How can I select all images that are not inside of a link element?
document.select("a img"); //selects all images inside a link
document.select(":not(a) img"); //images not inside a link (does not work)
Ok, so the problem here is that :not(a) img needs just one element around the <img> which is not an <a> containing an <img>. For example <body> matches for :not(a). So your selector matches nearly all <img> tags. Even if you pass an HTML string to Jsoup.parse() which doesn't have a <body> or <html> tag. Jsoup automatically generates it.
Let's assume we have the following HTML:
<html>
<body>
<a><div><img id="a-div-img"></div></a>
<a><img id="a-img"></a>
<img id="img">
</body>
</html>
If you just want to exclude direct <img> childs in <a> you can use :not(a) > img as selector:
Elements images = document.select(":not(a) > img");
The result will be this:
<img id="a-div-img">
<img id="img">
The problem with this is that it also prints the first <img> of the example, which is actually inside an <a> (#a-div-img). If this in enough to fit your needs you can go with this solution.
Excluding all <a> tags from the selection is not possible with a pure CSS (at least I didn't find a solution yet). But you can just remove all <a> tags from the document before selecting all <img> tags:
document.select("a").remove();
Elements images = document.select("img");
The result will be just this:
<img id="img">
If you need the original document without modifications you can use Document.clone() before:
Document tempDocument = document.clone();
tempDocument.select("a").remove();
Elements images = tempDocument.select("img");
Using this the original document is never modified.