SearchInputParsingUtils sanitize should replace special characters with spaces, not remove them

SearchInputParsingUtils removes some special characters to sanitize the query.
Most special characters are word boundaries, by just removing these the information on word boundary is lost, creating unwanted concatenated words.
Example Bagels&Beans -> BagelsBeans
When replacing with a space query will conserve the information that query has two words: Bagels Beans
Suggestion to make a new method so original behavior is not changed

Reference
http://unicode.org/reports/tr29/#Word_Boundaries