Homophone Detection

Search…
Homophone Detection
One of the problems that arises with alpha ordinals is that some English letters sound like other words. So for example, pronouncing "C" sounds like "see" or "sea". The letter "B" sounds like "bee" or "be". These are called homophones.
If your user is using a voice assistant to dictate their responses to your bot, it's likely the voice assistant will substitute one of these homophones for the actual letter and your bot would fail to match the user's message.
This is where homophone detection helps.
Homophone detection is currently only supported for the English language.
Example
Given this code snippet from the Alpha Ordinal Matcher docs:
1
def get_response
2
  handle_message(
3
    'Red' => proc { current_user.update_attributes!(favorite_color: 'red') },
4
    'Blue' => proc { current_user.update_attributes!(favorite_color: 'blue') },
5
    'Green' => proc { current_user.update_attributes!(favorite_color: 'green') },
6
    'Yellow' => proc { current_user.update_attributes!(favorite_color: 'yellow') }
7
  )
8
end
Copied!
If a user enters the letter "b", then the "Blue" match arm will automatically be selected as expected. However, Xip will also natively accept the input "be" and "bee" from the user and the "Blue" match arm will still be selected. Similarly, "see" and "sea" will select the "Green" match arm.
For a full list of homophones Xip detects, you can inspect the array constant Xip::Controller::Messages::HOMOPHONES
If you attempt to use a homophone as your match expression, Xip will raise a Xip::Errors::ReservedHomophoneUsed exception.
Alpha Ordinal Matcher
NLP Matcher
Last modified 10mo ago
Copy link
Contents
Example