Error returned for gl_speech_op() about googlelanguager HOT 3 CLOSED

Thanks @smmurphy - I have never tried it on an audio file that long, so will need to check it to reproduce. That error is when the API response is unexpected format, so perhaps its empty or with a message saying something like "Your file is too large" - nevertheless it would be good to track down, will take a closer look.

from googlelanguager.

Thanks for looking into this @MarkEdmondson1234 . I got a little further with this... what I did is regenerate call_api() with the parse function as: parse_async instead of parse_op. This initially threw an error due to the if statement:

if(!is.null(x$metadata$startTime)){
    my_message("Speech transcription running - started at ", x$metadata$startTime,
               " - last update: ", x$metadata$lastUpdateTime, level = 3)
  }

So I made a local copy of the function that was just...

# parse asynchronous speech calls responses
parse_async <- function(x) {
  structure(x, class = "gl_speech_op")
}

Running that returned the request (though not in as convenient a format).

I'm not sure exactly why, likely something to do with metadata handling, but it appears that parse_op is defaulting to parse_speech, which is what I was getting for the 5 minute files and working. For the longer file it still defaults to parse_speech, but that fails whereas parse_async works.

from googlelanguager.

Thanks @smmurphy that is exactly what I would have done - it would be really useful if you can report the results of str(x) of the object you get with your local copy of parse_async() - then I can modify the existing one to be more forgiving and accept the response you are getting. It sounds odd, that in your case the x$metadata$startTime isn't present, but a quick enough fix to make it conditional.

from googlelanguager.