte <tt class="xref py py-class docutils literal">str</tt> to in order to be
safe, (including things as ubiquitous as <a class="reference external" href="http://docs.python.org/library/functions.html#print" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">print()</tt></a> as we&#8217;ll see in the
next section), it&#8217;s getting easier and easier to use <tt class="xref py py-class docutils literal">unicode</tt>
strings with most APIs.
</div>
</div>
<div class="section" id="frustration-3-inconsistent-treatment-of-output">
<h2>Frustration #3: Inconsistent treatment of output<a class="headerlink" href="#frustration-3-inconsistent-treatment-of-output" title="Permalink to this headline">¶</a></h2>
Alright, since the python community is moving to using <tt class="xref py py-class docutils literal">unicode</tt>
strings everywhere, we might as well convert everything to <tt class="xref py py-class docutils literal">unicode</tt>
strings and use that by default, right? Sounds good most of the time but
there&#8217;s at least one huge caveat to be aware of. Anytime you output text to
the terminal or to a file, the text has to be converted into a byte
<tt class="xref py py-class docutils literal">str</tt>. Python will try to implicitly convert from <tt class="xref py py-class docutils literal">unicode</tt> to
byte <tt class="xref py py-class docutils literal">str</tt>... but it will throw an exception if the bytes are
non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a>:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; string = unicode(raw_input(), &#39;utf8&#39;)
café
&gt;&gt;&gt; log = open(&#39;/var/tmp/debug.log&#39;, &#39;w&#39;)
&gt;&gt;&gt; log.write(string)
Traceback (most recent call last):
 File &quot;&lt;stdin&gt;&quot;, line 1, in &lt;module&gt;
UnicodeEncodeError: &#39;ascii&#39; codec can&#39;t encode character u&#39;\xe9&#39; in position 3: ordinal not in range(128)
</pre></div>
</div>
Okay, this is simple enough to solve: Just convert to a byte <tt class="xref py py-class docutils literal">str</tt> and
we&#8217;re all set:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; string = unicode(raw_input(), &#39;utf8&#39;)
café
&gt;&gt;&gt; string_for_output = string.encode(&#39;utf8&#39;, &#39;replace&#39;)
&gt;&gt;&gt; log = open(&#39;/var/tmp/debug.log&#39;, &#39;w&#39;)
&gt;&gt;&gt; log.write(string_for_output)
&gt;&gt;&gt;
</pre></div>
</div>
So that was simple, right? Well... there&#8217;s one gotcha that makes things a bit
harder to debug sometimes. When you attempt to write non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a>
<tt class="xref py py-class docutils literal">unicode</tt> strings to a file-like object you get a traceback everytime.
But what happens when you use <a class="reference external" href="http://docs.python.org/library/functions.html#print" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">print()</tt></a>? The terminal is a file-like object
so it should raise an exception right? The answer to that is....
sometimes:
<div class="highlight-pycon"><div class="highlight"><pre>$ python
&gt;&gt;&gt; print u&#39;café&#39;
café
</pre></div>
</div>
No exception. Okay, we&#8217;re fine then?
We are until someone does one of the following:
<ul>
<li>Runs the script in a different locale:
<div class="highlight-pycon"><div class="highlight"><pre>$ LC_ALL=C python
&gt;&gt;&gt; # Note: if you&#39;re using a good terminal program when running in the C locale
&gt;&gt;&gt; # The terminal program will prevent you from entering non-ASCII characters
&gt;&gt;&gt; # python will still recognize them if you use the codepoint instead:
&gt;&gt;&gt; print u&#39;caf\xe9&#39;
Traceback (most recent call last):
 File &quot;&lt;stdin&gt;&quot;, line 1, in &lt;module&gt;
UnicodeEncodeError: &#39;ascii&#39; codec can&#39;t encode character u&#39;\xe9&#39; in position 3: ordinal not in range(128)
</pre></div>
</div>
</li>
<li>Redirects output to a file:
<div class="highlight-pycon"><div class="highlight"><pre>$ cat test.py
#!/usr/bin/python -tt
# -*- coding: utf-8 -*-
print u&#39;café&#39;
$ ./test.py &gt;t
Traceback (most recent call last):
 File &quot;./test.py&quot;, line 4, in &lt;module&gt;
 print u&#39;café&#39;
UnicodeEncodeError: &#39;ascii&#39; codec can&#39;t encode character u&#39;\xe9&#39; in position 3: ordinal not in range(128)
</pre></div>
</div>
</li>
</ul>
Okay, the locale thing is a pain but understandable: the C locale doesn&#8217;t
understand any characters outside of <a class="reference internal" href="glossary.html#term-ascii">ASCII</a> so naturally attempting to
display those won&#8217;t work. Now why does redirecting to a file cause problems?
It&#8217;s because <a class="reference external" href="http://docs.python.org/library/functions.html#print" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">print()</tt></a> in python2 is treated specially. Whereas the other
file-like objects in python always convert to <a class="reference internal" href="glossary.html#term-ascii">ASCII</a> unless you set
them up differently, using <a class="reference external" href="http://docs.python.org/library/functions.html#print" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">print()</tt></a> to output to the terminal will use
the user&#8217;s locale to convert before sending the output to the terminal. When
<a class="reference external" href="http://docs.python.org/library/functions.html#print" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">print()</tt></a> is not outputting to the terminal (being redirected to a file,
for instance), <a class="reference external" href="http://docs.python.org/library/functions.html#print" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">print()</tt></a> decides that it doesn&#8217;t know what locale to use
for that file and so it tries to convert to <a class="reference internal" href="glossary.html#term-ascii">ASCII</a> instead.
So what does this mean for you, as a programmer? Unless you have the luxury
of controlling how your users use your code, you should always, always, always
convert to a byte <tt class="xref py py-class docutils literal">str</tt> before outputting strings to the terminal or to
a file. Python even provides you with a facility to do just this. If you
know that every <tt class="xref py py-class docutils literal">unicode</tt> string you send to a particular file-like
object (for instance, <a class="reference external" href="http://docs.python.org/library/sys.html#sys.stdout" title="(in Python v2.7)"><tt class="xref py py-data docutils literal">stdout</tt></a>) should be converted to a particular
encoding you can use a <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.StreamWriter" title="(in Python v2.7)"><tt class="xref py py-class docutils literal">codecs.StreamWriter</tt></a> object to convert from
a <tt class="xref py py-class docutils literal">unicode</tt> string into a byte <tt class="xref py py-class docutils literal">str</tt>. In particular,
<a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.getwriter" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">codecs.getwriter()</tt></a> will return a <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.StreamWriter" title="(in Python v2.7)"><tt class="xref py py-class docutils literal">StreamWriter</tt></a> class
that will help you to wrap a file-like object for output. Using our
<a class="reference external" href="http://docs.python.org/library/functions.html#print" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">print()</tt></a> example:
<div class="highlight-python"><pre>$ cat test.py
#!/usr/bin/python -tt
# -*- coding: utf-8 -*-
import codecs
import sys

UTF8Writer = codecs.getwriter('utf8')
sys.stdout = UTF8Writer(sys.stdout)
print u'café'
$ ./test.py &gt;t
$ cat t
café</pre>
</div>
</div>
<div class="section" id="frustrations-4-and-5-the-other-shoes">
<h2>Frustrations #4 and #5 &#8211; The other shoes<a class="headerlink" href="#frustrations-4-and-5-the-other-shoes" title="Permalink to this headline">¶</a></h2>
In English, there&#8217;s a saying &#8220;waiting for the other shoe to drop&#8221;. It means
that when one event (usually bad) happens, you come to expect another event
(usually worse) to come after. In this case we have two other shoes.
<div class="section" id="frustration-4-now-it-doesn-t-take-byte-strings">
<h3>Frustration #4: Now it doesn&#8217;t take byte strings?!<a class="headerlink" href="#frustration-4-now-it-doesn-t-take-byte-strings" title="Permalink to this headline">¶</a></h3>
If you wrap <a class="reference external" href="http://docs.python.org/library/sys.html#sys.stdout" title="(in Python v2.7)"><tt class="xref py py-data docutils literal">sys.stdout</tt></a> using <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.getwriter" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">codecs.getwriter()</tt></a> and think you
are now safe to print any variable without checking its type I am afraid
I must inform you that you&#8217;re not paying enough attention to <a class="reference internal" href="glossary.html#term-murphy-s-law">Murphy&#8217;s
Law</a>. The <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.StreamWriter" title="(in Python v2.7)"><tt class="xref py py-class docutils literal">StreamWriter</tt></a> that <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.getwriter" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">codecs.getwriter()</tt></a>
provides will take <tt class="xref py py-class docutils literal">unicode</tt> strings and transform them into byte
<tt class="xref py py-class docutils literal">str</tt> before they get to <a class="reference external" href="http://docs.python.org/library/sys.html#sys.stdout" title="(in Python v2.7)"><tt class="xref py py-data docutils literal">sys.stdout</tt></a>. The problem is if you
give it something that&#8217;s already a byte <tt class="xref py py-class docutils literal">str</tt> it tries to transform
that as well. To do that it tries to turn the byte <tt class="xref py py-class docutils literal">str</tt> you give it
into <tt class="xref py py-class docutils literal">unicode</tt> and then transform that back into a byte <tt class="xref py py-class docutils literal">str</tt>...
and since it uses the <a class="reference internal" href="glossary.html#term-ascii">ASCII</a> codec to perform those conversions,
chances are that it&#8217;ll blow up when making them:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; import codecs
&gt;&gt;&gt; import sys
&gt;&gt;&gt; UTF8Writer = codecs.getwriter(&#39;utf8&#39;)
&gt;&gt;&gt; sys.stdout = UTF8Writer(sys.stdout)
&gt;&gt;&gt; print &#39;café&#39;
Traceback (most recent call last):
 File &quot;&lt;stdin&gt;&quot;, line 1, in &lt;module&gt;
 File &quot;/usr/lib64/python2.6/codecs.py&quot;, line 351, in write
 data, consumed = self.encode(object, self.errors)
UnicodeDecodeError: &#39;ascii&#39; codec can&#39;t decode byte 0xc3 in position 3: ordinal not in range(128)
</pre></div>
</div>
To work around this, kitchen provides an alternate version of
<a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.getwriter" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">codecs.getwriter()</tt></a> that can deal with both byte <tt class="xref py py-class docutils literal">str</tt> and
<tt class="xref py py-class docutils literal">unicode</tt> strings. Use <a class="reference internal" href="api-text-converters.html#kitchen.text.converters.getwriter" title="kitchen.text.converters.getwriter"><tt class="xref py py-func docutils literal">kitchen.text.converters.getwriter()</tt></a> in
place of the <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs" title="(in Python v2.7)"><tt class="xref py py-mod docutils literal">codecs</tt></a> version like this:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; import sys
&gt;&gt;&gt; from kitchen.text.converters import getwriter
&gt;&gt;&gt; UTF8Writer = getwriter(&#39;utf8&#39;)
&gt;&gt;&gt; sys.stdout = UTF8Writer(sys.stdout)
&gt;&gt;&gt; print u&#39;café&#39;
café
&gt;&gt;&gt; print &#39;café&#39;
café
</pre></div>
</div>
</div>
<div class="section" id="frustration-5-exceptions">
<h3>Frustration #5: Exceptions<a class="headerlink" href="#frustration-5-exceptions" title="Permalink to this headline">¶</a></h3>
Okay, so we&#8217;ve gotten ourselves this far. We convert everything to
<tt class="xref py py-class docutils literal">unicode</tt> strings. We&#8217;re aware that we need to convert back into byte
<tt class="xref py py-class docutils literal">str</tt> before we write to the terminal. We&#8217;ve worked around the
inability of the standard <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.getwriter" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">getwriter()</tt></a> to deal with both byte
<tt class="xref py py-class docutils literal">str</tt> and <tt class="xref py py-class docutils literal">unicode</tt> strings. Are we all set? Well, there&#8217;s at
least one more gotcha: raising exceptions with a <tt class="xref py py-class docutils literal">unicode</tt> message.
Take a look:
<div class="highlight-pycon"><pre>&gt;&gt;&gt; class MyException(Exception):
&gt;&gt;&gt; pass
&gt;&gt;&gt;
&gt;&gt;&gt; raise MyException(u'Cannot do this')
Traceback (most recent call last):
 File "&lt;stdin&gt;", line 1, in &lt;module&gt;
__main__.MyException: Cannot do this
&gt;&gt;&gt; raise MyException(u'Cannot do this while at a café')
Traceback (most recent call last):
 File "&lt;stdin&gt;", line 1, in &lt;module&gt;
__main__.MyException:
&gt;&gt;&gt;</pre>
</div>
No, I didn&#8217;t truncate that last line; raising exceptions really cannot handle
non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> characters in a <tt class="xref py py-class docutils literal">unicode</tt> string and will output an
exception without the message if the message contains them. What happens if
we try to use the handy dandy <a class="reference internal" href="api-text-converters.html#kitchen.text.converters.getwriter" title="kitchen.text.converters.getwriter"><tt class="xref py py-func docutils literal">getwriter()</tt></a> trick
to work around this?
<div class="highlight-pycon"><pre>&gt;&gt;&gt; import sys
&gt;&gt;&gt; from kitchen.text.converters import getwriter
&gt;&gt;&gt; sys.stderr = getwriter('utf8')(sys.stderr)
&gt;&gt;&gt; raise MyException(u'Cannot do this')
Traceback (most recent call last):
 File "&lt;stdin&gt;", line 1, in &lt;module&gt;
__main__.MyException: Cannot do this
&gt;&gt;&gt; raise MyException(u'Cannot do this while at a café')
Traceback (most recent call last):
 File "&lt;stdin&gt;", line 1, in &lt;module&gt;
__main__.MyException&gt;&gt;&gt;</pre>
</div>
Not only did this also fail, it even swallowed the trailing newline that&#8217;s
normally there.... So how to make this work? Transform from <tt class="xref py py-class docutils literal">unicode</tt>
strings to byte <tt class="xref py py-class docutils literal">str</tt> manually before outputting:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; from kitchen.text.converters import to_bytes
&gt;&gt;&gt; raise MyException(to_bytes(u&#39;Cannot do this while at a café&#39;))
Traceback (most recent call last):
 File &quot;&lt;stdin&gt;&quot;, line 1, in &lt;module&gt;
__main__.MyException: Cannot do this while at a café
&gt;&gt;&gt;
</pre></div>
</div>
<div class="admonition warning">
Warning
If you use <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.getwriter" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">codecs.getwriter()</tt></a> on <a class="reference external" href="http://docs.python.org/library/sys.html#sys.stderr" title="(in Python v2.7)"><tt class="xref py py-data docutils literal">sys.stderr</tt></a>, you&#8217;ll find
that raising an exception with a byte <tt class="xref py py-class docutils literal">str</tt> is broken by the
default <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.StreamWriter" title="(in Python v2.7)"><tt class="xref py py-class docutils literal">StreamWriter</tt></a> as well. Don&#8217;t do that or you&#8217;ll
have no way to output non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> characters. If you want to use
a <a class="reference external" href="http://docs.python.org/library/codecs.html#codecs.StreamWriter" title="(in Python v2.7)"><tt class="xref py py-class docutils literal">StreamWriter</tt></a> to encode other things on stderr while
still having working exceptions, use
<a class="reference internal" href="api-text-converters.html#kitchen.text.converters.getwriter" title="kitchen.text.converters.getwriter"><tt class="xref py py-func docutils literal">kitchen.text.converters.getwriter()</tt></a>.
</div>
</div>
</div>
<div class="section" id="frustration-6-inconsistent-apis-part-deux">
<h2>Frustration #6: Inconsistent APIs Part deux<a class="headerlink" href="#frustration-6-inconsistent-apis-part-deux" title="Permalink to this headline">¶</a></h2>
Sometimes you do everything right in your code but other people&#8217;s code fails
you. With unicode issues this happens more often than we want. A glaring
example of this is when you get values back from a function that aren&#8217;t
consistently <tt class="xref py py-class docutils literal">unicode</tt> string or byte <tt class="xref py py-class docutils literal">str</tt>.
An example from the <a class="reference external" href="http://docs.python.org/library">python standard library</a> is <a class="reference external" href="http://docs.python.org/library/gettext.html#gettext" title="(in Python v2.7)"><tt class="xref py py-mod docutils literal">gettext</tt></a>. The <a class="reference external" href="http://docs.python.org/library/gettext.html#gettext" title="(in Python v2.7)"><tt class="xref py py-mod docutils literal">gettext</tt></a> functions
are used to help translate messages that you display to users in the users&#8217;
native languages. Since most languages contain letters outside of the
<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> range, the values that are returned contain unicode characters.
<a class="reference external" href="http://docs.python.org/library/gettext.html#gettext" title="(in Python v2.7)"><tt class="xref py py-mod docutils literal">gettext</tt></a> provides you with <a class="reference external" href="http://docs.python.org/library/gettext.html#gettext.GNUTranslations.ugettext" title="(in Python v2.7)"><tt class="xref py py-meth docutils literal">ugettext()</tt></a> and
<a class="reference external" href="http://docs.python.org/library/gettext.html#gettext.GNUTranslations.ungettext" title="(in Python v2.7)"><tt class="xref py py-meth docutils literal">ungettext()</tt></a> to return these translations as
<tt class="xref py py-class docutils literal">unicode</tt> strings and <a class="reference external" href="http://docs.python.org/library/gettext.html#gettext.GNUTranslations.gettext" title="(in Python v2.7)"><tt class="xref py py-meth docutils literal">gettext()</tt></a>,
<a class="reference external" href="http://docs.python.org/library/gettext.html#gettext.GNUTranslations.ngettext" title="(in Python v2.7)"><tt class="xref py py-meth docutils literal">ngettext()</tt></a>,
<a class="reference external" href="http://docs.python.org/library/gettext.html#gettext.GNUTranslations.lgettext" title="(in Python v2.7)"><tt class="xref py py-meth docutils literal">lgettext()</tt></a>, and
<a class="reference external" href="http://docs.python.org/library/gettext.html#gettext.GNUTranslations.lngettext" title="(in Python v2.7)"><tt class="xref py py-meth docutils literal">lngettext()</tt></a> to return them as encoded byte
<tt class="xref py py-class docutils literal">str</tt>. Unfortunately, even though they&#8217;re documented to return only
one type of string or the other, the implementation has corner cases where the
wrong type can be returned.
This means that even if you separate your <tt class="xref py py-class docutils literal">unicode</tt> string and byte
<tt class="xref py py-class docutils literal">str</tt> correctly before you pass your strings to a <a class="reference external" href="http://docs.python.org/library/gettext.html#gettext" title="(in Python v2.7)"><tt class="xref py py-mod docutils literal">gettext</tt></a>
function, afterwards, you might have to check that you have the right sort of
string type again.
<div class="admonition note">
Note
<a class="reference internal" href="api-i18n.html#module-kitchen.i18n" title="kitchen.i18n"><tt class="xref py py-mod docutils literal">kitchen.i18n</tt></a> provides alternate gettext translation objects that
return only byte <tt class="xref py py-class docutils literal">str</tt> or only <tt class="xref py py-class docutils literal">unicode</tt> string.
</div>
</div>
<div class="section" id="a-few-solutions">
<h2>A few solutions<a class="headerlink" href="#a-few-solutions" title="Permalink to this headline">¶</a></h2>
Now that we&#8217;ve identified the issues, can we define a comprehensive strategy
for dealing with them?
<div class="section" id="convert-text-at-the-border">
<h3>Convert text at the border<a class="headerlink" href="#convert-text-at-the-border" title="Permalink to this headline">¶</a></h3>
If you get some piece of text from a library, read from a file, etc, turn it
into a <tt class="xref py py-class docutils literal">unicode</tt> string immediately. Since python is moving in the
direction of <tt class="xref py py-class docutils literal">unicode</tt> strings everywhere it&#8217;s going to be easier to
work with <tt class="xref py py-class docutils literal">unicode</tt> strings within your code.
If your code is heavily involved with using things that are bytes, you can do
the opposite and convert all text into byte <tt class="xref py py-class docutils literal">str</tt> at the border and
only convert to <tt class="xref py py-class docutils literal">unicode</tt> when you need it for passing to another
library or performing string operations on it.
In either case, the important thing is to pick a default type for strings and
stick with it throughout your code. When you mix the types it becomes much
easier to operate on a string with a function that can only use the other type
by mistake.
<div class="admonition note">
Note
In python3, the abstract unicode type becomes much more prominent.
The type named <tt class="docutils literal">str</tt> is the equivalent of python2&#8217;s <tt class="xref py py-class docutils literal">unicode</tt> and
python3&#8217;s <tt class="docutils literal">bytes</tt> type replaces python2&#8217;s <tt class="xref py py-class docutils literal">str</tt>. Most APIs deal
in the unicode type of string with just some pieces that are low level
dealing with bytes. The implicit conversions between bytes and unicode
is removed and whenever you want to make the conversion you need to do so
explicitly.
</div>
</div>
<div class="section" id="when-the-data-needs-to-be-treated-as-bytes-or-unicode-use-a-naming-convention">
<h3>When the data needs to be treated as bytes (or unicode) use a naming convention<a class="headerlink" href="#when-the-data-needs-to-be-treated-as-bytes-or-unicode-use-a-naming-convention" title="Permalink to this headline">¶</a></h3>
Sometimes you&#8217;re converting nearly all of your data to <tt class="xref py py-class docutils literal">unicode</tt>
strings but you have one or two values where you have to keep byte
<tt class="xref py py-class docutils literal">str</tt> around. This is often the case when you need to use the value
verbatim with some external resource. For instance, filenames or key values
in a database. When you do this, use a naming convention for the data you&#8217;re
working with so you (and others reading your code later) don&#8217;t get confused
about what&#8217;s being stored in the value.
If you need both a textual string to present to the user and a byte value for
an exact match, consider keeping both versions around. You can either use two
variables for this or a <a class="reference external" href="http://docs.python.org/library/stdtypes.html#dict" title="(in Python v2.7)"><tt class="xref py py-class docutils literal">dict</tt></a> whose key is the byte value.
<div class="admonition note">
Note
You can use the naming convention used in kitchen as a guide for
implementing your own naming convention. It prefixes byte <tt class="xref py py-class docutils literal">str</tt>
variables of unknown encoding with <tt class="docutils literal">b_</tt> and byte <tt class="xref py py-class docutils literal">str</tt> of known
encoding with the encoding name like: <tt class="docutils literal">utf8_</tt>. If the default was to
handle <tt class="xref py py-class docutils literal">str</tt> and only keep a few <tt class="xref py py-class docutils literal">unicode</tt> values, those
variables would be prefixed with <tt class="docutils literal">u_</tt>.
</div>
</div>
<div class="section" id="when-outputting-data-convert-back-into-bytes">
<h3>When outputting data, convert back into bytes<a class="headerlink" href="#when-outputting-data-convert-back-into-bytes" title="Permalink to this headline">¶</a></h3>
When you go to send your data back outside of your program (to the filesystem,
over the network, displaying to the user, etc) turn the data back into a byte
<tt class="xref py py-class docutils literal">str</tt>. How you do this will depend on the expected output format of
the data. For displaying to the user, you can use the user&#8217;s default encoding
using <a class="reference external" href="http://docs.python.org/library/locale.html#locale.getpreferredencoding" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">locale.getpreferredencoding()</tt></a>. For entering into a file, you&#8217;re best
bet is to pick a single encoding and stick with it.
<div class="admonition warning">
Warning
When using the encoding that the user has set (for instance, using
<a class="reference external" href="http://docs.python.org/library/locale.html#locale.getpreferredencoding" title="(in Python v2.7)"><tt class="xref py py-func docutils literal">locale.getpreferredencoding()</tt></a>, remember that they may have their
encoding set to something that can&#8217;t display every single unicode
character. That means when you convert from <tt class="xref py py-class docutils literal">unicode</tt> to a byte
<tt class="xref py py-class docutils literal">str</tt> you need to decide what should happen if the byte value is
not valid in the user&#8217;s encoding. For purposes of displaying messages to
the user, it&#8217;s usually okay to use the <tt class="docutils literal">replace</tt> encoding error handler
to replace the invalid characters with a question mark or other symbol
meaning the character couldn&#8217;t be displayed.
</div>
You can use <a class="reference internal" href="api-text-converters.html#kitchen.text.converters.getwriter" title="kitchen.text.converters.getwriter"><tt class="xref py py-func docutils literal">kitchen.text.converters.getwriter()</tt></a> to do this automatically
for <a class="reference external" href="http://docs.python.org/library/sys.html#sys.stdout" title="(in Python v2.7)"><tt class="xref py py-data docutils literal">sys.stdout</tt></a>. When creating exception messages be sure to convert
to bytes manually.
</div>
<div class="section" id="when-writing-unittests-include-non-ascii-values-and-both-unicode-and-str-type">
<h3>When writing unittests, include non-ASCII values and both unicode and str type<a class="headerlink" href="#when-writing-unittests-include-non-ascii-values-and-both-unicode-and-str-type" title="Permalink to this headline">¶</a></h3>
Unless you know that a specific portion of your code will only deal with
<a class="reference internal" href="glossary.html#term-ascii">ASCII</a>, be sure to include non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> values in your unittests.
Including a few characters from several different scripts is highly advised as
well because some code may have special cased accented roman characters but
not know how to handle characters used in Asian alphabets.
Similarly, unless you know that that portion of your code will only be given
<tt class="xref py py-class docutils literal">unicode</tt> strings or only byte <tt class="xref py py-class docutils literal">str</tt> be sure to try variables
of both types in your unittests. When doing this, make sure that the
variables are also non-<a class="reference internal" href="glossary.html#term-ascii">ASCII</a> as python&#8217;s implicit conversion will mask
problems with pure <a class="reference internal" href="glossary.html#term-ascii">ASCII</a> data. In many cases, it makes sense to check
what happens if byte <tt class="xref py py-class docutils literal">str</tt> and <tt class="xref py py-class docutils literal">unicode</tt> strings that won&#8217;t
decode in the present locale are given.
</div>
<div class="section" id="be-vigilant-about-spotting-poor-apis">
<h3>Be vigilant about spotting poor APIs<a class="headerlink" href="#be-vigilant-about-spotting-poor-apis" title="Permalink to this headline">¶</a></h3>
Make sure that the libraries you use return only <tt class="xref py py-class docutils literal">unicode</tt> strings or
byte <tt class="xref py py-class docutils literal">str</tt>. Unittests can help you spot issues here by running many
variations of data through your functions and checking that you&#8217;re still
getting the types of string that you expect.
</div>
<div class="section" id="example-putting-this-all-together-with-kitchen">
<h3>Example: Putting this all together with kitchen<a class="headerlink" href="#example-putting-this-all-together-with-kitchen" title="Permalink to this headline">¶</a></h3>
The kitchen library provides a wide array of functions to help you deal with
byte <tt class="xref py py-class docutils literal">str</tt> and <tt class="xref py py-class docutils literal">unicode</tt> strings in your program. Here&#8217;s
a short example that uses many kitchen functions to do its work:
<div class="highlight-python"><div class="highlight"><pre>#!/usr/bin/python -tt
# -*- coding: utf-8 -*-
import locale
import os
import sys
import unicodedata

from kitchen.text.converters import getwriter, to_bytes, to_unicode
from kitchen.i18n import get_translation_object

if __name__ == &#39;__main__&#39;:
 # Setup gettext driven translations but use the kitchen functions so
 # we don&#39;t have the mismatched bytes-unicode issues.
 translations = get_translation_object(&#39;example&#39;)
 # We use _() for marking strings that we operate on as unicode
 # This is pretty much everything
 _ = translations.ugettext
 # And b_() for marking strings that we operate on as bytes.
 # This is limited to exceptions
 b_ = translations.lgettext

# Setup stdout
 encoding = locale.getpreferredencoding()
 Writer = getwriter(encoding)
 sys.stdout = Writer(sys.stdout)

# Load data. Format is filename\0description
 # description should be utf-8 but filename can be any legal filename
 # on the filesystem
 # Sample datafile.txt:
 # /etc/shells\x00Shells available on caf\xc3\xa9.lan
 # /var/tmp/file\xff\x00File with non-utf8 data in the filename
 #
 # And to create /var/tmp/file\xff (under bash or zsh) do:
 # echo &#39;Some data&#39; &gt; /var/tmp/file$&#39;\377&#39;
 datafile = open(&#39;datafile.txt&#39;, &#39;r&#39;)
 data = {}
 for line in datafile:
 # We&#39;re going to keep filename as bytes because we will need the
 # exact bytes to access files on a POSIX operating system.
 # description, we&#39;ll immediately transform into unicode type.
 b_filename, description = line.split(&#39;\0&#39;, 1)

# to_unicode defaults to decoding output from utf-8 and replacing
 # any problematic bytes with the unicode replacement character
 # We accept mangling of the description here knowing that our file
 # format is supposed to use utf-8 in that field and that the
 # description will only be displayed to the user, not used as
 # a key value.
 description = to_unicode(description, &#39;utf-8&#39;).strip()
 data[b_filename] = description
 datafile.close()

# We&#39;re going to add a pair of extra fields onto our data to show the
 # length of the description and the filesize. We put those between
 # the filename and description because we haven&#39;t checked that the
 # description is free of NULLs.
 datafile = open(&#39;newdatafile.txt&#39;, &#39;w&#39;)

# Name filename with a b_ prefix to denote byte string of unknown encoding
 for b_filename in data:
 # Since we have the byte representation of filename, we can read any
 # filename
 if os.access(b_filename, os.F_OK):
 size = os.path.getsize(b_filename)
 else:
 size = 0
 # Because the description is unicode type, we know the number of
 # characters corresponds to the length of the normalized unicode
 # string.
 length = len(unicodedata.normalize(&#39;NFC&#39;, description))

# Print a summary to the screen
 # Note that we do not let implici type conversion from str to
 # unicode transform b_filename into a unicode string. That might
 # fail as python would use the ASCII filename. Instead we use
 # to_unicode() to explictly transform in a way that we know will
 # not traceback.
 print _(u&#39;filename: %s&#39;) % to_unicode(b_filename)
 print _(u&#39;file size: %s&#39;) % size
 print _(u&#39;desc length: %s&#39;) % length
 print _(u&#39;description: %s&#39;) % data[b_filename]

# First combine the unicode portion
 line = u&#39;%s\0%s\0%s&#39; % (size, length, data[b_filename])
 # Since the filenames are bytes, turn everything else to bytes before combining
 # Turning into unicode first would be wrong as the bytes in b_filename
 # might not convert
 b_line = &#39;%s\0%s\n&#39; % (b_filename, to_bytes(line))

# Just to demonstrate that getwriter will pass bytes through fine
 print b_(&#39;Wrote: %s&#39;) % b_line
 datafile.write(b_line)
 datafile.close()

# And just to show how to properly deal with an exception.
 # Note two things about this:
 # 1) We use the b_() function to translate the string. This returns a
 # byte string instead of a unicode string
 # 2) We&#39;re using the b_() function returned by kitchen. If we had
 # used the one from gettext we would need to convert the message to
 # a byte str first
 message = u&#39;Demonstrate the proper way to raise exceptions. Sincerely, \u3068\u3057\u304a&#39;
 raise Exception(b_(message))
</pre></div>
</div>
<div class="admonition-see-also admonition seealso">
See also
<a class="reference internal" href="api-text-converters.html#module-kitchen.text.converters" title="kitchen.text.converters"><tt class="xref py py-mod docutils literal">kitchen.text.converters</tt></a>
</div>
</div>
</div>
</div>

</div>
 </div>
 </div>
 <div class="sphinxsidebar">
 <div class="sphinxsidebarwrapper">
 <h3><a href="index.html">Table Of Contents</a></h3>
 <ul>
<li><a class="reference internal" href="#">Overcoming frustration: Correctly using unicode in python2</a><ul>
<li><a class="reference internal" href="#frustration-1-inconsistent-errors">Frustration #1: Inconsistent Errors</a></li>
<li><a class="reference internal" href="#frustration-2-inconsistent-apis">Frustration #2: Inconsistent APIs</a></li>
<li><a class="reference internal" href="#frustration-3-inconsistent-treatment-of-output">Frustration #3: Inconsistent treatment of output</a></li>
<li><a class="reference internal" href="#frustrations-4-and-5-the-other-shoes">Frustrations #4 and #5 &#8211; The other shoes</a><ul>
<li><a class="reference internal" href="#frustration-4-now-it-doesn-t-take-byte-strings">Frustration #4: Now it doesn&#8217;t take byte strings?!</a></li>
<li><a class="reference internal" href="#frustration-5-exceptions">Frustration #5: Exceptions</a></li>
</ul>
</li>
<li><a class="reference internal" href="#frustration-6-inconsistent-apis-part-deux">Frustration #6: Inconsistent APIs Part deux</a></li>
<li><a class="reference internal" href="#a-few-solutions">A few solutions</a><ul>
<li><a class="reference internal" href="#convert-text-at-the-border">Convert text at the border</a></li>
<li><a class="reference internal" href="#when-the-data-needs-to-be-treated-as-bytes-or-unicode-use-a-naming-convention">When the data needs to be treated as bytes (or unicode) use a naming convention</a></li>
<li><a class="reference internal" href="#when-outputting-data-convert-back-into-bytes">When outputting data, convert back into bytes</a></li>
<li><a class="reference internal" href="#when-writing-unittests-include-non-ascii-values-and-both-unicode-and-str-type">When writing unittests, include non-ASCII values and both unicode and str type</a></li>
<li><a class="reference internal" href="#be-vigilant-about-spotting-poor-apis">Be vigilant about spotting poor APIs</a></li>
<li><a class="reference internal" href="#example-putting-this-all-together-with-kitchen">Example: Putting this all together with kitchen</a></li>
</ul>
</li>
</ul>
</li>
</ul>

<h4>Previous topic</h4>
 <a href="tutorial.html"
 title="previous chapter">Using kitchen to write good code</a>
 <h4>Next topic</h4>
 <a href="designing-unicode-apis.html"
 title="next chapter">Designing Unicode Aware APIs</a>
 <h3>This Page</h3>
 <ul class="this-page-menu">
 <li><a href="_sources/unicode-frustrations.txt"
 rel="nofollow">Show Source</a></li>
 </ul>
<div id="searchbox" style="display: none">
 <h3>Quick search</h3>
 <form class="search" action="search.html" method="get">
 <input type="text" name="q" />
 <input type="submit" value="Go" />
 <input type="hidden" name="check_keywords" value="yes" />
 <input type="hidden" name="area" value="default" />
 </form>
 
 Enter search terms or a module, class or function name.
 
</div>
<script type="text/javascript">$('#searchbox').show(0);</script>
 </div>
 </div>
 <div class="clearer"></div>
 </div>
 <div class="related">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="genindex.html" title="General Index"
 >index</a></li>
 <li class="right" >
 <a href="py-modindex.html" title="Python Module Index"
 >modules</a> |</li>
 <li class="right" >
 <a href="designing-unicode-apis.html" title="Designing Unicode Aware APIs"
 >next</a> |</li>
 <li class="right" >
 <a href="tutorial.html" title="Using kitchen to write good code"
 >previous</a> |</li>
 <li><a href="index.html">kitchen 1.1.1 documentation</a> &raquo;</li>
 <li><a href="tutorial.html" >Using kitchen to write good code</a> &raquo;</li> 
 </ul>
 </div>
 <div class="footer">
 &copy; Copyright 2011 Red Hat, Inc. and others.
 Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 1.1.3.
 </div>
 </body>
</html>