fergyfresh
commented
6 years ago You could definitely do it with this repository here without changing too much. You would need to make a German version of the https://github.com/stevenleeg/geemusic/blob/master/speech_assets/sampleUtterances.txt on a branch on your fork. The second thing we would have to do is some sort or YAML files with all the strings for every function or what I would like to do is just a Python dictionary that we put at the top of each file with the English/German versions of each string and we could add an environment variable for what language you want to use.
So instead of the code below you would substitute all of the English phrases for german phrases and then eventually we would need to use the dictionary/YAML method for swapping between languages. So the quick and dirty approach is to translate all of the English phrases in the selection.py file and modify all of the speech assets to be German.
You could do a relatively simple test by only changing a few things:
Make a fork and modify the sampleUtterances.txt file lines below after GeeMusicPlayThumbsUpIntent to be German so that your German Alexa knows what you're saying.
79 GeeMusicPlayThumbsUpIntent Play promoted songs
80 GeeMusicPlayThumbsUpIntent Play up voted songs
81 GeeMusicPlayThumbsUpIntent Play thumbs up songs The code that goes with the Intent above is in selection.py and you'd have to modify the English strings on lines 88, 92, and 98 to German.
86 @ask.intent("GeeMusicPlayThumbsUpIntent")
87 def play_promoted_songs():
88 app.logger.debug("Fetching songs that you have up voted.")
89
90 promoted_songs = api.get_promoted_songs()
91 if promoted_songs is False:
92 return statement("Sorry, I couldnt find any up voted songs.")
93
94 first_song_id = queue.reset(promoted_songs)
95 stream_url = api.get_stream_url(first_song_id)
96
97 thumbnail = api.get_thumbnail(queue.current_track()['albumArtRef'][0]['url'])
98 speech_text = "Playing your upvoted songs."
99 return audio(speech_text).play(stream_url) \
100 .standard_card(title=speech_text,
101 text='',
102 small_image_url=thumbnail,
103 large_image_url=thumbnail) I hope this works and if it doesn't work it probably has something to do with the fact that I don't know how to tell Alexa if an app is German or English, but either way you should try what I suggested above and test it out and come back with the results, we can work through this.
simonszu
Hi,
i have some basic skills with developing skills for the Echo. I think it could be possible to adapt your bridge for other languages, but at first, i need to know how language-agnostic the bridge daemon is. I have not checked it for myself, yet, but does it report other things than music back to the Alexa service?
My idea is to just change the utterances in the AWS skill settings, so that Alexa knows how to handle german utterances (i am from germany), and if your bridge just reports back the music to the device, all would be fine. However, the Amazon Music integration for Alexa makes Alexa say something, like a radio moderator, for example: "Alexa, play me some Metallica music" - "Here is some Metallica music".
I think it would be weird, if a german voice model tries to speak english words...but could it be generally possible?