Commit Graph

234 Commits

Author SHA1 Message Date
Adrian Cowham
79c8aa2c4a ruff formatting 2024-10-11 11:35:02 -07:00
Adrian Cowham
522d931950 better interruption handling by moving the processors after the transport output 2024-10-11 10:33:12 -07:00
Adrian Cowham
4d81a2ebfe nuked the code that marks user audio in favor for InputAudioRawFrame. also moving to stereo instead of mono with the human and bot on their own channel. 2024-10-03 14:10:03 -07:00
Adrian Cowham
2d82702e04 merge from main 2024-10-03 09:42:06 -07:00
Aleix Conchillo Flaqué
75d27aeb9f examples(storytelling): update packages 2024-10-02 12:00:00 -07:00
Aleix Conchillo Flaqué
8c01a900cd google: allow using GOOGLE_APPLICATION_CREDENTIALS 2024-10-02 11:56:01 -07:00
Allenmylath
70c3ff31fd Update env.example
elevenlabs is not used in code instead cartesian is used hence changed
2024-10-02 21:59:51 +05:30
Mark Backman
3d642df2b0 Revert aligning voice_id name in TTS service constructor 2024-10-02 11:07:48 -04:00
Mark Backman
28643b453d Update to use LLM, STT, TTS subclasses and remove setter methods 2024-10-01 20:30:27 -04:00
Kwindla Hultman Kramer
92d8b37229 implement vision for openai 2024-09-30 21:49:29 -07:00
Kwindla Hultman Kramer
0d5cb84531 function calling testing and improvements 2024-09-30 20:59:28 -07:00
Kwindla Hultman Kramer
a5c73ec829 handle openai multiple function calls 2024-09-30 16:08:30 -07:00
Kwindla Hultman Kramer
5d63615b1b Merge pull request #528 from pipecat-ai/khk/sentence-splits
TTS sentence aggregation fix
2024-09-30 16:07:21 -07:00
Kwindla Hultman Kramer
ed607d5c4b typo fix 2024-09-30 16:05:32 -07:00
Kwindla Hultman Kramer
37da7e44cd whitespace fix 2024-09-30 16:05:32 -07:00
Aleix Conchillo Flaqué
c8995b82e5 all frame processors are asynchrnous
In this commit we make all frame processors asynchronous, that is, they have an
internal queue and they push frames using a task from that queue.
2024-09-30 15:11:21 -07:00
Kwindla Hultman Kramer
6b7f924af6 tts sentence aggregation fix 2024-09-30 14:33:08 -07:00
Mark Backman
97cba92fa5 Merge pull request #516 from pipecat-ai/mb/google-tts
Add Google TTS
2024-09-30 12:25:16 -04:00
Mark Backman
830d2df671 Add Google TTS 2024-09-27 07:36:20 -04:00
Aleix Conchillo Flaqué
4501dca133 Merge pull request #467 from joachimchauvet/main
Add LiveKit audio transport
2024-09-26 22:58:25 -07:00
Mark Backman
55a70cde8f Merge pull request #514 from pipecat-ai/mb/aws-polly-tts
Add AWS Polly TTS support
2024-09-26 22:20:13 -04:00
Mark Backman
706c00d897 Code review feedback 2024-09-26 22:13:37 -04:00
Mark Backman
a018112a13 Merge pull request #510 from pipecat-ai/mb/deepgram-tts-http
Improve usability of Deepgram TTS: use Deepgram client, remove aiohttp
2024-09-26 13:38:42 -04:00
Mark Backman
6a6ea251ae Add AWS Polly TTS support 2024-09-26 13:35:59 -04:00
Mark Backman
73da8c1910 Improve usability of Deepgram TTS: use Deepgram client, remove aiohttp 2024-09-25 22:43:10 -04:00
joachimchauvet
b6e1d6e6ae format with ruff 2024-09-24 10:21:02 +03:00
joachimchauvet
fa609f1afc adjust output sample rate and create user token 2024-09-24 10:16:54 +03:00
joachimchauvet
2e5b0c1d6b add tenacity dependency 2024-09-24 10:16:54 +03:00
joachimchauvet
a9390d96a1 add LiveKit audio transport 2024-09-24 10:16:54 +03:00
Jin Kim
cf72129852 Merge remote-tracking branch 'upstream/main' 2024-09-24 07:18:22 +09:00
Aleix Conchillo Flaqué
eeb8338dce introduce Ruff formatting 2024-09-23 09:53:37 -07:00
Jin Kim
68cc4186ad Merge remote-tracking branch 'upstream/main' 2024-09-23 16:34:31 +09:00
Mark Backman
55c645c614 Add voice_settings and optimize_streaming_latency to ElevenLabs 2024-09-22 13:58:50 -04:00
Mark Backman
c73111afea Add extra input param to LLMs 2024-09-21 00:01:25 -04:00
Mark Backman
e8f8a49646 Merge pull request #484 from pipecat-ai/mb/llm-input-params
Add input params for OpenAI, Anthropic, Together AI LLMs
2024-09-20 20:35:49 -04:00
Mark Backman
f3fd312b83 Add Together AI interruptible example 2024-09-20 20:21:19 -04:00
Aleix Conchillo Flaqué
14acf05a26 Merge pull request #480 from pipecat-ai/aleix/input-output-frames
introduce input/output audio and image frames
2024-09-20 14:44:37 -07:00
Aleix Conchillo Flaqué
7e39d9ad3d introduce input/output audio and image frames
We now distinguish between input and output audio and image frames. We introduce
`InputAudioRawFrame`, `OutputAudioRawFrame`, `InputImageRawFrame` and
`OutputImageRawFrame` (and other subclasses of those). The input frames usually
come from an input transport and are meant to be processed inside the pipeline
to generate new frames. However, the input frames will not be sent through an
output transport. The output frames can also be processed by any frame processor
in the pipeline and they are allowed to be sent by the output transport.
2024-09-19 23:11:03 -07:00
mattie ruth backman
a4edb3dab1 Cleanup on aisle METRICS. Note: See below, this is a breaking change
1. Fleshed out MetricsFrames and broke it into a proper set of types
2. Add model_name as a property to the AIService so that it can be
   automatically included in metrics and also remove that
   overhead from all the various services themselves

Breaking change!

Because of the types improvements, the MetricsFrame type has
changed. Each frame will have a list of metrics simlilar to before
except each item in the list will only contain one type of metric:
"ttfb", "tokens", "characters", or "processing". Previously these
fields would be in every entry but set to None if they didn't apply.

While this changes internal handling of the MetricsFrame, it does NOT
break the RTVI/daily messaging of metrics. That format remains the same.

Also. Remember to use model_name for accessing a service's current
model and set_model_name for setting it.
2024-09-19 21:30:34 -04:00
mattie ruth backman
50b45ac2da get the test infrastructure running again
disable broken tests for now
2024-09-19 20:58:17 -04:00
Aleix Conchillo Flaqué
4f1b06e6b2 pipeline: renamed ParallelTask to SyncParallelPipeline 2024-09-19 01:32:17 -07:00
Aleix Conchillo Flaqué
62e9a33a70 examples: use CartesiaHttpTTSService to synchronize frames 2024-09-19 01:32:17 -07:00
Aleix Conchillo Flaqué
337f048864 introduce synchronous and asynchronous frame processors
Pipecat has a pipeline-based architecture. The pipeline consists of frame
processors linked to each other. The elements travelling across the pipeline are
called frames.

To have a deterministic behavior the frames travelling through the pipeline
should always be ordered, except system frames which are out-of-band frames. To
achieve that, each frame processor should only output frames from a single task.

There are synchronous and asynchronous frame processors. The synchronous
processors push output frames from the same task that they receive input frames,
and therefore only pushing frames from one task. Asynchrnous frame processors
can have internal tasks to perform things asynchrnously (e.g. receiving data
from a websocket) but they also have a single task where they push frames from.
2024-09-19 01:31:10 -07:00
Jin Kim
c7f814b2dc Merge remote-tracking branch 'upstream/main' 2024-09-18 00:33:29 +09:00
Adrian Cowham
2e02ab740d PR feedback 2024-09-15 20:59:17 -07:00
Aleix Conchillo Flaqué
5acc4928fe examples: add 07d-interruptible-elevenlabs.py 2024-09-13 09:43:18 -07:00
Aleix Conchillo Flaqué
434493b8aa services(elevenlabs): implement word-by-word support through websockets 2024-09-13 09:31:35 -07:00
Aleix Conchillo Flaqué
f08b25dbb2 examples: assistant aggregator should always goes after transport 2024-09-12 00:37:34 -07:00
Adrian Cowham
b4eff2028f Merge branch 'main' into recording 2024-09-10 10:18:57 -07:00
Adrian Cowham
f411bf33fd adding a frame processor with the ability to save a conversation to a buffer and another frame processor to upload audio to Canonical for evaluation and metrics collection. Examples included 2024-09-10 10:15:48 -07:00