阿里云语音合成对接接口
2018-12-11 本文已影响0人
逍遥_yjz
1.第一种方式-阿里云语音合成，录音文件识别，自然语言分析，rest 调用 python实现
自然语言分析官网文档：
https://help.aliyun.com/document_detail/61378.html?spm=a2c4g.11186623.6.547.9q3U1C
智能语音官网文档：
https://help.aliyun.com/product/30413.html?spm=a2c4g.11186623.3.1.yS0DIK
决定采用阿里云提供的restful api接口，主要利用python urllib库来实现调用，整合里官方给出的demo，总结出调用代码。话不多说，贴出代码：
python环境
python3.5.2
代码结构所示：
结构图
http请求代理工具类
在工程目录中建立utils目录，在目录中建立init.py文件代码如下
# -*- coding:utf-8 -*-
import hashlib
import urllib.request
import hmac
import base64
import datetime
import ssl
import uuid
from urllib.error import HTTPError
class http_proxy:
"""
Http工具类，封装了鉴权
"""
def __init__(self, ak_id, ak_secret):
self.__ak_id = ak_id
self.__ak_secret = ak_secret
def __current_gmt_time(self):
date = datetime.datetime.strftime(datetime.datetime.utcnow(), "%a, %d %b %Y %H:%M:%S GMT")
return date
def __md5_base64(self, strbody):
hash = hashlib.md5()
hash.update(strbody.encode('utf-8'))
print(hash.digest())
return base64.b64encode(hash.digest()).decode('utf-8')
def __sha1_base64(self, str_to_sign, secret):
hmacsha1 = hmac.new(secret.encode('utf-8'), str_to_sign.encode('utf-8'), hashlib.sha1)
return base64.b64encode(hmacsha1.digest()).decode('utf-8')
def send_request(self, url, body):
gmtnow = self.__current_gmt_time()
print(gmtnow)
body_md5 = self.__md5_base64(body)
print(body_md5)
str_to_sign = "POST\napplication/json\n" + body_md5 + "\napplication/json\n" + gmtnow
print(str_to_sign)
signature = self.__sha1_base64(str_to_sign, self.__ak_secret)
print(signature)
auth_header = "Dataplus " + self.__ak_id + ":" + signature
print(auth_header)
ssl._create_default_https_context = ssl._create_unverified_context
headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:23.0) Gecko/20100101 Firefox/23.0'}
req = urllib.request.Request(url, headers=headers)
req.add_header("Accept", "application/json")
req.add_header("Content-Type", "application/json")
req.add_header("Date", gmtnow)
req.add_header("Authorization", auth_header)
data = body.encode('utf-8')
f = urllib.request.urlopen(req, data)
return f.read().decode('utf-8')
def send_get(self, url, task_id):
gmtnow = self.__current_gmt_time()
print(gmtnow)
accept = "application/json"
content_type = "application/json"
str_to_sign = "GET\n" + accept + "\n" + "" + "\n" + content_type + "\n" + gmtnow
print(str_to_sign)
signature = self.__sha1_base64(str_to_sign, self.__ak_secret)
print(signature)
auth_header = "Dataplus " + self.__ak_id + ":" + signature
print(auth_header)
ssl._create_default_https_context = ssl._create_unverified_context
url += "/" + task_id
req = urllib.request.Request(url)
req.add_header("Accept", "application/json")
req.add_header("Content-Type", "application/json")
req.add_header("Date", gmtnow)
req.add_header("Authorization", auth_header)
f = urllib.request.urlopen(req)
return f.read().decode('utf-8')
# try:
# f = urllib.request.urlopen(req)
# except HTTPError as e:
# print(e)
# return None
# else:
# return f.read().decode('utf-8')
def send_requestForNlp(self, path, content):
method = "POST"
content_type = "application/json;chrset=utf-8"
accept = "application/json"
host = "nlp.cn-shanghai.aliyuncs.com"
gmtnow = self.__current_gmt_time()
print(gmtnow)
body_md5 = self.__md5_base64(content)
print(body_md5)
uuidstr = uuid.uuid4().hex
str_to_sign = method+"\n"+accept+"\n" + body_md5 + "\n"+content_type+"\n" + gmtnow+ "\nx-acs-signature-method:HMAC-SHA1\n" + "x-acs-signature-nonce:" + uuidstr + "\n" + path;
print(str_to_sign)
signature = self.__sha1_base64(str_to_sign, self.__ak_secret)
print(signature)
auth_header = "acs " + self.__ak_id + ":" + signature
print(auth_header)
ssl._create_default_https_context = ssl._create_unverified_context
req = urllib.request.Request("http://"+host+path)
req.add_header("Accept",accept)
req.add_header("Content-Type", content_type)
req.add_header("Content-MD5", body_md5)
req.add_header("Date", gmtnow)
req.add_header("Host", host)
req.add_header("x-acs-signature-nonce", uuidstr)
req.add_header("x-acs-signature-method", "HMAC-SHA1")
req.add_header("Authorization", auth_header)
data = content.encode('utf-8')
headers = {
'User-Agent': 'Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.1.6) Gecko/20091201 Firefox/3.5.6'}
f = urllib.request.urlopen(req, data, headers=headers)
return f.read().decode('utf-8')
def sendTtsPost(self, textData, ttsRequestParam,fileRootPath):
url = 'http://nlsapi.aliyun.com/speak?'#语音合成接口
#拼接
paramArray = []
for key in ttsRequestParam:
paramArray.append(key+"="+ttsRequestParam[key])
url+=url+'&'.join(paramArray)
method = "POST"
content_type = "text/plain"
accept = "audio/" + ttsRequestParam['encode_type'] + ",application/json"
gmtnow = self.__current_gmt_time()
body_md5 = self.__md5_base64(textData)
print(body_md5)
str_to_sign = method + "\n" + accept + "\n" + body_md5 + "\n" + content_type + "\n" + gmtnow
print(str_to_sign)
signature = self.__sha1_base64(str_to_sign, self.__ak_secret)
print(signature)
auth_header = "Dataplus " + self.__ak_id + ":" + signature
print(auth_header)
ssl._create_default_https_context = ssl._create_unverified_context
req = urllib.request.Request(url)
req.add_header("accept", accept)
req.add_header("content-Type", content_type)
req.add_header("date", gmtnow)
req.add_header("Authorization", auth_header)
req.add_header("Content-Length", len(textData))
data = textData.encode('utf-8')
f = urllib.request.urlopen(req, data)
if f.status ==200:
file = 'D:\PythonFiles\\2018second\\testTTS\demo\\' + uuid.uuid4().hex + ".wav"
content = f.read()
with open(file, 'wb') as f:
f.write(content)
print("success"+file)
else:
print('失败！')
调用demo
在目录中建立demo.py ，注意和上面的utils目录同级
# -*- coding:utf-8 -*-
import sys
import utils
import json
# app_key 语音数据格式 领域
# nls-service-realtime-8k 8kHz采样率 智能客服服务领域，比如电话客服等
# nls-service-multi-domain 16kHz采样率 汉语通用识别
# nls-service 16kHz采样率 输入法，社交聊天
# nls-service-tv 16kHz采样率 家庭娱乐
# nls-service-shopping 16kHz采样率 电商购物
# nls-realtime-fangyan 16kHz采样率 支持东北、河南、四川等方言
# nls-service-yue-streaming 16kHz采样率 粤语
# nls-service-en 16kHz采样率 英语
ak_id = ""; ##数加管控台获得的accessId
ak_secret = "" ## 数加管控台获得的accessSecret
url = "https://nlsapi.aliyun.com/transcriptions"
# 录音文件提交
def request():
body = {
'app_key': 'nls-service-multi-domain',
'oss_link': 'http://网址/audio/zl4.mp3',
}
bodyStr = json.dumps(body)
httpProxy = utils.http_proxy(ak_id, ak_secret)
result = httpProxy.send_request(url, bodyStr)
return result
# 录音文件识别结果查询
def query(id):
httpProxy = utils.http_proxy(ak_id, ak_secret)
result = httpProxy.send_get(url, id)
return result
# 自然语音分析 分词
def nlpTest():
path = '/nlp/api/wordpos/general'
postBody = {
'text': '为什么世界是这个样子，人们都不诚实，我要努力，获得成功，让别人尊敬',
'lang': 'ZH',
}
bodyStr = json.dumps(postBody)
httpProxy = utils.http_proxy(ak_id, ak_secret)
result = httpProxy.send_requestForNlp(path,bodyStr)
return result
#语音合成
def ttsTest():
text = '阿彪阿彪阿彪阿彪阿彪阿彪阿彪阿彪阿彪阿彪阿彪阿彪，'
fileRootPath='g:audio/'
ttsRequestParam ={
'encode_type':'wav',#合成语音的编码格式，支持pcm/wav/mp3/alaw
'voice_name':'ruoxi',#xiaogang - 男，xiaoyun - 女
'volume':'33',#0～100
'sample_rate':'16000',#抽样频率率 8000／16000
'speech_rate':'0',#语速 -500～500
'pitch_rate':'0',#语调 -500～500
'tts_nus':'1',#0 - 通过参数合成语音，1 - 拼接原始录音
'background_music_id':'1',#播放语音时可选背景音乐，0，1
'background_music_offset':'0',#背景音乐播放偏移时长，毫秒。当启用背景音乐时生效
'background_music_volume':'100'#背景音乐音量，当启用背景音乐时生效，0～100
}
httpProxy = utils.http_proxy(ak_id, ak_secret)
filepath = httpProxy.sendTtsPost(text,ttsRequestParam,fileRootPath)
print(filepath)
if __name__ == '__main__':
#print(request())
#print(query('2324ec1ed63549318b9477f1bf3eaf8a'))
#print(nlpTest())
print(ttsTest())
转发来源于：https://blog.csdn.net/huangmingleiluo/article/details/81301209
2. 第二种方式
定义speechSynthesis.py
# -*- coding:utf-8 -*-
import hashlib
import requests
import hmac
import base64
import datetime
import tempfile
class Alibaba:
def __init__(self):
self.__ak_id = "" # 替换个人用户 id
self.__ak_secret = "" # 替换个人用户 secret
def __current_gmt_time(self):
date = datetime.datetime.strftime(datetime.datetime.utcnow(), "%a, %d %b %Y %H:%M:%S GMT")
return date
def __md5_base64(self, body):
hash = hashlib.md5()
hash.update(body)
str_hex = hash.digest()
return base64.b64encode(str_hex)
def __sha1_base64(self, str_to_sign, secret):
hmacsha1 = hmac.new(secret.encode('utf-8'), str_to_sign.encode('utf-8'), hashlib.sha1)
return base64.b64encode(hmacsha1.digest()).decode('utf-8')
def send_request(self, phrase):
options = {'method': 'POST',
'url': 'http://nlsapi.aliyun.com/speak?encode_type=wav&voice_name=xiaoyun&volume=33&volume=50',
'body': phrase.encode('utf-8')
}
headers = {
'Authorization': '',
'Content-type': 'text/plain',
'Accept': 'audio/mp3,application/json',
'Date': self.__current_gmt_time()
}
bodyMd5 = self.__md5_base64(options['body']).decode('utf-8')
feature = options['method'] + '\n' + headers['Accept'] + '\n' + bodyMd5 + '\n' \
+ headers['Content-type'] + '\n' + headers['Date']
signature = self.__sha1_base64(feature, self.__ak_secret)
headers['Authorization'] = "Dataplus " + self.__ak_id + ":" + signature
req = requests.post(options['url'], data=options['body'], headers=headers)
#print(req.json())
with tempfile.NamedTemporaryFile(suffix='.mp3', delete=False) as f:
f.write(req.content)
tmpfile = f.name
return tmpfile
if __name__ == '__main__':
a = Alibaba()
a.send_request('今天天气怎天天气怎么样?') # 输入要合成语音的文字
3. 第三种
定义successsfulExample.py
涉及到文件的转移和文件的重命名
# -*- coding:utf-8 -*-
import hashlib
import requests
import hmac
import base64
import datetime
import tempfile
import shutil, os
class aliyun:
def __init__(self):
self.__aliyun_id = "" # 个人用户 id
self.__aliyun_secret = "" # 个人用户 secret
def __time(self):
time = datetime.datetime.strftime(datetime.datetime.utcnow(), "%a, %d %b %Y %H:%M:%S GMT")
return time
def __md5_base64(self, body):
hash = hashlib.md5()
hash.update(body)
str_hex = hash.digest()
return base64.b64encode(str_hex)
def __sha1_base64(self, str_to_sign, secret):
hmacsha1 = hmac.new(secret.encode('utf-8'), str_to_sign.encode('utf-8'), hashlib.sha1)
return base64.b64encode(hmacsha1.digest()).decode('utf-8')
def send_request(self, phrase):
options = {'method': 'POST',
'url': 'http://nlsapi.aliyun.com/speak?encode_type=wav&voice_name=ruoxi&volume=35',
'body': phrase.encode('utf-8')}
headers = {'Authorization': '', 'voice_name': 'xiaogang', 'Content-type': 'text/plain', 'Accept': 'audio/mp3,application/json',
'Date': self.__time()}
MD5 = self.__md5_base64(options['body']).decode('utf-8')
feature = options['method'] + '\n' + headers['Accept'] + '\n' + MD5 + '\n' \
+ headers['Content-type'] + '\n' + headers['Date']
signa = self.__sha1_base64(feature, self.__aliyun_secret)
headers['Authorization'] = "Dataplus " + self.__aliyun_id + ":" + signa
request = requests.post(options['url'], data=options['body'], headers=headers)
print(type(request.content))
path = r'D:\PythonFiles\2018second\testTTS\successsfulExample'
with tempfile.NamedTemporaryFile(suffix='.wav', prefix='3333', dir=path, delete=False) as file:
file.write(request.content)
tmpfile = file.name
return tmpfile
if __name__ == '__main__':
a = aliyun()
text = "李先生吗"
file = a.send_request(text)
#time = datetime.datetime.strftime(datetime.datetime.utcnow(), "%a, %d %b %Y-%b-%d %H:%M:%S GMT")
slash = '\\'
time = datetime.datetime.strftime(datetime.datetime.utcnow(), "%Y-%m-%d-%H%M%S")
print(time)
print(type(time))
time1 = "123454.wav"
main_path = r'D:\PythonFiles\2018second\testTTS\successsfulExample'
print(file)
video_name = os.path.join(main_path + slash, str(time) + '.wav')
#shutil.move(file, video_name) # 文件位置
os.rename(file, os.path.join(main_path + slash, time + '.wav'))
