adding spider args as attrs to docs

[demelziraptor]

No description provided.

[codecov-io]

Current coverage is 82.99%

Merging #1719 into master will not affect coverage as of f72017a

Powered by Codecov. Updated on successful CI builds.

[eliasdorneles]

Hey @demelziraptor !
This is about documenting the feature of spider arguments being exposed as attributes, right? (this line: https://github.com/scrapy/scrapy/blob/master/scrapy/spiders/__init__.py#L30)

From the example it isn't clear what this is trying to show.

[demelziraptor]

Exactly @eliasdorneles, from the current docs it seems like the only way to access spider input arguments is from the constructor arguments.

Maybe this example is better?

    import scrapy

    class MySpider(scrapy.Spider):
        name = 'myspider'

        def parse(self, response):
            products = response.css('.%s' % self.category)

[eliasdorneles]

Hm, I think it'll need more than an example.
Looking at the example alone it looks like broken code. :)

Perhaps adding a default as a class attribute, like:

import scrapy

class MySpider(scrapy.Spider):
    name = 'myspider'
    category = 'default-category'  # use -a category=another from the cmdline to override this

    def parse(self, response):
        products = response.css('.%s' % self.category)

This may deserve its own section in the docs really. :)

[nramirezuy]

I would suggest to add a spider that can be copied, ran with runspider and get a result; something somehow interactive. Doesn't need to browse, it can do some comparison and close, maybe using start_requests or adding a signal.

[kmike]

Closing it in favor of #2483. Thanks @demelziraptor!