How to programmatically determine the uniqueness of text in search engines?

Question

Search Engine Optimization

I wonder how services like copyscape, antiplagiat.ru determine the uniqueness of the text?

Answer 1

Most likely, they are looking for similar documents. And if the studied text is very similar to any other by some metric, then it is considered a copy. Perhaps the same is done at the paragraph level.
How to find similar documents quickly - LSH (locality sensitive hashing) and clustering.

Answer 2

Use shingles. That is, they randomly take a shingle from the text (usually they use shingles, I don’t remember exactly, from 5 to 9 words) and request it in quotes in the search. If there are more than 1 results, then someone copy-pasted someone. And here the algorithm of the search engines themselves begins to work to determine the original, moreover, it does not always correctly determine the original source.