An input module of ODAS beamformed sources for Rhasspy and a sensor with DOA capabilities and activity identification. In the direction of providing a more robust signal to noise ratio for infering specific intents with rhasspy.
This module is based on the rhasspy-pyaudio module and records audio from ODAS and publishes WAV chunks according to the Hermes protocol. Furthermore, specific messages for the DOA from ODAS of localized (potential) sources and tracked sources are produced. See lisa rhasppy messages
VERY IMPORTANT, in all cases at today the lisa-odas library must be build separately
Adjustment in the path could be necessary
- Python 3.7
$ docker run -it rhasspy/rhasspy-lisa-odas-hermes:<VERSION> <ARGS>Clone (including submodules) the repository and create the virtual environment:
$ git clone --recurse-submodules https://github.com/lawrence-iviani/rhasspy-lisa-odas-hermes.git
$ cd rhasspy-lisa-odas-hermes
## rhasspy run in virtual env
$ ./configure
## or use the system environment
$ ./configure --enable-in-place
$ make
$ make installRun the rhasspy-lisa-odas-hermes script to access the command-line interface:
$ ./rhasspy-lisa-odas-hermes --helpNOT TESTED
Run scripts/build-docker.sh with a local docker registry:
$ make dist$ bin/rhasspy-lisa-odas-hermes <ARGS>Requires Docker Buildx. Set PLATFORMS environment to only build for specific platforms (e.g., linux/amd64).
This will create a Docker image tagged rhasspy/rhasspy-lisa-odas-hermes:<VERSION> where VERSION comes from the file of the same name in the source root directory.
NOTE: If you add things to the Docker image, make sure to whitelist them in .dockerignore.
Based on the pyaudio module, although they are tunable with several personalization they must be equal as defined in the ODAS revceiver. The default values are defined in config_file/lisa.cfg and must match the ODAS configuration file as defined in the field [ODAS] odas_config.
usage: rhasspy-lisa-odas-hermes [-h] [--sample-rate SAMPLE_RATE]
[--sample-width SAMPLE_WIDTH]
[--channels CHANNELS] [--demux]
[--output-site-id OUTPUT_SITE_ID]
[--udp-audio-host UDP_AUDIO_HOST]
[--udp-audio-port UDP_AUDIO_PORT]
[--odas-config ODAS_CONFIG] [--host HOST]
[--port PORT] [--username USERNAME]
[--password PASSWORD] [--tls]
[--tls-ca-certs TLS_CA_CERTS]
[--tls-certfile TLS_CERTFILE]
[--tls-keyfile TLS_KEYFILE]
[--tls-cert-reqs {CERT_REQUIRED,CERT_OPTIONAL,CERT_NONE}]
[--tls-version TLS_VERSION]
[--tls-ciphers TLS_CIPHERS]
[--site-id SITE_ID] [--debug]
[--log-format LOG_FORMAT]
optional arguments:
-h, --help show this help message and exit
--sample-rate SAMPLE_RATE
Sample rate of recorded audio in hertz (e.g., 16000)
--sample-width SAMPLE_WIDTH
Sample width of recorded audio in bytes (e.g., 2)
--channels CHANNELS Number of channels in recorded audio (e.g., 1)
--demux Stream always one channel out by selecting the one
with higher priority (priority mode latest source). If
channels is provided is then discarded and only one
streamed
--output-site-id OUTPUT_SITE_ID
If set, output audio data to a different site id
--udp-audio-host UDP_AUDIO_HOST
Host for UDP audio (default: 127.0.0.1)
--udp-audio-port UDP_AUDIO_PORT
Send raw audio to UDP port outside ASR listening
--odas-config ODAS_CONFIG
ODAS configuration, default is: lisa-
odas/config/matrix-lisa/matrix_voice_LISA_1.cfg
--host HOST MQTT host (default: localhost)
--port PORT MQTT port (default: 1883)
--username USERNAME MQTT username
--password PASSWORD MQTT password
--tls Enable MQTT TLS
--tls-ca-certs TLS_CA_CERTS
MQTT TLS Certificate Authority certificate files
--tls-certfile TLS_CERTFILE
MQTT TLS certificate file (PEM)
--tls-keyfile TLS_KEYFILE
MQTT TLS key file (PEM)
--tls-cert-reqs {CERT_REQUIRED,CERT_OPTIONAL,CERT_NONE}
MQTT TLS certificate requirements (default:
CERT_REQUIRED)
--tls-version TLS_VERSION
MQTT TLS version (default: highest)
--tls-ciphers TLS_CIPHERS
MQTT TLS ciphers to use
--site-id SITE_ID Hermes site id(s) to listen for (default: all)
--debug Print DEBUG messages to the console
--log-format LOG_FORMAT
Python logger format
https://arxiv.org/abs/1412.5567
As an example from the forked repos, CI not implemented yet
