Find phonetic siblings

Validate slot values with Double Metaphone algorithm to find phonetic equivalents was already introduced in the last release. However, this algorithm is optimized for English language. AlexaInput now has two more helper methods to:

• validate slot values with Cologne Phonetic algorithm - optimized for German language
• validate slot values and automatically leverage the best algorithm for the current locale. this is where the SDK automatically picks either Double Metaphone or Cologne Phonetic algorithm depending on the locale coming in with a speechlet request

Minor improvements

• Wildcard intent to listen for "any" intent handler
• Support YAML-internal placeholders
• Support multiphrases in placeholders
• Support placeholders in multiphrases
• New slot value verification methods in AlexaInput

Initial release

Key features

• Your entire code is released from having output speech hardcoded in your skill, because ..
• Response utterances and reprompts are sourced out to YAML files
• YAML files can be part of your JAR or can be stored in an AWS S3 bucket
• Have multiple YAML-files - one for each locale. That makes it easy to build
  multi-language skills without code redundancy
• Have slots in those utterances which the engine fills up with values
  from your POJO models or from an output given by your intent handlers
• Have multi-phrase-collections in your utterance to vary Alexa's replies
• Compatible with state models from Alexa States SDK which does all the
  state management of your POJO models by using S3, Dynamo or Alexa session as a store.
• Speechlet handlers (either Lambda or Servlet) configurable with annotations
• Auto-registered intent listeners subscribe for one to many intents matching
  custom criteria - it was never that flexible to react on intents
• Explicit exception handling in intent handlers to have Alexa react accordingly
  in any situation