term_count is slow

think about rewriting it with hashmaps

Right now we are using regex which is slow 

Here is a simple solution:
put elements of the input string in a hash map (container map)
compare the contents of that container map with the keys of the header
if exists output the values
