News sources images getting resized

[alvarob]

Hi. I was trying to fix a .recipe file that was getting small images in my kindle paperwhite. So I made a new function so it downloads the biggest size. Then added some img css so it used the full width. But it didn't work. When I converted the .mobi file to .epub I noticed that all the images had some custom css:

.calibre_13 {
  height: 222px;
  vertical-align: baseline;
  width: 414px;
}
.calibre_14 {
  height: 252px;
  vertical-align: baseline;
  width: 414px;
}
.calibre_15 {
  height: 240px;
  vertical-align: baseline;
  width: 414px;
}

And many more like that. All with that width, 414px, preventing it from using the full screen. Why is calibre doing that? How do I stop it?

Thank you.

[kovidgoyal]

There is no 414 in calibre, its coming from the source html/css

[alvarob]

You're right, I was looking at the css but it seems that comes from the html.

Any idea on how to parse that from the source?
The website looks like this:

<img height="0" width="414" src="">

And I would like to remove both height and width from that tag, whatever size they specify (it can vary), leaving only

<img src="">

I guess I coud try with soup, something like

def postprocess_html(self, soup, first_fetch):
while len(soup.find_all('width')) > 0:
    soup.width.extract()
while len(soup.find_all('height')) > 0:
    soup.height.extract()
return soup

[kovidgoyal]

Just add remove_attributes = ['width']

to the recipe

[alvarob]

Thank you!