New features and improvements
- Alexa Speak directive “caption” support: Starting with the SpeechSynthesizer v1.3 directive, a “caption” property was added to the Alexa responses. After cleaning and formatting it a little bit, it is now included in our Alexa Virtual devices response (also as a “caption” property).
In turn, this means that you can also access it via our e2e tests like this:
We’ll still support the transcript as our main property to do assertions on (as that is available across multiple voice platforms) but, if you are testing Alexa skills with the SpeechSynthesizer v1.3 directive, we encourage you to try this property as it allows for more reliable results. For comparison, this is how an assertion against a transcript might look like with the same example:
- Lenient mode: Another way in which you can improve your test success rate while dealing with transcripts is to use our new lenient mode configuration inside your testing.json file and set it to true. By doing this, you are specifically stating that, even if the transcript does include punctuation signs, you want them to be ignored. Also, it will ignore any extra whitespace that might appear as a product of doing speech to text transformation. More info here.
Fixed an issue that prevented the tests from running correctly if the specified testing.json file was in a parent directory from the running directory. bst#649