Xinonria's picture
compress upload data
0ce4f54
import asyncio
from datetime import datetime
import aiohttp
import pickle
import pandas as pd
from utils import normalize_audio_loudness
import os
from dotenv import load_dotenv
from pymongo import MongoClient
from bson import Binary, ObjectId
import zlib
# 尝试直接获取环境变量
BASE_URL = os.environ.get("BASE_URL")
AUDIO_URL = os.environ.get("AUDIO_URL")
MONGO_URI = os.environ.get("MONGO_URI")
DATABASE_NAME = os.environ.get("DATABASE_NAME")
COLLECTION_NAME = os.environ.get("COLLECTION_NAME")
CREATE_COLLECTION = os.environ.get("CREATE_COLLECTION")
# 如果直接获取不到,则从.env文件加载
if BASE_URL is None or AUDIO_URL is None or MONGO_URI is None or DATABASE_NAME is None or COLLECTION_NAME is None or CREATE_COLLECTION is None:
print("从.env文件加载环境变量")
load_dotenv()
BASE_URL = os.getenv("BASE_URL")
AUDIO_URL = os.getenv("AUDIO_URL")
MONGO_URI = os.getenv("MONGO_URI")
DATABASE_NAME = os.getenv("DATABASE_NAME")
COLLECTION_NAME = os.getenv("COLLECTION_NAME")
CREATE_COLLECTION = os.getenv("CREATE_COLLECTION")
client = MongoClient(MONGO_URI)
db = client[DATABASE_NAME]
collection = db[COLLECTION_NAME]
create_collection = db[CREATE_COLLECTION]
async def generate_api(voice_ids, text):
timeout = aiohttp.ClientTimeout(total=10) # 设置10秒的总超时时间
try:
async with aiohttp.ClientSession(timeout=timeout) as session:
async with session.post(BASE_URL+"tts", json={"ids": voice_ids, "text": text}) as response:
if response.status == 200:
# 读取响应内容
audio_data = await response.read()
# print(type(audio_data))
audio_data = normalize_audio_loudness(audio_data)
return audio_data
else:
print(response)
return f"合成失败: {response.status}"
except asyncio.TimeoutError:
return "请求超时,请稍后重试"
except aiohttp.ClientError as e:
return f"网络错误: {str(e)}"
async def get_audio(voice_id):
url = AUDIO_URL + voice_id + ".ogg"
try:
async with aiohttp.ClientSession() as session:
async with session.get(url) as response:
if response.status == 200:
return await response.read()
else:
return f"获取音频失败: {response.status}"
except asyncio.TimeoutError:
return "请求超时,请稍后重试"
except aiohttp.ClientError as e:
return f"网络错误: {str(e)}"
def load_characters_csv(lang):
# 从MongoDB集合中获取数据
cursor = collection.find({"language": lang, "is_public": True})
# 将查询结果转换为列表
data = list(cursor)
# 创建一个空的DataFrame
df = pd.DataFrame(columns=["类别", "id", "名称", "情绪", "头像", "voice_id"])
# 遍历数据并填充DataFrame
for item in data:
df = pd.concat([df, pd.DataFrame({
"类别": [item["category"]],
"id": [str(item["id"])], # 确保id是字符串类型
"名称": [item["name"]],
"情绪": [item["emotion"]],
"头像": [item["avatar"]],
"voice_id": [item["voice_id"]]
})], ignore_index=True)
指定顺序 = {
"zh": ["原神", "崩坏星穹铁道", "绝区零", "鸣潮"],
"en": ["Genshin Impact", "Honkai: Star Rail", "Zenless Zone Zero", "Wuthering Waves"],
"ja": ["原神[げんしん", "崩壊:スターレイル", "ゼンレスゾーンゼロ", "Wuthering Waves"],
"ko": ["원신", "붕괴: 스타레일", "젠레스 존 제로", "Wuthering Waves"]
}
当前语言顺序 = 指定顺序.get(lang, 指定顺序["en"])
其他类别 = sorted(set(df['类别'].unique()) - set(当前语言顺序))
unique_categories = 当前语言顺序 + 其他类别
return df, unique_categories
async def generate_voice(avatar, name, emotion, tags, gender, audio_data, language):
# 将图像数据转换为二进制
avatar_binary = zlib.compress(pickle.dumps(avatar))
# 将音频数据转换为二进制
audio_binary = zlib.compress(pickle.dumps(audio_data))
# 创建声音对象
voice = {
"avatar": Binary(avatar_binary),
"name": name,
"emotion": emotion,
"tags": tags,
"gender": gender,
"audio_data": Binary(audio_binary),
"language": language,
"create_at": datetime.now().isoformat(),
"is_public": False,
"is_reviewed": False
}
result = create_collection.insert_one(voice)
return result.inserted_id