File size: 3,666 Bytes
b6b3361 fb3c56f 07631e1 6aa3952 07631e1 ac79266 6234905 07631e1 4674077 fb3c56f 07631e1 6aa3952 ea8f6b9 6aa3952 07631e1 6234905 07631e1 6aa3952 fb3c56f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 |
import os
import time
from pydantic import BaseModel
from fastapi import FastAPI, HTTPException, Query
from fastapi.responses import FileResponse
from fastapi.middleware.cors import CORSMiddleware
from langchain.chains import LLMChain
from langchain.prompts import PromptTemplate
from TextGen.suno import custom_generate_audio, get_audio_information
from langchain_google_genai import (
ChatGoogleGenerativeAI,
HarmBlockThreshold,
HarmCategory,
)
from TextGen import app
from gradio_client import Client
song_base_api=os.environ["VERCEL_API"]
my_hf_token=os.environ["HF_TOKEN"]
tts_client = Client("https://jofthomas-xtts.hf.space/",hf_token=my_hf_token)
class Generate(BaseModel):
text:str
def generate_text(prompt: str):
if prompt == "":
return {"detail": "Please provide a prompt."}
else:
prompt = PromptTemplate(template=prompt, input_variables=['Prompt'])
# Initialize the LLM
llm = ChatGoogleGenerativeAI(
model="gemini-pro",
safety_settings={
HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE,
},
)
llmchain = LLMChain(
prompt=prompt,
llm=llm
)
llm_response = llmchain.run({"Prompt": prompt})
return Generate(text=llm_response)
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
@app.get("/", tags=["Home"])
def api_home():
return {'detail': 'Welcome to FastAPI TextGen Tutorial!'}
@app.post("/api/generate", summary="Generate text from prompt", tags=["Generate"], response_model=Generate)
def inference(input_prompt: str):
return generate_text(prompt=input_prompt)
@app.get("/generate_wav")
async def generate_wav(text: str, language: str = "en"):
try:
# Use the Gradio client to generate the wav file
result = tts_client.predict(
text, # str in 'Text Prompt' Textbox component
language, # str in 'Language' Dropdown component
"./narator_out.wav", # str (filepath on your computer (or URL) of file) in 'Reference Audio' Audio component
"./narator_out.wav", # str (filepath on your computer (or URL) of file) in 'Use Microphone for Reference' Audio component
False, # bool in 'Use Microphone' Checkbox component
False, # bool in 'Cleanup Reference Voice' Checkbox component
False, # bool in 'Do not use language auto-detect' Checkbox component
True, # bool in 'Agree' Checkbox component
fn_index=1
)
# Get the path of the generated wav file
wav_file_path = result[1]
# Return the generated wav file as a response
return FileResponse(wav_file_path, media_type="audio/wav", filename="output.wav")
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))
@app.get("/generate_song")
async def generate_song(text: str):
try:
data = custom_generate_audio({
"prompt": f"{text}",
"make_instrumental": False,
"wait_audio": False
})
ids = f"{data[0]['id']},{data[1]['id']}"
print(f"ids: {ids}")
for _ in range(60):
data = get_audio_information(ids)
if data[0]["status"] == 'streaming':
print(f"{data[0]['id']} ==> {data[0]['audio_url']}")
print(f"{data[1]['id']} ==> {data[1]['audio_url']}")
break
# sleep 5s
time.sleep(5)
except:
print("Error") |