keep_only_tags and the order of the related contents

[njpig]

Hi,

The order of the classes in keep_only_tags controls the order how the related contents display on the html page, right?

e.g, these two keep_only_tags will get contents on the page displayed in different order.

keep_only_tags = [
dict(attrs={'class': re.compile('^SplitScreenContentHeaderHed')}), <---
dict(attrs={'class': re.compile('^SplitScreenContentHeaderDek')}),
dict(attrs={'class': re.compile('^SplitScreenContentHeaderByline')}),
dict(attrs={'class': re.compile('^SplitScreenContentHeaderPublishDate') }),
dict(attrs={'class': re.compile('^SplitScreenContentHeaderLedeBlock')}) ,
dict(attrs={'class': re.compile('^SplitScreenContentHeaderCaption')}),
]

keep_only_tags = [
dict(attrs={'class': re.compile('^SplitScreenContentHeaderDek')}),
dict(attrs={'class': re.compile('^SplitScreenContentHeaderByline')}),
dict(attrs={'class': re.compile('^SplitScreenContentHeaderPublishDate') }),
dict(attrs={'class': re.compile('^SplitScreenContentHeaderLedeBlock')}) ,
dict(attrs={'class': re.compile('^SplitScreenContentHeaderCaption')}),
dict(attrs={'class': re.compile('^SplitScreenContentHeaderHed')}), <---
]

[kovidgoyal]

yes they do. combine them all into one if you dont care about order with this function

def prefixed_classes(classes):
    q = frozenset(classes.split(' '))

    def matcher(x):
        if x:
            for candidate in frozenset(x.split()):
                for x in q:
                    if candidate.startswith(x):
                        return True
        return False
    return {'attrs': {'class': matcher}}

keep_onlly_tags = [prefixed_classes('prefix1 prefix2 prefix3')]

[njpig]

Originally Posted by kovidgoyal

yes they do. combine them all into one if you dont care about order with this function

def prefixed_classes(classes):
    q = frozenset(classes.split(' '))

    def matcher(x):
        if x:
            for candidate in frozenset(x.split()):
                for x in q:
                    if candidate.startswith(x):
                        return True
        return False
    return {'attrs': {'class': matcher}}

keep_onlly_tags = [prefixed_classes('prefix1 prefix2 prefix3')]

Thanks a lot!