ItemLoader: support non-TextResponse

At the moment, ItemLoader(response=response) fails if response is not a TextResponse instance.

Passing a binary response can still be useful, though. For example, to allow processors to access the response from their loader context, and hence be able to report the source URL (response.url) when reporting input issues.

[Gallaecio]

@GeorgeA92 You are testing the upstream ItemLoader, Scrapy subclasses it to make some Scrapy-specific adjustments:

>>> from scrapy.http.response import Response
>>> from scrapy.loader import ItemLoader
>>> ItemLoader(response=Response(url='https://example.org'))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/adrian/proxectos/scrapy/scrapy/loader/__init__.py", line 86, in __init__
    selector = self.default_selector_class(response)
  File "/home/adrian/proxectos/scrapy/scrapy/selector/unified.py", line 78, in __init__
    text = response.text
  File "/home/adrian/proxectos/scrapy/scrapy/http/response/__init__.py", line 129, in text
    raise AttributeError("Response content isn't text")
AttributeError: Response content isn't text

@calicomills That avoids the exception, but prevents the response from being available from a loader context, so it does not really solve the root issue.

[Gallaecio]

Thar’s great! No need to ask for permission 🙂