Convert and compress pictures before adding them to the ZIM

Open benoit74 opened this issue 10 months ago • 1 comments

Experienced with other scrapers showed that converting pictures to WEBP and compressing them can save a significant ZIM size.

This is probably true as well for warc2zim.

Suggested in https://github.com/openzim/libzim/issues/951

Feb 10 '25 14:02 benoit74

In my own limited testing, webp and avif seem to get the best compression, but webp certainly wins when it comes to speed and platform support.

libwebp has support for jpeg, png, gif, tiff, bmp, and netpbm. This should cover most of what would be scraped into a zim. avif and jpegxl aren't supported by libwebp from what I've been able to find, but those formats are still pretty rare to see.

Feb 13 '25 00:02 ThisIsFineTM