Scrapy: why does my response object not have a bod

Question

I wrote a spider, that worked brilliantly the first time. The second time I tried to run it, it didn't venture beyond the start_urls. I tried to fetch the url in scrapy shell and create a HtmlXPathSelector object from the returned response. That is when I got the error

So the steps were: `

[scrapy shell] fetch('http://example.com') #its something other than example.
[scrapy shell] from scrapy.selector import HtmlXPathSelector
[scrapy shell] hxs = HtmlXPathSelector(response)

---------------------------------------------------------------------------

Traceback:

AttributeError                            Traceback (most recent call last)
<ipython-input-3-a486208adf1e> in <module>()
----> 1 HtmlXPathSelector(response)

/home/codefreak/project-r42catalog/env-r42catalog/lib/python2.7/site-packages/scrapy/selector/lxmlsel.pyc in __init__(self, response, text, namespaces, _root, _expr)
     29                 body=unicode_to_str(text, 'utf-8'), encoding='utf-8')
     30         if response is not None:
---> 31             _root = LxmlDocument(response, self._parser)
     32 
     33         self.namespaces = namespaces

/home/codefreak/project-r42catalog/env-r42catalog/lib/python2.7/site-packages/scrapy/selector/lxmldocument.pyc in __new__(cls, response, parser)
     25         if parser not in cache:
     26             obj = object_ref.__new__(cls)
---> 27             cache[parser] = _factory(response, parser)
     28         return cache[parser]
     29 

/home/codefreak/project-r42catalog/env-r42catalog/lib/python2.7/site-packages/scrapy/selector/lxmldocument.pyc in _factory(response, parser_cls)
     11 def _factory(response, parser_cls):
     12     url = response.url
---> 13     body = response.body_as_unicode().strip().encode('utf8') or '<html/>'
     14     parser = parser_cls(recover=True, encoding='utf8')
     15     return etree.fromstring(body, parser=parser, base_url=url)

Error:

AttributeError: 'Response' object has no attribute 'body_as_unicode'

Am I overlooking something very obvious or stumbled upon a bug in scrapy?

Answer 1

body_as_unicode is a method of TextResponse. TextResponse, or one of its subclasses such as HtmlResponse, will be created by scrapy if the http response contains textual content.

In [1]: fetch('http://scrapy.org')
...
In [2]: type(response)
Out[2]: scrapy.http.response.html.HtmlResponse
...
In [3]: fetch('http://www.scrapy.org/site-media/images/logo.png')
...
In [4]: type(response)
Out[4]: scrapy.http.response.Response

In your case, the most likely explanation is that scrapy believes the response does not contain text.

Does the HTTP response from the server correctly set the Content-Type header? Does it render correctly in a browser? These questions will help understand if it's expected behavior or a bug.