Currently, a title is matched in the search phrase, so a title "He" will match "The" in an utterance. This should be updated to only search words, not substrings.
Initial Implementation Requirements
Update search logic to search on boundaries
Implement unit tests
Other Considerations
Does this realistically only apply to one-word titles?
Can/should we boost or lower confidence based on percent of substring matched?
Relating to confidence based on substrings matched, confidence could be reduced with track/album/artist length since a longer string must have matched more of the utterance
Objective
Currently, a title is matched in the search phrase, so a title "He" will match "The" in an utterance. This should be updated to only search words, not substrings.
Initial Implementation Requirements
Other Considerations