Rhasspy only recognises one intent ignoring all the others

Help
1

Rhaspy version: 2.5.9
Setup:

{
    "dialogue": {
        "system": "rhasspy"
    },
    "handle": {
        "system": "hass"
    },
    "home_assistant": {
        "access_token": :)...,
        "handle_type": "event",
        "url": "http://"
    },
    "intent": {
        "fuzzywuzzy": {
            "min_confidence": "0"
        },
        "system": "fsticuffs"
    },
    "microphone": {
        "system": "pyaudio"
    },
    "sounds": {
        "system": "aplay"
    },
    "speech_to_text": {
        "system": "pocketsphinx"
    },
    "text_to_speech": {
        "espeak": {
            "voice": "ru",
            "volume": "1"
        },
        "system": "espeak"
    },
    "wake": {
        "porcupine": {
            "keyword_path": "terminator_linux.ppn"
        },
        "system": "porcupine"
    }
}

Hello, I have recorded a very strange BUG or maybe I am doing something wrong, but rhasspy does not do intent recognition as it is supposed to do.
So, I have 2 intents for my test - one is a big one, which works fine, this was my first intent I was working on, and later I had to add another one. When I started to test I the second one, I discovered, that no metter what I do, not STT is going on, but when I test the first intent, STT goes as it has to, and intent recognition takes place. Below is the text of the intent file and the logs (do not pay attention to the word in russian, it does not metter what language):


[test_scene_request]

test_data = (мама | папа | дочь){person}

test_action = (мыла раму | ела кашу | пел песню)

<test_data><test_action>

[lighting_scene_set_devices]
scene_device = ($light_devices){scene}
scene_mode = ($lighting_modes){scene_mode} 
role = (0..5999 | постоянно:0){time} [секунд]  
presence = ((присутствие):in | (отсутствие):out | (перехода):transition){presence}

a = ($lighting_actions){action}
b = освещение [в] ($rooms ){room}
c = [для: ] ($light_devices){scene}
d = [((во | на) время | когда) <presence>]
e = [на <role>]
f = [в] [режим] [<scene_mode>] [режим]
 
<a><b><c><d><e><f>
<a><b><c><d><f><e>
<a><b><c><e><d><f>
<a><b><c><e><f><d>
<a><b><c><f><e><d>
<a><b><c><f><d><e>
<a><b><d><c><e><f>
<a><b><d><c><f><e>
<a><b><d><f><c><e>
<a><b><d><f><e><c>
<a><b><d><e><f><c>
<a><b><d><e><c><f>
<a><b><f><c><d><e>
<a><b><f><c><e><d>
<a><b><f><d><c><e>
<a><b><f><d><e><c>
<a><b><f><e><c><d>
<a><b><f><e><d><c>
<a><b><e><c><d><f>
<a><b><e><c><f><d>
<a><b><e><d><c><f>
<a><b><e><d><f><c>
<a><b><e><f><c><d>
<a><b><e><f><d><c>

Logs when I call 'test_scene_request' speech:

[DEBUG:2021-02-02 02:49:46,287] rhasspyserver_hermes: Handling NluIntentNotRecognized (topic=hermes/nlu/intentNotRecognized, id=4759292f-1b87-4db3-8ecb-9b59661160c2)
[DEBUG:2021-02-02 02:49:46,287] rhasspyserver_hermes: <- NluIntentNotRecognized(input='', site_id='default', id='27335fd8-3676-498c-bc9b-4ea216eec8eb', custom_data=None, session_id='27335fd8-3676-498c-bc9b-4ea216eec8eb')
[DEBUG:2021-02-02 02:49:46,285] rhasspyserver_hermes: Publishing 189 bytes(s) to hermes/nlu/query
[DEBUG:2021-02-02 02:49:46,285] rhasspyserver_hermes: -> NluQuery(input='', site_id='default', id='27335fd8-3676-498c-bc9b-4ea216eec8eb', intent_filter=None, session_id='27335fd8-3676-498c-bc9b-4ea216eec8eb', wakeword_id=None, lang=None)
[DEBUG:2021-02-02 02:49:46,278] rhasspyserver_hermes: Handling AsrTextCaptured (topic=hermes/asr/textCaptured, id=ee04f80a-ddfd-42e6-b940-36afc65cf073)
[DEBUG:2021-02-02 02:49:46,036] rhasspyserver_hermes: Publishing 74 bytes(s) to hermes/asr/stopListening
[DEBUG:2021-02-02 02:49:46,036] rhasspyserver_hermes: -> AsrStopListening(site_id='default', session_id='2708748d-220f-4e53-b3b0-5b56c25a8fd9')
[DEBUG:2021-02-02 02:49:46,036] rhasspyserver_hermes: Sent 123540 byte(s) of WAV data
[DEBUG:2021-02-02 02:49:46,035] rhasspyserver_hermes: Publishing 181 bytes(s) to hermes/asr/startListening
[DEBUG:2021-02-02 02:49:46,035] rhasspyserver_hermes: -> AsrStartListening(site_id='default', session_id='2708748d-220f-4e53-b3b0-5b56c25a8fd9', lang=None, stop_on_silence=False, send_audio_captured=True, wakeword_id=None, intent_filter=None)
[DEBUG:2021-02-02 02:49:45,902] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-02 02:49:42,057] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-02 02:49:41,923] rhasspynlu.ini_jsgf: Loaded ini file

Logs when I call 'lighting_scene_set_devices' speech:
[DEBUG:2021-02-02 02:49:46,278] rhasspyserver_hermes: Handling AsrTextCaptured (topic=hermes/asr/textCaptured, id=ee04f80a-ddfd-42e6-b940-36afc65cf073)
[DEBUG:2021-02-02 02:49:46,036] rhasspyserver_hermes: Publishing 74 bytes(s) to hermes/asr/stopListening
[DEBUG:2021-02-02 02:49:46,036] rhasspyserver_hermes: -> AsrStopListening(site_id='default', session_id='2708748d-220f-4e53-b3b0-5b56c25a8fd9')
[DEBUG:2021-02-02 02:49:46,036] rhasspyserver_hermes: Sent 123540 byte(s) of WAV data
[DEBUG:2021-02-02 02:49:46,035] rhasspyserver_hermes: Publishing 181 bytes(s) to hermes/asr/startListening
[DEBUG:2021-02-02 02:49:46,035] rhasspyserver_hermes: -> AsrStartListening(site_id='default', session_id='2708748d-220f-4e53-b3b0-5b56c25a8fd9', lang=None, stop_on_silence=False, send_audio_captured=True, wakeword_id=None, intent_filter=None)
[DEBUG:2021-02-02 02:49:45,902] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-02 02:49:42,057] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-02 02:49:41,923] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-02 02:49:33,593] rhasspyserver_hermes: Handling NluIntent (topic=hermes/intent/lighting_scene_set_devices, id=fa29d95f-2107-439e-b727-fa75b3c8c834)
[DEBUG:2021-02-02 02:49:33,593] rhasspyserver_hermes: <- NluIntent(input='activate освещение в столовая на 30 секунд режим rose когда in потолочные светильники', intent=Intent(intent_name='lighting_scene_set_devices', confidence_score=1.0), site_id='default', id='b661c5f7-46b6-4e97-894e-f6f070f357a6', slots=[Slot(entity='lighting_actions', value={'kind': 'Unknown', 'value': 'activate'}, slot_name='action', raw_value='установить', confidence=1.0, range=SlotRange(start=0, end=8, raw_start=0, raw_end=10)), Slot(entity='rooms', value={'kind': 'Unknown', 'value': 'столовая'}, slot_name='room', raw_value='столовая', confidence=1.0, range=SlotRange(start=21, end=29, raw_start=23, raw_end=31)), Slot(entity='rhasspy/number', value={'kind': 'Number', 'value': 30}, slot_name='time', raw_value='тридцать', confidence=1.0, range=SlotRange(start=33, end=35, raw_start=35, raw_end=43)), Slot(entity='lighting_modes', value={'kind': 'Unknown', 'value': 'rose'}, slot_name='scene_mode', raw_value='розовый', confidence=1.0, range=SlotRange(start=49, end=53, raw_start=57, raw_end=64)), Slot(entity='presence', value={'kind': 'Unknown', 'value': 'in'}, slot_name='presence', raw_value='присутствие', confidence=1.0, range=SlotRange(start=60, end=62, raw_start=71, raw_end=82)), Slot(entity='light_devices', value={'kind': 'Unknown', 'value': 'потолочные светильники'}, slot_name='scene', raw_value='потолочные светильники', confidence=1.0, range=SlotRange(start=63, end=85, raw_start=83, raw_end=105))], session_id='b661c5f7-46b6-4e97-894e-f6f070f357a6', custom_data=None, asr_tokens=[[AsrToken(value='activate', confidence=1.0, range_start=0, range_end=8, time=None), AsrToken(value='освещение', confidence=1.0, range_start=9, range_end=18, time=None), AsrToken(value='в', confidence=1.0, range_start=19, range_end=20, time=None), AsrToken(value='столовая', confidence=1.0, range_start=21, range_end=29, time=None), AsrToken(value='на', confidence=1.0, range_start=30, range_end=32, time=None), AsrToken(value='30', confidence=1.0, range_start=33, range_end=35, time=None), AsrToken(value='секунд', confidence=1.0, range_start=36, range_end=42, time=None), AsrToken(value='режим', confidence=1.0, range_start=43, range_end=48, time=None), AsrToken(value='rose', confidence=1.0, range_start=49, range_end=53, time=None), AsrToken(value='когда', confidence=1.0, range_start=54, range_end=59, time=None), AsrToken(value='in', confidence=1.0, range_start=60, range_end=62, time=None), AsrToken(value='потолочные', confidence=1.0, range_start=63, range_end=73, time=None), AsrToken(value='светильники', confidence=1.0, range_start=74, range_end=85, time=None)]], asr_confidence=None, raw_input='установить освещение в столовая на тридцать секунд режим розовый когда присутствие потолочные светильники', wakeword_id=None, lang=None)
[DEBUG:2021-02-02 02:49:30,603] rhasspyserver_hermes: Publishing 294 bytes(s) to hermes/nlu/query
[DEBUG:2021-02-02 02:49:30,602] rhasspyserver_hermes: -> NluQuery(input='установить освещение в столовая на тридцать секунд режим розовый когда присутствие потолочные светильники', site_id='default', id='b661c5f7-46b6-4e97-894e-f6f070f357a6', intent_filter=None, session_id='b661c5f7-46b6-4e97-894e-f6f070f357a6', wakeword_id=None, lang=None)
[DEBUG:2021-02-02 02:49:30,458] rhasspyserver_hermes: Handling AsrTextCaptured (topic=hermes/asr/textCaptured, id=8f674fb7-5669-4bbf-b5e0-9246c9ca67ff)
[DEBUG:2021-02-02 02:49:29,538] rhasspyserver_hermes: Publishing 74 bytes(s) to hermes/asr/stopListening
[DEBUG:2021-02-02 02:49:29,538] rhasspyserver_hermes: -> AsrStopListening(site_id='default', session_id='bad07fd5-ca1e-4e59-8bf8-a5f842224938')
[DEBUG:2021-02-02 02:49:29,537] rhasspyserver_hermes: Sent 414404 byte(s) of WAV data
[DEBUG:2021-02-02 02:49:29,531] rhasspyserver_hermes: Publishing 181 bytes(s) to hermes/asr/startListening
[DEBUG:2021-02-02 02:49:29,530] rhasspyserver_hermes: -> AsrStartListening(site_id='default', session_id='bad07fd5-ca1e-4e59-8bf8-a5f842224938', lang=None, stop_on_silence=False, send_audio_captured=True, wakeword_id=None, intent_filter=None)
[DEBUG:2021-02-02 02:49:29,223] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-02 02:49:16,347] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-02 02:49:16,271] rhasspynlu.ini_jsgf: Loaded ini file

As you can see, when I say something that calls big intent - it processes the STT, but when I say something different no STT processing happends.

Is there something I can do about it?

2

Looks like you intents are so unbalanced that the complex one completely overruns the small one.

You should simplify your big intent (split it into multiple smaller intents, avoid huge numbers range, large slot values list, etc.).

3

Thanks for the suggestion. At the moment I tried some variants, but it works awfull… I shall keep trying, but I’d like to question one thing - how do you understand that that this is intent’s problem? As a metter of fact I use mobile app to send commands. It has wonderfull visualisation feature - after you stop recording, it recieves back the recognised text of your speach and prints it into the text window. When I say something relevant to the big intent - there is a text, when irrelevant - there is no text. Hence, I concluded that it is some STT problem, which does not generate text string for all instances. And yet it would be helpfull to know the rules for intent design - where does the intent start being ‘big’? As I said,
the big intent was tested in all possible ways thus it’s syntax is formally correct - what are the criteria?
And another question - maybe there is a better way to solve my problem with big intent? What it does - I have a function which takes 6 parameters - this function understands everything you may think about the light. From text message I have to derrive these 6 parameters. There is a formal guide that these 6 parameters should be in a sentence, but their order is arbitrary, because otherwise no human being would be able to use it. So, I split a sentence into 6 pieces containing eachb, of the parameters and made all possible (those should be 120 actually, but vital are only 24 of those) variants those 6 segments could be composed togather in real life. Thus I have 24 sentences with chunks a, b, c, d, e, f. As I understand, if I only keep one order - <a><b><c><d><e><f>, all other orders wont be recognised. Is there built in methods to tell the progrann that all possible permutations of the elements should be considered?

Thanks alot, if you have read this to this point!

4

You should try to use Kaldi instead of Pocketsphinx (which is pretty awful).

The issue is very similar to one I had a year ago with Rhasspy:

Maybe this can help:

5

I would, but it is even worse - it does not support my language and the table says that so far the Pocketsphinx is the only choise…

6

What language are you using?

7

russian. I’ve read your topic from a year ago. You have big experience in this area. Do you think using RasaNLU for the intent recognition woud make things better? Because I tried fuzzywuzzy with threshold 0.5 + Pocketsphinx today and got this for all attemts to recognise anything:

[ERROR:2021-02-02 23:48:28,412] rhasspyserver_hermes: 41
Traceback (most recent call last):
  File "/usr/lib/rhasspy/.venv/lib/python3.7/site-packages/quart/app.py", line 1821, in full_dispatch_request
    result = await self.dispatch_request(request_context)
  File "/usr/lib/rhasspy/.venv/lib/python3.7/site-packages/quart/app.py", line 1869, in dispatch_request
    return await handler(**request_.view_args)
  File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/__main__.py", line 1455, in api_text_to_intent
    user_entities=user_entities,
  File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/__main__.py", line 2705, in text_to_intent_dict
    result = await core.recognize_intent(text, intent_filter=intent_filter)
  File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/__init__.py", line 524, in recognize_intent
    raise NluException(result.error)
rhasspyserver_hermes.NluException: 41
[ERROR:2021-02-02 23:48:28,412] rhasspyserver_hermes: NluError(error='41', site_id='mobile-app1', context='установить освещение столовая в режим', session_id='30a0fefe-62ad-4f48-836f-e086e1cbffbc')
8

Russian is available for Kaldi in Rhasspy (according to the doc and Github repo).

This is more an issue with the ASR (STT) than with the NLU part so using Rasa should not improve things.

Can you try using Kaldi and fiddling with language model type?

9

Yes, I will try again. Last time it did not recognise the words well, so I switched to Pocketsphinx. This time it downloaded some new file when I set it to run with fuzzuwuzzy and mixed model. Maybe this would improove the performance…

10

When I use fuzzywuzzy I get this error, does anyone have any idea what it is?

[ERROR:2021-02-03 00:57:13,985] rhasspyserver_hermes: 46
Traceback (most recent call last):
  File "/usr/lib/rhasspy/.venv/lib/python3.7/site-packages/quart/app.py", line 1821, in full_dispatch_request
    result = await self.dispatch_request(request_context)
  File "/usr/lib/rhasspy/.venv/lib/python3.7/site-packages/quart/app.py", line 1869, in dispatch_request
    return await handler(**request_.view_args)
  File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/__main__.py", line 1455, in api_text_to_intent
    user_entities=user_entities,
  File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/__main__.py", line 2705, in text_to_intent_dict
    result = await core.recognize_intent(text, intent_filter=intent_filter)
  File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/__init__.py", line 524, in recognize_intent
    raise NluException(result.error)
rhasspyserver_hermes.NluException: 46
[ERROR:2021-02-03 00:57:13,985] rhasspyserver_hermes: NluError(error='46', site_id='mobile-app1', context='сделай в кухня розовый на тридцать секунд потолочные светильники', session_id='8eae5b5e-97a0-4fcb-a8f9-5a4b290f5d42')

When I use fsticuffs, there is no errors, but it does not recognise words properly again:
example: I say “set led lights in the diningroom’ and it STTs it into 'set led lights in the kitchen” - why!!! The words are miles away from each other!

11

By the way, do you by chance have any knowledge of how to set up rasa with portainer? They have this text command to start rasa in the container:
docker run -it -v “$(pwd):/app” -p 5005:5005 rasa/rasa:latest-spacy-en run --enable-api

but I cannot understand how to run this comand in the portainer. Since I am under hossos I can only access the docker via portainer, cant I?

12

It turn out, the Kaldi with fuzzywuzzy does not finish thraining and return error. I am trying to use Kaldi with fsticuffs but the result is far away from accaptible as far as proper word detection is concerned. For some reason insted of ‘diningroom’ it hears ‘kitchen’… I looked into the log after training and discoverd some ‘Skip’ actions for downlod request. I am concerned that something was not downloaded and that is the reason why I am having problems with Kaldi. What is you oppininon? I have poseted the log for the training part and my Rhasspy set up file.

[DEBUG:2021-02-03 13:46:24,122] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-03 13:46:22,260] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-03 13:46:22,107] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-03 13:43:52,552] rhasspyprofile.download: text_to_speech.system larynx marytts = False
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: speech_to_text.kaldi.mix_weight >0 0 = False
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: speech_to_text.kaldi.open_transcription True False = False
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: speech_to_text.pocketsphinx.mix_weight >0 0 = False
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: speech_to_text.pocketsphinx.open_transcription True False = False
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/acoustic_model/ivectors_test_hires/conf/splice.conf (/share/rhasspy/profiles/ru/kaldi/acoustic_model/ivectors_test_hires/conf/splice.conf)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/acoustic_model/ivectors_test_hires/conf/ivector_extractor.conf (/share/rhasspy/profiles/ru/kaldi/acoustic_model/ivectors_test_hires/conf/ivector_extractor.conf)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/ivectors_test_hires/conf/online_cmvn.conf (/share/rhasspy/profiles/ru/kaldi/model/ivectors_test_hires/conf/online_cmvn.conf)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/extractor/final.ie (/share/rhasspy/profiles/ru/kaldi/model/extractor/final.ie)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/extractor/final.dubm (/share/rhasspy/profiles/ru/kaldi/model/extractor/final.dubm)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/extractor/splice_opts (/share/rhasspy/profiles/ru/kaldi/model/extractor/splice_opts)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/extractor/global_cmvn.stats (/share/rhasspy/profiles/ru/kaldi/model/extractor/global_cmvn.stats)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/extractor/online_cmvn.conf (/share/rhasspy/profiles/ru/kaldi/model/extractor/online_cmvn.conf)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/extractor/final.mat (/share/rhasspy/profiles/ru/kaldi/model/extractor/final.mat)
[DEBUG:2021-02-03 13:43:52,551] rhasspyprofile.download: Skipping kaldi/model/conf/mfcc_hires.conf (/share/rhasspy/profiles/ru/kaldi/model/conf/mfcc_hires.conf)
[DEBUG:2021-02-03 13:43:52,550] rhasspyprofile.download: Skipping kaldi/model/conf/mfcc.conf (/share/rhasspy/profiles/ru/kaldi/model/conf/mfcc.conf)
[DEBUG:2021-02-03 13:43:52,550] rhasspyprofile.download: Skipping kaldi/model/conf/online_cmvn.conf (/share/rhasspy/profiles/ru/kaldi/model/conf/online_cmvn.conf)
[DEBUG:2021-02-03 13:43:52,550] rhasspyprofile.download: Skipping kaldi/model/path.sh (/share/rhasspy/profiles/ru/kaldi/model/path.sh)
[DEBUG:2021-02-03 13:43:52,550] rhasspyprofile.download: Skipping kaldi/model/phones/silence_phones.txt (/share/rhasspy/profiles/ru/kaldi/model/phones/silence_phones.txt)
[DEBUG:2021-02-03 13:43:52,550] rhasspyprofile.download: Skipping kaldi/model/phones/optional_silence.txt (/share/rhasspy/profiles/ru/kaldi/model/phones/optional_silence.txt)
[DEBUG:2021-02-03 13:43:52,549] rhasspyprofile.download: Skipping kaldi/model/phones/extra_questions.txt (/share/rhasspy/profiles/ru/kaldi/model/phones/extra_questions.txt)
[DEBUG:2021-02-03 13:43:52,549] rhasspyprofile.download: Skipping kaldi/model/phones/nonsilence_phones.txt (/share/rhasspy/profiles/ru/kaldi/model/phones/nonsilence_phones.txt)
[DEBUG:2021-02-03 13:43:52,549] rhasspyprofile.download: Skipping kaldi/model/model/normalization.fst (/share/rhasspy/profiles/ru/kaldi/model/model/normalization.fst)
[DEBUG:2021-02-03 13:43:52,549] rhasspyprofile.download: Skipping kaldi/model/model/final.mdl (/share/rhasspy/profiles/ru/kaldi/model/model/final.mdl)
[DEBUG:2021-02-03 13:43:52,548] rhasspyprofile.download: Skipping kaldi/model/model/den.fst (/share/rhasspy/profiles/ru/kaldi/model/model/den.fst)
[DEBUG:2021-02-03 13:43:52,548] rhasspyprofile.download: Skipping kaldi/model/model/tree (/share/rhasspy/profiles/ru/kaldi/model/model/tree)
[DEBUG:2021-02-03 13:43:52,548] rhasspyprofile.download: Skipping kaldi/model/model/cmvn_opts (/share/rhasspy/profiles/ru/kaldi/model/model/cmvn_opts)
[DEBUG:2021-02-03 13:43:52,548] rhasspyprofile.download: Skipping kaldi/g2p.fst (/share/rhasspy/profiles/ru/kaldi/g2p.fst)
[DEBUG:2021-02-03 13:43:52,547] rhasspyprofile.download: Skipping kaldi/g2p.corpus (/share/rhasspy/profiles/ru/kaldi/g2p.corpus)
[DEBUG:2021-02-03 13:43:52,547] rhasspyprofile.download: Skipping kaldi/base_dictionary.txt (/share/rhasspy/profiles/ru/kaldi/base_dictionary.txt)
[DEBUG:2021-02-03 13:43:52,547] rhasspyprofile.download: speech_to_text.system kaldi kaldi = True
[DEBUG:2021-02-03 13:43:52,547] rhasspyprofile.download: speech_to_text.system pocketsphinx kaldi = False
[DEBUG:2021-02-03 13:43:52,350] rhasspyserver_hermes: [AsrTrainSuccess(id='ac2b2789-2242-452e-acc5-9e45c6bd7ec0'), NluTrainSuccess(id='ac2b2789-2242-452e-acc5-9e45c6bd7ec0')]
[DEBUG:2021-02-03 13:43:51,733] rhasspyserver_hermes: Handling AsrTrainSuccess (topic=rhasspy/asr/mobile-app1/trainSuccess, id=ce4f1ab9-8200-4a92-b645-f28392a0d33d)
[DEBUG:2021-02-03 13:43:13,730] rhasspyserver_hermes: Handling NluTrainSuccess (topic=rhasspy/nlu/mobile-app1/trainSuccess, id=ce4f1ab9-8200-4a92-b645-f28392a0d33d)
[DEBUG:2021-02-03 13:43:10,174] rhasspyserver_hermes: Publishing 48625 bytes(s) to rhasspy/nlu/mobile-app1/train
[DEBUG:2021-02-03 13:43:10,174] rhasspyserver_hermes: -> NluTrain
[DEBUG:2021-02-03 13:43:10,159] rhasspyserver_hermes: Publishing 48625 bytes(s) to rhasspy/asr/mobile-app1/train
[DEBUG:2021-02-03 13:43:10,159] rhasspyserver_hermes: -> AsrTrain
[DEBUG:2021-02-03 13:43:10,135] rhasspyserver_hermes: Subscribed to hermes/error/nlu
[DEBUG:2021-02-03 13:43:10,135] rhasspyserver_hermes: Subscribed to rhasspy/nlu/mobile-app1/trainSuccess
[DEBUG:2021-02-03 13:43:10,135] rhasspyserver_hermes: Subscribed to hermes/error/asr
[DEBUG:2021-02-03 13:43:10,135] rhasspyserver_hermes: Subscribed to rhasspy/asr/mobile-app1/trainSuccess
[DEBUG:2021-02-03 13:43:10,134] rhasspyserver_hermes: Finished writing /share/rhasspy/profiles/ru/intent_graph.pickle.gz
[DEBUG:2021-02-03 13:42:42,645] rhasspyserver_hermes: Writing /share/rhasspy/profiles/ru/intent_graph.pickle.gz
[DEBUG:2021-02-03 13:42:34,938] rhasspynlu.slots: Running program for slot rhasspy/number,0,5999,1: ['/usr/lib/rhasspy/rhasspy-profile/rhasspyprofile/profiles/ru/slot_programs/rhasspy/number', '0', '5999', '1']
[DEBUG:2021-02-03 13:42:34,938] rhasspynlu.slots: Loading slot light_devices from /share/rhasspy/profiles/ru/slots/light_devices
[DEBUG:2021-02-03 13:42:34,937] rhasspynlu.slots: Loading slot rooms from /share/rhasspy/profiles/ru/slots/rooms
[DEBUG:2021-02-03 13:42:34,937] rhasspynlu.slots: Loading slot lighting_actions from /share/rhasspy/profiles/ru/slots/lighting_actions
[DEBUG:2021-02-03 13:42:34,936] rhasspynlu.slots: Loading slot lighting_modes from /share/rhasspy/profiles/ru/slots/lighting_modes
[DEBUG:2021-02-03 13:42:34,913] rhasspynlu.ini_jsgf: Loaded ini file
[DEBUG:2021-02-03 13:42:34,912] rhasspyserver_hermes: Generating intent graph
[DEBUG:2021-02-03 13:42:34,912] rhasspyserver_hermes: Loading sentences from [PosixPath('/share/rhasspy/profiles/ru/sentences.ini')]
[INFO:2021-02-03 13:42:34,912] rhasspyserver_hermes: Starting training

{
    "command": {
        "webrtcvad": {
            "max_sec": "30",
            "silence_method": "all",
            "vad_mode": "3"
        }
    },
    "dialogue": {
        "system": "rhasspy"
    },
    "handle": {
        "system": "hass"
    },
    "home_assistant": {
        "access_token": ".token",
        "handle_type": "event",
        "url": "http://192.168.1.3:8123/"
    },
    "intent": {
        "fsticuffs": {
            "fuzzy": false
        },
        "fuzzywuzzy": {
            "min_confidence": "0"
        },
        "system": "fsticuffs"
    },
    "microphone": {
        "system": "pyaudio"
    },
    "mqtt": {
        "enabled": "true",
        "host": "192.168.1.3",
        "password": "pass",
        "site_id": "mobile-app1",
        "username": "name"
    },
    "sounds": {
        "system": "aplay"
    },
    "speech_to_text": {
        "kaldi": {
            "mix_weight": "0"
        },
        "pocketsphinx": {
            "mix_weight": "0"
        },
        "system": "kaldi"
    },
    "text_to_speech": {
        "espeak": {
            "voice": "ru",
            "volume": "1"
        },
        "marytts": {
            "locale": "ru",
            "url": "http://192.168.1.3:59125/process",
            "voice": "ac-irina-hsmm"
        },
        "system": "marytts"
    },
    "wake": {
        "porcupine": {
            "keyword_path": "porcupine_linux.ppn",
            "sensitivity": "0.9",
            "udp_audio": "0.0.0.0:12333:mobile-app1"
        },
        "system": "porcupine"
    }
}