Dear Aasim
If you remember, Self- information of any symbol x is defined as I(x) = log2(1/p(x))
it means, it does not depend on CODES, or huffman codes. It says how many bits it should be given
In short: you just need to know probability of symbol x to find its self information, that's given in question (mostly)
Refer to Lec-2 Handouts, example after self -information
I hope it clarifies your doubt.
by the way attached chapter on Huffman is quite good.