Smallest dictionary segmentation

Segment a string with no spaces into a sequence of dictionary tokens. Given s and words (a list of allowed tokens), return the lexicographically SMALLEST valid tokenization as a list of tokens, where 'lexicographically smallest' compares the resulting list of tokens element by element (shorter list is NOT automatically smaller; standard Python list comparison of the token lists). If no segmentation exists, return an empty list. s is lowercase letters; words may contain duplicates; the empty string segments to [].