6b5d45e507
Do not create configs folders until writing
...
This replaces save_*_path with usage of the xdg_*_home when handling
config files. This means the config folders will not be created unless
actually written to.
The check for whether a directory needs to be created is handled behind
a lock to avoid race conditions
2021-11-30 22:50:09 +01:00
3fd96cf71b
WIP Review comments
...
- Add TODO for 22.02 to remove the compatibility code
- Make Warning a single Log statement
- mycroft-config script now uses XDG-environment variable
- Remove redundant code
- Replace hard coded references to ~/.config
- Explicitly remove new path before move of "filesystem" (if needed)
2021-08-05 10:04:45 +02:00
9029dc1f41
Slight cleanup
...
- Make XDG usage more visible by using the xdg module prefix
- fix overloaded python keywords
- remove unused imports
2021-08-05 10:04:45 +02:00
e20443b824
Use XDG Base directories for settings, cache and runtime data
...
Improve deprecation warning message
2021-08-05 10:04:45 +02:00
0e2a78068a
Merge pull request #2910 from MycroftAI/feature/avx-check
...
Check for the AVX instruction set during install
2021-06-02 06:54:08 +09:30
305b93625a
Update hotword_factory.py
...
Fix long line string
2021-05-28 21:48:49 +09:30
011c3899a9
Fix line too long
2021-05-28 21:48:49 +09:30
4ca656caac
Issue-2232 Add warning when AVX is not detected and fallback.
2021-05-28 21:48:49 +09:30
e75a05557b
docstring cleanup
2021-05-11 15:10:32 +09:30
1273425799
add ProcessStatus to key services
2021-01-28 17:24:21 +01:00
73f0299cfa
Merge pull request #2722 from forslund/bugfix/latest-porcupine
...
Update for newer Porcupine engines
2021-01-25 16:29:48 +09:30
ad32a7a873
move bus connect method to process_utils
2021-01-11 16:57:25 +09:30
0423daeb4c
Voice service to use common bus cilent method
2021-01-11 16:57:25 +09:30
4747dd0e59
Fix UnboundLocalError in AudioProducer
...
==== Fixed Issues ====
==== Tech Notes ====
The variable should be declared outside the loop.
2020-12-15 10:16:34 +01:00
4b1efc7db9
Add deprecation notice
2020-12-13 21:53:01 +01:00
23cf197eb0
Pylint cleanup of imports and Porcupine class
2020-12-09 12:38:54 +01:00
cb8c3fe7de
Switch to using the pvporcupine pip package
...
This is the same set of bindings as previously but packaged and
published on pypi
Picovoce library finder functions are used to determine the location of
libraries.
2020-12-09 12:37:31 +01:00
32279894f2
Update for latest Porcupine release
...
- Update the creation of Porcupine object
- Update the detection to accept >= 0 (index of detected hot word)
2020-12-09 12:37:31 +01:00
50f049032b
Add docstrings for Hotword base class
2020-11-04 08:03:15 +01:00
0a9de0ab7c
Update to pyee 8.1.0
...
pyee 8.1.0 adds a small change to make the once call to be more safe in
multithreaded environments.
This switches back from the now deprecated BaseEventEmitter to the
standard EventEmitter.
2020-10-22 11:43:06 +02:00
02cb489f63
wake word plugins
2020-10-10 18:03:40 +01:00
6e197d0138
Bugfix/issue 2563 ( #2692 )
...
* Issue-2563 Added fallback to english model
* Added test case for checking config set up
* Added more descriptive log message
2020-09-22 15:01:01 +09:30
1202417f86
Handle ntp syncs during precise Startup
...
Use MonotonicEvent instead of the builtin Event class
2020-09-10 13:27:21 +02:00
d1ea9088fb
Merge pull request #2670 from MycroftAI/remove/startListening
...
remove deprecated startListening signal
2020-08-21 00:31:30 +00:00
d4d7d09e40
remove hardcoded timeouts; already moved to conf
2020-08-19 20:35:23 +09:30
23d1d35ea9
remove deprecated startListening signal
2020-08-18 22:21:38 +09:30
5dac7fe54c
disable wakeword uploads before deprecating API
2020-08-13 16:49:54 +09:30
7aa80e2b7d
remove duplicate and unused method
2020-08-13 16:46:08 +09:30
236a2ed594
Add hook for stopping
2020-07-22 11:06:47 +02:00
a1fdae3415
Add support for watchdog in skills and speech client
...
If no watchdog is provided a dummy function will be called
2020-07-22 10:52:32 +02:00
09b1deb511
Add possibility for notification hooks for services
...
- skills
- audio
- speech client
- messagebus service
- enclosure
The main functions now accepts the arguments ready_hook and error_hook
allowing a service or runner script to catch these states and perform
actions accordingly.
This is useful for things like systemd or a desktop launcher.
Fix audio service startup
2020-07-22 10:52:27 +02:00
0fc1ba325a
Update pyee to v7.0.1
...
pyee has added ThreadExecutor variant of the EventEmitter which can
replace the home-spun threadpool based emitter
2020-07-07 08:21:42 +02:00
926d3ff1c1
Extract noise-tracking from _record_phrase
...
_record_phrase now uses a helper class to track the noise in a sentence
to determine when the user has stopped speaking.
The logic should be the same but the same but allow for easier testing
and manipulation.
2020-05-26 12:12:12 +02:00
9e73180206
Make sec_to_bytes a member of MutableMic
...
The function is based on the parameters of the Mic and is much more a
method of the mic than a static function
2020-05-26 12:12:12 +02:00
225f22fe82
Remove unused member variables
...
Move wakeword duration to the wakeword engine since all wakeword engines
doesn't really use phonemes
2020-05-26 12:12:12 +02:00
dbf2bf45b1
Return status from wakeword as tuple
...
- Return the data from the wakeword as a tuple.
- Extract post wakeword actions from wait for wakeword method
2020-05-26 12:12:12 +02:00
567e08809a
Refactor _wait_for_wakeword
...
- Add RollingMean data structure
- Add simple CyclicBuffer data structure
- Move out post find actions into separate method
2020-05-26 12:12:12 +02:00
7586de86d9
Fix MutableStream __init__ with mute=True
...
If the stream is restarted while muted this would cause an Exception
leading to an unresponsive voice client due to the read_lock being
called before creation.
2020-04-20 09:26:31 +02:00
c5c0881d94
Add configuration option to set microphone timeout
...
listener->recording_timeout: Maximum length of recording
listener->recording_timeout: Maximum length of silence before stopping
recording
2020-04-17 11:23:25 +02:00
003f2920a3
Enable snowboy to use several wakeword models
...
Snowboy could use several wakeword models, not only the first one.
2020-03-11 08:34:50 +01:00
49c6b298f1
Add path checks
2020-02-27 08:33:11 +01:00
af6b7353e6
Added a "save_path" configuration for saved wake words and utterances.
2020-02-27 08:33:06 +01:00
75c796a29d
add context to all listener messages
2020-02-15 15:35:38 +01:00
d16964c580
bus message targetting
2020-02-15 12:37:25 +01:00
308b0c5111
Fix incorrect porcupine shutdown and Exception
...
Flagged by flake8 after merge
2019-12-12 17:51:47 +01:00
3315d98582
Merge pull request #2139 from alumae/feature/issue-1491
...
Add support for the Porcupine wakeword engine.
2019-12-12 16:59:12 +01:00
f9585e47ba
Clean up of create_wake_word_recognizer
...
- Proper docstring
- " -> '
2019-12-07 20:22:20 +01:00
2c2c933832
Make sure hotword entry has phonemes and threshold
...
If hotword entry is missing either add from old listener location
2019-12-07 20:18:47 +01:00
026f769028
Fix using of hotword structure for custom wakeword
...
Fixes overriding settings from listener config
2019-11-26 09:07:15 +01:00
a1ff3fb355
Upgrade from pycodestyle to flake8
2019-11-19 23:07:52 +01:00
Åke Forslund
Bart Ribbers
Kris Gesling
mathmauney
dalgwen
jarbasal
katridi
Kris Gesling
Aditya Mehra
ludwhe
Christian Clauss