How to design an efficient algorithm to support random access in LZ77 compressed sequences?

I'm working on a project in text compression and I need to design an efficient algorithm in a LZ77 compressed sequences. Specially, Given a LZ77 compressed sequence and an index i, we can recover a single symbol S[i] of the input sequence S. The space consumed by the algorithm and the time of random access to a symbol are what we pursue。

Thanks in advance for your suggestions.