jyapayne/vocode-python
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

remove pyq goodbye model and rime synthesizer and fix environment loading

Browse source
This commit is contained in:
Ajay Raj 2023-03-28 10:20:36 -07:00
commit 1dc7bc74c3
28 changed files with 143 additions and 285 deletions
Download patch file Download diff file Expand all files Collapse all files

7
vocode/streaming/synthesizer/azure_synthesizer.py
View file

@ -4,7 +4,7 @@ import re
from typing import Any, Optional
from xml.etree import ElementTree
import azure.cognitiveservices.speech as speechsdk
from dotenv import load_dotenv
from vocode import getenv
from vocode.streaming.agent.bot_sentiment_analyser import BotSentiment
from vocode.streaming.models.message import BaseMessage, SSMLMessage
@ -20,7 +20,6 @@ from vocode.streaming.synthesizer.base_synthesizer import (
from vocode.streaming.models.synthesizer import AzureSynthesizerConfig
from vocode.streaming.models.audio_encoding import AudioEncoding
load_dotenv()
NAMESPACES = {
"mstts": "https://www.w3.org/2001/mstts",
@ -59,8 +58,8 @@ class AzureSynthesizer(BaseSynthesizer):
self.synthesizer_config = synthesizer_config
# Instantiates a client
speech_config = speechsdk.SpeechConfig(
subscription=os.environ.get("AZURE_SPEECH_KEY"),
region=os.environ.get("AZURE_SPEECH_REGION"),
subscription=getenv("AZURE_SPEECH_KEY"),
region=getenv("AZURE_SPEECH_REGION"),
)
if self.synthesizer_config.audio_encoding == AudioEncoding.LINEAR16:
if self.synthesizer_config.sampling_rate == 44100:

7
vocode/streaming/synthesizer/eleven_labs_synthesizer.py
View file

@ -1,7 +1,6 @@
from typing import Any, Optional
import os
from dotenv import load_dotenv
import requests
from vocode import getenv
from vocode.streaming.synthesizer.base_synthesizer import (
BaseSynthesizer,
@ -11,9 +10,7 @@ from vocode.streaming.models.synthesizer import ElevenLabsSynthesizerConfig
from vocode.streaming.agent.bot_sentiment_analyser import BotSentiment
from vocode.streaming.models.message import BaseMessage
load_dotenv()
ELEVEN_LABS_API_KEY = os.environ.get("ELEVEN_LABS_API_KEY")
ELEVEN_LABS_BASE_URL = "https://api.elevenlabs.io/v1/"
ADAM_VOICE_ID = "pNInz6obpgDQGcFmaJgB"
OBAMA_VOICE_ID = "vLITIS0SH2an5iQGxw5C"
@ -22,7 +19,7 @@ OBAMA_VOICE_ID = "vLITIS0SH2an5iQGxw5C"
class ElevenLabsSynthesizer(BaseSynthesizer):
def __init__(self, config: ElevenLabsSynthesizerConfig):
super().__init__(config)
self.api_key = config.api_key
self.api_key = getenv("ELEVEN_LABS_API_KEY")
self.voice_id = config.voice_id or ADAM_VOICE_ID
self.words_per_minute = 150

3
vocode/streaming/synthesizer/google_synthesizer.py
View file

@ -2,7 +2,6 @@ import io
import wave
from typing import Any, Optional
from dotenv import load_dotenv
from google.cloud import texttospeech_v1beta1 as tts
from vocode.streaming.agent.bot_sentiment_analyser import BotSentiment
@ -16,8 +15,6 @@ from vocode.streaming.models.synthesizer import GoogleSynthesizerConfig
from vocode.streaming.models.audio_encoding import AudioEncoding
from vocode.streaming.utils import convert_wav
load_dotenv()
class GoogleSynthesizer(BaseSynthesizer):
OFFSET_SECONDS = 0.5

78
vocode/streaming/synthesizer/rime_synthesizer.py
View file

@ -1,78 +0,0 @@
import audioop
import base64
from vocode.streaming.agent.bot_sentiment_analyser import BotSentiment
from vocode.streaming.models.audio_encoding import AudioEncoding
from vocode.streaming.models.message import BaseMessage
from .base_synthesizer import BaseSynthesizer, SynthesisResult, encode_as_wav
from typing import Any, Optional
import os
import io
import wave
from dotenv import load_dotenv
import requests
from ..utils import convert_linear_audio, convert_wav
from ..models.synthesizer import ElevenLabsSynthesizerConfig, RimeSynthesizerConfig
load_dotenv()
RIME_API_KEY = os.getenv("RIME_API_KEY")
RIME_BASE_URL = os.getenv("RIME_BASE_URL")
class RimeSynthesizer(BaseSynthesizer):
def __init__(self, config: RimeSynthesizerConfig):
super().__init__(config)
self.speaker = config.speaker
def create_speech(
self,
message: BaseMessage,
chunk_size: int,
bot_sentiment: Optional[BotSentiment] = None,
) -> SynthesisResult:
url = RIME_BASE_URL
headers = {"Authorization": f"Bearer {RIME_API_KEY}"}
body = {"inputs": {"text": message.text, "speaker": self.speaker}}
response = requests.post(url, headers=headers, json=body)
def chunk_generator(audio, chunk_transform=lambda x: x):
for i in range(0, len(audio), chunk_size):
chunk = audio[i : i + chunk_size]
yield SynthesisResult.ChunkResult(
chunk_transform(chunk), len(chunk) != chunk_size
)
assert response.ok, response.text
data = response.json().get("data")
assert data
audio_file = io.BytesIO(base64.b64decode(data))
if self.synthesizer_config.audio_encoding == AudioEncoding.LINEAR16:
output_bytes = convert_wav(
audio_file,
output_sample_rate=self.synthesizer_config.sampling_rate,
output_encoding=AudioEncoding.LINEAR16,
)
elif self.synthesizer_config.audio_encoding == AudioEncoding.MULAW:
output_bytes = convert_wav(
audio_file,
output_sample_rate=self.synthesizer_config.sampling_rate,
output_encoding=AudioEncoding.MULAW,
)
if self.synthesizer_config.should_encode_as_wav:
output_generator = chunk_generator(
output_bytes, chunk_transform=encode_as_wav
)
else:
output_generator = chunk_generator(output_bytes)
return SynthesisResult(
output_generator,
lambda seconds: self.get_message_cutoff_from_total_response_length(
message, seconds, len(output_bytes)
),
)