Page MenuHomePhabricator
Files F4880334

No OneTemporary
Actions

View Options

diff --git a/AIMeiSheng/docker_demo/common.py b/AIMeiSheng/docker_demo/common.py
index 3ede6a6..ca42a27 100644
--- a/AIMeiSheng/docker_demo/common.py
+++ b/AIMeiSheng/docker_demo/common.py
@@ -1,101 +1,105 @@
import os
import time
# import logging
import urllib, urllib.request
# 测试/正式环境
gs_prod = True
gs_tmp_dir = "/tmp/ai_meisheng_tmp"
gs_model_dir = "/tmp/ai_meisheng_models"
gs_resource_cache_dir = "/tmp/ai_meisheng_resource_cache"
gs_svc_model_path = os.path.join(gs_model_dir,
"weights/xusong_v2_org_version_alldata_embed1_enzx_diff_fi_e15_s244110.pth")
gs_embed_model_path = os.path.join(gs_model_dir, "RawNet3/models/weights/model.pt")
gs_hubert_model_path = os.path.join(gs_model_dir, "hubert.pt")
gs_rmvpe_model_path = os.path.join(gs_model_dir, "rmvpe.pt")
# errcode
gs_err_code_success = 0
gs_err_code_download_vocal = 100
gs_err_code_download_svc_url = 101
gs_err_code_svc_process = 102
gs_err_code_transcode = 103
gs_err_code_volume_adjust = 104
gs_err_code_upload = 105
gs_err_code_params = 106
gs_err_code_pending = 107
gs_err_code_target_silence = 108
gs_err_code_too_many_connections = 429
gs_redis_conf = {
"host": "av-credis.starmaker.co",
"port": 6379,
"pwd": "lKoWEhz%jxTO",
}
gs_server_redis_conf = {
"producer": "test_ai_meisheng_producer", # 输入的队列
"ai_meisheng_key_prefix": "test_ai_meisheng_key_", # 存储结果情况
}
if gs_prod:
gs_server_redis_conf = {
"producer": "ai_meisheng_producer", # 输入的队列
"ai_meisheng_key_prefix": "ai_meisheng_key_", # 存储结果情况
}
def download2disk(url, dst_path):
- urllib.request.urlretrieve(url, dst_path)
- return os.path.exists(dst_path)
+ try:
+ urllib.request.urlretrieve(url, dst_path)
+ return os.path.exists(dst_path)
+ except Exception as ex:
+ print(f"download url={url} error", ex)
+ return False
def exec_cmd(cmd):
# gs_logger.info(cmd)
print(cmd)
ret = os.system(cmd)
if ret != 0:
return False
return True
def exec_cmd_and_result(cmd):
r = os.popen(cmd)
text = r.read()
r.close()
return text
def upload_file2cos(key, file_path, region='ap-singapore', bucket_name='av-audit-sync-sg-1256122840'):
"""
将文件上传到cos
:param key: 桶上的具体地址
:param file_path: 本地文件地址
:param region: 区域
:param bucket_name: 桶地址
:return:
"""
gs_coscmd = "coscmd"
gs_coscmd_conf = "~/.cos.conf"
cmd = "{} -c {} -r {} -b {} upload {} {}".format(gs_coscmd, gs_coscmd_conf, region, bucket_name, file_path, key)
if exec_cmd(cmd):
cmd = "{} -c {} -r {} -b {} info {}".format(gs_coscmd, gs_coscmd_conf, region, bucket_name, key) \
+ "| grep Content-Length |awk \'{print $2}\'"
res_str = exec_cmd_and_result(cmd)
# logging.info("{},res={}".format(key, res_str))
size = float(res_str)
if size > 0:
return True
return False
return False
def check_input(input_data):
key_list = ["record_song_url", "target_url", "start", "end", "vocal_loudness", "female_recording_url",
"male_recording_url"]
for key in key_list:
if key not in input_data.keys():
return False
return True
diff --git a/AIMeiSheng/docker_demo/redis_helper.py b/AIMeiSheng/docker_demo/redis_helper.py
index 2cac13a..07c4676 100644
--- a/AIMeiSheng/docker_demo/redis_helper.py
+++ b/AIMeiSheng/docker_demo/redis_helper.py
@@ -1,45 +1,46 @@
# -*- coding: UTF-8 -*-
import time
import redis
class RedisHelper:
def __init__(self, redis_conf, esp_sec=60):
self.redis_conf = redis_conf
self.last_tm = time.time()
self.esp_sec = esp_sec
self.redis_client = redis.StrictRedis(host=self.redis_conf["host"],
port=int(self.redis_conf["port"]), password=self.redis_conf["pwd"])
redis_client = redis.StrictRedis(host="av-credis.starmaker.co", port=6379, password="stamaker_av20240709!@#")
+
def get_client(self):
if time.time() - self.last_tm > self.esp_sec:
self.redis_client = redis.StrictRedis(host=self.redis_conf["host"],
port=int(self.redis_conf["port"]), password=self.redis_conf["pwd"])
self.last_tm = time.time()
return self.redis_client
def get(self, key):
return self.get_client().get(key)
def set(self, key, value):
return self.get_client().set(key, value)
def exists(self, key):
return self.get_client().exists(key)
def expire(self, key, expire_time_sec):
return self.get_client().expire(key, expire_time_sec)
def del_key(self, key):
return self.get_client().delete(key)
def lpush(self, key, value):
return self.get_client().lpush(key, value)
def rpop(self, key):
return self.get_client().rpop(key)
def llen(self, key):
return self.get_client().llen(key)
diff --git a/AIMeiSheng/docker_demo/svc_online.py b/AIMeiSheng/docker_demo/svc_online.py
index 76a1fe9..db0f255 100644
--- a/AIMeiSheng/docker_demo/svc_online.py
+++ b/AIMeiSheng/docker_demo/svc_online.py
@@ -1,181 +1,188 @@
# -*- coding: UTF-8 -*-
"""
SVC的核心处理逻辑
"""
import os
import time
import socket
import shutil
import hashlib
from AIMeiSheng.meisheng_svc_final import load_model, process_svc_online
from AIMeiSheng.meisheng_env_preparex import meisheng_env_prepare
from AIMeiSheng.voice_classification.online.voice_class_online_fang import VoiceClass, download_volume_balanced
from AIMeiSheng.docker_demo.common import *
import logging
hostname = socket.gethostname()
log_file_name = f"{os.path.dirname(os.path.abspath(__file__))}/av_meisheng_{hostname}.log"
# 设置logger
svc_offline_logger = logging.getLogger("svc_offline")
file_handler = logging.FileHandler(log_file_name)
file_handler.setLevel(logging.INFO)
formatter = logging.Formatter('%(asctime)s %(levelname)s %(message)s', datefmt='%Y-%m-%d %I:%M:%S')
file_handler.setFormatter(formatter)
if gs_prod:
svc_offline_logger.addHandler(file_handler)
if os.path.exists(gs_tmp_dir):
shutil.rmtree(gs_tmp_dir)
os.makedirs(gs_model_dir, exist_ok=True)
os.makedirs(gs_resource_cache_dir, exist_ok=True)
# 预设参数
gs_gender_models_url = "https://av-audit-sync-sg-1256122840.cos.ap-singapore.myqcloud.com/hub/voice_classification/models.zip"
gs_volume_bin_url = "https://av-audit-sync-sg-1256122840.cos.ap-singapore.myqcloud.com/dataset/AIMeiSheng/ebur128_tool"
class GSWorkerAttr:
def __init__(self, input_data):
# 取出输入资源
vocal_url = input_data["record_song_url"]
target_url = input_data["target_url"]
start = input_data["start"] # 单位是ms
end = input_data["end"] # 单位是ms
vocal_loudness = input_data["vocal_loudness"]
female_recording_url = input_data["female_recording_url"]
male_recording_url = input_data["male_recording_url"]
self.distinct_id = hashlib.md5(vocal_url.encode()).hexdigest()
self.tmp_dir = os.path.join(gs_tmp_dir, self.distinct_id)
if os.path.exists(self.tmp_dir):
shutil.rmtree(self.tmp_dir)
os.makedirs(self.tmp_dir)
self.vocal_url = vocal_url
self.target_url = target_url
ext = vocal_url.split(".")[-1]
self.vocal_path = os.path.join(self.tmp_dir, self.distinct_id + f"_in.{ext}")
self.target_wav_path = os.path.join(self.tmp_dir, self.distinct_id + "_out.wav")
self.target_wav_ad_path = os.path.join(self.tmp_dir, self.distinct_id + "_out_ad.wav")
self.target_path = os.path.join(self.tmp_dir, self.distinct_id + "_out.m4a")
self.female_svc_source_url = female_recording_url
self.male_svc_source_url = male_recording_url
ext = female_recording_url.split(".")[-1]
self.female_svc_source_path = os.path.join(gs_resource_cache_dir,
hashlib.md5(female_recording_url.encode()).hexdigest() + "." + ext)
ext = male_recording_url.split(".")[-1]
self.male_svc_source_path = os.path.join(gs_resource_cache_dir,
hashlib.md5(male_recording_url.encode()).hexdigest() + "." + ext)
self.st_tm = start
self.ed_tm = end
self.target_loudness = vocal_loudness
def log_info_name(self):
return f"d_id={self.distinct_id}, vocal_url={self.vocal_url}"
def rm_cache(self):
if os.path.exists(self.tmp_dir):
shutil.rmtree(self.tmp_dir)
def init_gender_model():
"""
下载模型
:return:
"""
dst_model_dir = os.path.join(gs_model_dir, "voice_classification")
if not os.path.exists(dst_model_dir):
dst_zip_path = os.path.join(gs_model_dir, "models.zip")
if not download2disk(gs_gender_models_url, dst_zip_path):
svc_offline_logger.fatal(f"download gender_model err={gs_gender_models_url}")
cmd = f"cd {gs_model_dir}; unzip {dst_zip_path}; mv models voice_classification; rm -f {dst_zip_path}"
os.system(cmd)
if not os.path.exists(dst_model_dir):
svc_offline_logger.fatal(f"unzip {dst_zip_path} err")
music_voice_pure_model = os.path.join(dst_model_dir, "voice_005_rec_v5.pth")
music_voice_no_pure_model = os.path.join(dst_model_dir, "voice_10_v5.pth")
gender_pure_model = os.path.join(dst_model_dir, "gender_8k_ratev5_v6_adam.pth")
gender_no_pure_model = os.path.join(dst_model_dir, "gender_8k_v6_adam.pth")
vc = VoiceClass(music_voice_pure_model, music_voice_no_pure_model, gender_pure_model, gender_no_pure_model)
return vc
def init_svc_model():
meisheng_env_prepare(logging, gs_model_dir)
embed_model, hubert_model = load_model()
return embed_model, hubert_model
def download_volume_adjustment():
"""
下载音量调整工具
:return:
"""
volume_bin_path = os.path.join(gs_model_dir, "ebur128_tool")
if not os.path.exists(volume_bin_path):
if not download2disk(gs_volume_bin_url, volume_bin_path):
svc_offline_logger.fatal(f"download volume_bin err={gs_volume_bin_url}")
os.system(f"chmod +x {volume_bin_path}")
def volume_adjustment(wav_path, target_loudness, out_path):
"""
音量调整
:param wav_path:
:param target_loudness:
:param out_path:
:return:
"""
volume_bin_path = os.path.join(gs_model_dir, "ebur128_tool")
cmd = f"{volume_bin_path} {wav_path} {target_loudness} {out_path}"
os.system(cmd)
class SVCOnline:
def __init__(self):
st = time.time()
self.gender_model = init_gender_model()
self.embed_model, self.hubert_model = init_svc_model()
download_volume_adjustment()
download_volume_balanced()
svc_offline_logger.info(f"svc init finished, sp = {time.time() - st}")
def gender_process(self, worker_attr):
st = time.time()
gender, female_rate, is_pure = self.gender_model.process(worker_attr.vocal_path)
svc_offline_logger.info(
f"{worker_attr.vocal_url}, gender={gender}, female_rate={female_rate}, is_pure={is_pure}, "
f"gender_process sp = {time.time() - st}")
if gender == 0:
gender = 'female'
elif gender == 1:
gender = 'male'
elif female_rate > 0.5:
gender = 'female'
else:
gender = 'male'
svc_offline_logger.info(f"{worker_attr.vocal_url}, modified gender={gender}")
- return gender
+
+ err = gs_err_code_success
+ if female_rate == -1:
+ err = gs_err_code_target_silence
+ return gender, err
def process(self, worker_attr):
- gender = self.gender_process(worker_attr)
+ gender, err = self.gender_process(worker_attr)
+ if err != gs_err_code_success:
+ return gender, err
+
song_path = worker_attr.female_svc_source_path
if gender == "male":
song_path = worker_attr.male_svc_source_path
params = {'gender': gender, 'tst': worker_attr.st_tm, "tnd": worker_attr.ed_tm, 'delay': 0, 'song_path': None}
st = time.time()
err_code = process_svc_online(song_path, worker_attr.vocal_path, worker_attr.target_wav_path, self.embed_model,
- self.hubert_model, params)
+ self.hubert_model, params)
svc_offline_logger.info(f"{worker_attr.vocal_url}, err_code={err_code} process svc sp = {time.time() - st}")
return gender, err_code

File Metadata

Mime Type
text/x-diff
Expires
Sun, Jan 12, 08:34 (1 d, 15 h)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
1347202
Default Alt Text
(13 KB)

Event Timeline