selfhostyourtech/reflector
2
0
Fork 0
mirror of https://github.com/Monadical-SAS/reflector.git synced 2025-12-20 12:19:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

server: remove reference to banana.dev

Browse Source
This commit is contained in:
Mathieu Virbel
2023-11-15 15:17:21 +01:00
committed by Mathieu Virbel
parent 84e425bd3b
commit 4dbcb80228
4 changed files with 6 additions and 169 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
server/env.example
View File

@@ -51,17 +51,6 @@
#TRANSLATE_URL=https://xxxxx--reflector-translator-web.modal.run #TRANSLATE_URL=https://xxxxx--reflector-translator-web.modal.run
#TRANSCRIPT_MODAL_API_KEY=xxxxx #TRANSCRIPT_MODAL_API_KEY=xxxxx
## Using serverless banana.dev (require reflector-gpu-banana deployed)
## XXX this service is buggy do not use at the moment
## XXX it also require the audio to be saved to S3
#TRANSCRIPT_BACKEND=banana
#TRANSCRIPT_URL=https://reflector-gpu-banana-xxxxx.run.banana.dev
#TRANSCRIPT_BANANA_API_KEY=xxx
#TRANSCRIPT_BANANA_MODEL_KEY=xxx
#TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID=xxx
#TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY=xxx
#TRANSCRIPT_STORAGE_AWS_BUCKET_NAME="reflector-bucket/chunks"
## ======================================================= ## =======================================================
## LLM backend ## LLM backend
## ##
@@ -78,13 +67,6 @@
#LLM_URL=https://xxxxxx--reflector-llm-web.modal.run #LLM_URL=https://xxxxxx--reflector-llm-web.modal.run
#LLM_MODAL_API_KEY=xxx #LLM_MODAL_API_KEY=xxx
## Using serverless banana.dev (require reflector-gpu-banana deployed)
## XXX this service is buggy do not use at the moment
#LLM_BACKEND=banana
#LLM_URL=https://reflector-gpu-banana-xxxxx.run.banana.dev
#LLM_BANANA_API_KEY=xxxxx
#LLM_BANANA_MODEL_KEY=xxxxx
## Using OpenAI ## Using OpenAI
#LLM_BACKEND=openai #LLM_BACKEND=openai
#LLM_OPENAI_KEY=xxx #LLM_OPENAI_KEY=xxx

54
server/reflector/llm/llm_banana.py
View File

@@ -1,54 +0,0 @@
import httpx
from reflector.llm.base import LLM
from reflector.settings import settings
from reflector.utils.retry import retry
class BananaLLM(LLM):
def __init__(self):
super().__init__()
self.timeout = settings.LLM_TIMEOUT
self.headers = {
"X-Banana-API-Key": settings.LLM_BANANA_API_KEY,
"X-Banana-Model-Key": settings.LLM_BANANA_MODEL_KEY,
}
async def _generate(
self, prompt: str, gen_schema: dict | None, gen_cfg: dict | None, **kwargs
):
json_payload = {"prompt": prompt}
if gen_schema:
json_payload["gen_schema"] = gen_schema
if gen_cfg:
json_payload["gen_cfg"] = gen_cfg
async with httpx.AsyncClient() as client:
response = await retry(client.post)(
settings.LLM_URL,
headers=self.headers,
json=json_payload,
timeout=self.timeout,
retry_timeout=300, # as per their sdk
)
response.raise_for_status()
text = response.json()["text"]
return text
LLM.register("banana", BananaLLM)
if __name__ == "__main__":
from reflector.logger import logger
async def main():
llm = BananaLLM()
prompt = llm.create_prompt(
instruct="Complete the following task",
text="Tell me a joke about programming.",
)
result = await llm.generate(prompt=prompt, logger=logger)
print(result)
import asyncio
asyncio.run(main())

86
server/reflector/processors/audio_transcript_banana.py
View File

@@ -1,86 +0,0 @@
"""
Implementation using the GPU service from banana.
API will be a POST request to TRANSCRIPT_URL:
```json
{
"audio_url": "https://...",
"audio_ext": "wav",
"timestamp": 123.456
"language": "en"
}
```
"""
from pathlib import Path
import httpx
from reflector.processors.audio_transcript import AudioTranscriptProcessor
from reflector.processors.audio_transcript_auto import AudioTranscriptAutoProcessor
from reflector.processors.types import AudioFile, Transcript, Word
from reflector.settings import settings
from reflector.storage import Storage
from reflector.utils.retry import retry
class AudioTranscriptBananaProcessor(AudioTranscriptProcessor):
def __init__(self, banana_api_key: str, banana_model_key: str):
super().__init__()
self.transcript_url = settings.TRANSCRIPT_URL
self.timeout = settings.TRANSCRIPT_TIMEOUT
self.storage = Storage.get_instance(
settings.TRANSCRIPT_STORAGE_BACKEND, "TRANSCRIPT_STORAGE_"
)
self.headers = {
"X-Banana-API-Key": banana_api_key,
"X-Banana-Model-Key": banana_model_key,
}
async def _transcript(self, data: AudioFile):
async with httpx.AsyncClient() as client:
print(f"Uploading audio {data.path.name} to S3")
url = await self._upload_file(data.path)
print(f"Try to transcribe audio {data.path.name}")
request_data = {
"audio_url": url,
"audio_ext": data.path.suffix[1:],
"timestamp": float(round(data.timestamp, 2)),
}
response = await retry(client.post)(
self.transcript_url,
json=request_data,
headers=self.headers,
timeout=self.timeout,
)
print(f"Transcript response: {response.status_code} {response.content}")
response.raise_for_status()
result = response.json()
transcript = Transcript(
text=result["text"],
words=[
Word(text=word["text"], start=word["start"], end=word["end"])
for word in result["words"]
],
)
# remove audio file from S3
await self._delete_file(data.path)
return transcript
@retry
async def _upload_file(self, path: Path) -> str:
upload_result = await self.storage.put_file(path.name, open(path, "rb"))
return upload_result.url
@retry
async def _delete_file(self, path: Path):
await self.storage.delete_file(path.name)
return True
AudioTranscriptAutoProcessor.register("banana", AudioTranscriptBananaProcessor)

17
server/reflector/settings.py
View File

@@ -41,7 +41,7 @@ class Settings(BaseSettings):
AUDIO_BUFFER_SIZE: int = 256 * 960 AUDIO_BUFFER_SIZE: int = 256 * 960
# Audio Transcription # Audio Transcription
# backends: whisper, banana, modal # backends: whisper, modal
TRANSCRIPT_BACKEND: str = "whisper" TRANSCRIPT_BACKEND: str = "whisper"
TRANSCRIPT_URL: str | None = None TRANSCRIPT_URL: str | None = None
TRANSCRIPT_TIMEOUT: int = 90 TRANSCRIPT_TIMEOUT: int = 90
@@ -50,10 +50,6 @@ class Settings(BaseSettings):
TRANSLATE_URL: str | None = None TRANSLATE_URL: str | None = None
TRANSLATE_TIMEOUT: int = 90 TRANSLATE_TIMEOUT: int = 90
# Audio transcription banana.dev configuration
TRANSCRIPT_BANANA_API_KEY: str | None = None
TRANSCRIPT_BANANA_MODEL_KEY: str | None = None
# Audio transcription modal.com configuration # Audio transcription modal.com configuration
TRANSCRIPT_MODAL_API_KEY: str | None = None TRANSCRIPT_MODAL_API_KEY: str | None = None
@@ -61,13 +57,16 @@ class Settings(BaseSettings):
TRANSCRIPT_STORAGE_BACKEND: str = "aws" TRANSCRIPT_STORAGE_BACKEND: str = "aws"
# Storage configuration for AWS # Storage configuration for AWS
TRANSCRIPT_STORAGE_AWS_BUCKET_NAME: str = "reflector-bucket/chunks" TRANSCRIPT_STORAGE_AWS_BUCKET_NAME: str = "reflector-bucket"
TRANSCRIPT_STORAGE_AWS_REGION: str = "us-east-1" TRANSCRIPT_STORAGE_AWS_REGION: str = "us-east-1"
TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID: str | None = None TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID: str | None = None
TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY: str | None = None TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY: str | None = None
# Transcript MP3 storage
TRANSCRIPT_MP3_STORAGE_BACKEND: str = "aws"
# LLM # LLM
# available backend: openai, banana, modal, oobabooga # available backend: openai, modal, oobabooga
LLM_BACKEND: str = "oobabooga" LLM_BACKEND: str = "oobabooga"
# LLM common configuration # LLM common configuration
@@ -82,10 +81,6 @@ class Settings(BaseSettings):
LLM_TEMPERATURE: float = 0.7 LLM_TEMPERATURE: float = 0.7
ZEPHYR_LLM_URL: str | None = None ZEPHYR_LLM_URL: str | None = None
# LLM Banana configuration
LLM_BANANA_API_KEY: str | None = None
LLM_BANANA_MODEL_KEY: str | None = None
# LLM Modal configuration # LLM Modal configuration
LLM_MODAL_API_KEY: str | None = None LLM_MODAL_API_KEY: str | None = None