qsniyg / maxurl

Finds larger/original versions of images and videos
https://qsniyg.github.io/maxurl/
Apache License 2.0
1.08k stars 65 forks source link

How do you find larger images? #5

Open rEnr3n opened 6 years ago

rEnr3n commented 6 years ago

I'm not a developer but I'm interested in finding it out.

I came across these links and I feel like they have larger versions. I don't know how to start looking. https://goss.vcg.com/editorial/vcg/800/new/VCG111152795068.jpg https://goss2.vcg.com/editorial/vcg/800/new/VCG111152804617.jpg http://pictures.icpress.cn/thumbs/imgs/2018/0511/20180511_83981.jpg

qsniyg commented 6 years ago

It depends on the site! I was thinking of writing a tutorial or something on it at some point, but here's the general idea:

There are the rather obvious ways, such as URLs ending with ".jpg?width=500", removing everything after "?" (which is by far the largest rule), or "-500x300.jpg" to ".jpg" (second largest rule), but after that, it's mainly just trying out a lot of different URLs.

For example, in the ones you linked:

https://goss.vcg.com/editorial/vcg/800/new/VCG111152795068.jpg

The 800 in the URL is the same as the height, which could mean it's a container. So I would personally first try removing /800/ (so now it's https://goss.vcg.com/editorial/vcg/new/VCG111152795068.jpg), but that didn't work in this case. Then try replacing 800 with a smaller/larger number, neither of which worked, which probably means it's not a dynamic resize.

If it was dynamic, I would try either 0 or -1 (many of them treat those as "give the original image") or rather large numbers (e.g. 2000) to see if they stretch it. If they don't (I still haven't found a good reason why the server should stretch it, but many servers don't check that, unfortunately), then simply replace it with a very large number (e.g. 9999999, works for surprisingly quite a few sites, including iTunes).

Otherwise, I try replacing it with full, orig, origin, original, raw, master, source, default, binary, etc.

The way I usually do it though is usually to google site:(site that hosts the image), and look for the largest images there, to see if I can find a pattern that way. Sometimes, googling inurl:(domain that hosts the image) is more helpful, but in some cases (such as The Guardian), the full version can be on another domain.

Recently Google has silently disabled the ability to search for images larger than a specific size, but you can use it by adding/replacing &tbs=isz:lt,islt:2mp (this will search for images >= 2MP). There's a guide on this too

I wasn't able to find anything that worked for vcg or icpress.cn though, even after using google. There might very well be a way for both though (especially considering "/thumbs/" is in the icpress.cn URL). I also stumbled across this modification: http://pictures.icpress.cn/thumbs/imgs/2018/0511/20180511_83981s.jpg, so by replacing "s.jpg" to ".jpg", you get a larger image (which is the image you linked).

On a sidenote, in case this is of any help, the pictures.icpress.cn comes from this site: http://www.imaginechina.com/showImageDetail.ic?type=NewsImage&id=pbu722442_06&story=pbu722442&reqid=null&typestr=news_story_image_en_view, so by replacing the ID, we can get this URL: http://www.imaginechina.com/showImageDetail.ic?type=NewsImage&id=20180511_83981&story=pbu722442&reqid=null&typestr=news_story_image_en_view. The "download" link only works when logged in though, and it seems to be a manual registration (so I admittedly shouldn't have used 123 456 7890 for my phone number for the site haha)

qsniyg commented 5 years ago

For what it's worth, since version 0.5.7, it now supports removing the watermark from vcg.com:

https://goss.vcg.com/editorial/vcg/800/new/VCG111152795068.jpg

to:

https://goss.vcg.com/editorial/vcg/nowarter800/new/VCG111152795068.jpg