Use contentType for images about epub-gen HOT 4 OPEN

Hi @Secretmapper!

Yap, I also think that makes sense... because some image URLs don't have the extension and so on. Nice one. I think we could advance with a PR for it soon, yes. Cool! 👍

However, I'm not quite sure the meaning of "baseURL"? Is it the base url of the remote images?
If so, I prefer to replacing the image url in the HTML string with regex or with cheerio or some library else before pass the HTML string to epub-gen. What do you think?

Regarding the baseUrl option, I also lean to have that out of the library scope to make it more flexible and to keep the library simpler. Like @cyrilis is saying, I see that more of a lib user manipulation, through a Regex, of the data that gets fed to epub-gen, than a config option.

But, the point you arise of the parsing (twice) has some sense. And also the idea of a callback function seems interesting to me - and if it could benefit you for that GridFS scenario the better.

Gonna wrap my head around it, think about this a bit and come back with more comments - but seems to me that it would be easy to work something out. 😏

Cheers!

from epub-gen.

Hi @Secretmapper ,

Thanks for your feed back! what you said about detect image type by contentType should be right, A following PR is on the way. 😄

However, I'm not quite sure the meaning of "baseURL"? Is it the base url of the remote images?
If so, I prefer to replacing the image url in the HTML string with regex or with cheerio or some library else before pass the HTML string to epub-gen. What do you think?

Let me know if you have any questions.

Thanks.

from epub-gen.

and hi @pedrosanta, What do you think? Should this be added to options?

from epub-gen.

Hello @cyrilis thanks for the feedback! Glad to hear a PR is along the way :)

The reason I would opt to have baseUrl as an option is to save a parse step. Epub-gen is already parsing the html/url src when it's collecting the urls, so it seems wasteful to parse beforehand.

In fact I just realized as I'm typing this, it would probably be great if we can pass a callback function for image resolving, to make it more general. That way, we can for example do custom transforms on the URLs like we did here, and even read the data from different storage mechanisms.

I'm using GridFS, and right now my server is doing a round trip to download the assets on my own server. It'll be great to eliminate that.

from epub-gen.