Feature Request: Use URL to specify Cover image

[dteviot]

(Request by "Guest")

On the topic of cover issues with the extension, could I make a suggestion, to allow the setting of covers using an image from any URL instead of just those available on the page?

For example, take this page

https://www.baka-tsuki.org/project/index.php?title=The_Zashiki_Warashi_of_Intellectual_Village:Volume9

As visible on that page itself (and therefore available for packing by the extension), the closest one can get to a cover would be

https://www.baka-tsuki.org/project/images/3/31/Zashiki_v09_000.jpg

However, this is too wide as it includes the front cover, the spine, and the back cover as well.
If on the other hand one were to check the main series page,

https://www.baka-tsuki.org/project/index.php?title=The_Zashiki_Warashi_of_Intellectual_Village

there is a much better option available to act as a cover, not present on the volume's full text page.

https://www.baka-tsuki.org/project/images/3/30/Zashiki_Volume_9_Cover.jpg

As-is, the extension does not allow for setting this as the cover, and therefore the epub needs to be manually tweaked after the fact to replace the cover.

Since I'm uncertain if there is any practical easy one-size-fits-all fix to somehow magically detect the presence of a cover image on a page other than the one being viewed, then a solution could be to allow entering an image URL to fetch a specific image to act as cover.

[belldandu]

I will look into this a bit later.

Some pages appear to have the smaller cover and the wide cover within the novel illustrations while others do not. Sure this can be fixed easily once imageInfo.isCover is actually set to true somewhere.

Aside from that i think this will be fairly easy to implement a simple check for if a smaller cover already exists for the book. (only issue here is inconsistent file names)

And of course url support for the cover if you find one on the net when there isn't one on the site your getting the book from.

[belldandu]

At this point after extensively looking into this. The best option would be to provide an option for a custom cover url. This url should be passed into the images array like the rest of the images however so that all the checks can go through.

I will be making a seperate pull request from #10 to do this.

[dteviot]

Initial version implemented in version 0.0.0.9.
Needs more work for extracting the file name from the supplied URL. Some cases don't work.
There are three cases for Baka Tsuki images.

• https://www.baka-tsuki.org/project/index.php?title=File:HSDxD_v01_cover.jpg
• https://www.baka-tsuki.org/project/thumb.php?f=HSDxD_v01_cover.jpg&width=427
• https://www.baka-tsuki.org/project/images/7/76/HSDxD_v01_cover.jpg

Only the first one works correctly.

[belldandu]

@dteviot I recommend using this http://blog.teamtreehouse.com/reading-files-using-the-html5-filereader-api here https://github.com/dteviot/WebToEpub/blob/sonako/plugin/js/parsers/ImageCollector.js#L306 to get the image name and type from the file itself along with its width and height.

I've added myself to assignees for this as well.

[belldandu]

Using the filereader api will also get rid of the need for complex regex's that might fail under certain conditions.

[belldandu]

The above might be a bad idea though. I am still looking into other options.

[dteviot]

@belldandu
This works for the failing cases, but it's kind of ugly. fc13f84
I guess it will do for the moment.

[dteviot]

I'm going to call this done.

[belldandu]

Yeah seems to work for now.