Extractive text summarization

You have two sentences as a post-processed word list. Compute their similarity score, according to the original TextRank sentence similarity formula:

$$\text{Sim}(S_i, S_j) = \frac{|{\{ w_k} \vert {w_k} \in S_i \& w_k \in S_j \}|}{\log(|S_i|) + \log(|S_j|)}$$

A natural $\log$ is assumed. It is guaranteed that the input's lengths are more than $2$. Return the value rounded up to four digits after the decimal point.