'commit'
This commit is contained in:
@@ -17,18 +17,16 @@ from Config.Config import XF_APPID, XF_APISECRET, XF_APIKEY
|
||||
|
||||
|
||||
class XunFeiAudioEvaluator_cn:
|
||||
"""讯飞语音评测类"""
|
||||
|
||||
def __init__(self, appid, api_key, api_secret, audio_file, language, txt):
|
||||
"""讯飞语音评测类（中文专用）"""
|
||||
def __init__(self, appid, api_key, api_secret, audio_file, txt):
|
||||
self.appid = appid
|
||||
self.api_key = api_key
|
||||
self.api_secret = api_secret
|
||||
self.audio_file = audio_file
|
||||
self.language = language
|
||||
self.txt = txt.strip() # 确保文本预处理
|
||||
self.host_url = "wss://ise-api.xfyun.cn/v2/open-ise"
|
||||
self.evaluation_results = {} # 统一结果存储变量
|
||||
self.websocket_url = ""
|
||||
self.evaluation_results = {}
|
||||
self.txt = txt
|
||||
|
||||
def generate_auth_url(self):
|
||||
"""生成鉴权URL"""
|
||||
@@ -98,10 +96,10 @@ class XunFeiAudioEvaluator_cn:
|
||||
print(f"WebSocket connection closed,{ws}")
|
||||
|
||||
def on_open(self, ws):
|
||||
"""连接建立处理"""
|
||||
"""连接建立处理（中文专用配置）"""
|
||||
print(f"WebSocket connection opened,{ws},ws连接建立成功...")
|
||||
|
||||
# 发送初始参数
|
||||
# 发送初始参数（中文专用配置）
|
||||
send_dict = {
|
||||
"common": {
|
||||
"app_id": self.appid
|
||||
@@ -110,13 +108,14 @@ class XunFeiAudioEvaluator_cn:
|
||||
"category": "read_sentence",
|
||||
"rstcd": "utf8",
|
||||
"sub": "ise",
|
||||
"group": "pupil",
|
||||
"ent": "cn_vip" if self.language == "chinese" else "en_vip",
|
||||
"group": "pupil", # 根据文档设置为小学组
|
||||
"ent": "cn_vip", # 固定中文引擎
|
||||
"tte": "utf-8",
|
||||
"cmd": "ssb",
|
||||
"auf": "audio/L16;rate=16000",
|
||||
"aue": "lame",
|
||||
"text": '\uFEFF' + "[content]\n" + self.txt
|
||||
"auf": "audio/L16;rate=16000", # 修复参数格式错误
|
||||
"aue": "raw", # 中文推荐使用raw格式
|
||||
"text": '\uFEFF' + "[content]\n" + self.txt, # 确保UTF8 BOM
|
||||
"extra_ability": "multi_dimension" # 启用多维度评分
|
||||
},
|
||||
"data": {
|
||||
"status": 0,
|
||||
@@ -167,21 +166,38 @@ class XunFeiAudioEvaluator_cn:
|
||||
time.sleep(0.04)
|
||||
|
||||
def parse_evaluation_results(self, xml_content):
|
||||
root = ET.fromstring(xml_content)
|
||||
# 兼容中英文不同的根节点
|
||||
read_node = root.find('.//read_sentence') or root.find('.//read_chapter')
|
||||
if read_node is not None:
|
||||
self.results = {
|
||||
'total_score': read_node.attrib.get('total_score', '0'),
|
||||
'accuracy': read_node.attrib.get('accuracy_score', '0'),
|
||||
'fluency': read_node.attrib.get('fluency_score', '0'),
|
||||
'integrity': read_node.attrib.get('integrity_score', '0')
|
||||
}
|
||||
else:
|
||||
# 打印未匹配的XML结构用于调试
|
||||
print(f"No evaluation nodes found in XML: {xml_content}")
|
||||
self.results = {'error': '未找到有效评分节点'}
|
||||
"""解析评测结果XML并提取得分信息"""
|
||||
"""解析中文评测XML结果（仅处理read_sentence节点）"""
|
||||
try:
|
||||
root = ET.fromstring(xml_content)
|
||||
# 中文专用：仅查找read_sentence节点
|
||||
read_sentence = root.find('.//read_sentence')
|
||||
if read_sentence is not None:
|
||||
# 中文评分字段映射
|
||||
self.evaluation_results = {
|
||||
'total_score': float(read_sentence.get('total_score', 0)),
|
||||
'accuracy_score': float(read_sentence.get('accuracy_score', 0)),
|
||||
'fluency_score': float(read_sentence.get('fluency_score', 0)),
|
||||
'integrity_score': float(read_sentence.get('integrity_score', 0)),
|
||||
'tone_score': float(read_sentence.get('tone_score', 0)), # 中文特有调型分
|
||||
'is_rejected': read_sentence.get('is_rejected', 'false') == 'true'
|
||||
}
|
||||
# 提取句子级得分（中文特有）
|
||||
sentences = []
|
||||
for sent in read_sentence.findall('.//sentence'):
|
||||
sentences.append({
|
||||
'content': sent.get('content', ''),
|
||||
'total_score': float(sent.get('total_score', 0)),
|
||||
'fluency_score': float(sent.get('fluency_score', 0))
|
||||
})
|
||||
self.evaluation_results['sentences'] = sentences
|
||||
print(f"中文评测结果解析成功: {self.evaluation_results}")
|
||||
else:
|
||||
print(f"未找到中文评分节点: {xml_content[:200]}") # 打印前200字符调试
|
||||
self.evaluation_results = {'error': '未找到read_sentence节点'}
|
||||
except ET.ParseError as e:
|
||||
print(f"XML解析失败: {e}, 内容: {xml_content[:200]}")
|
||||
self.evaluation_results = {'error': f'解析失败: {str(e)}'}
|
||||
"""解析评测结果XML并提得分信息"""
|
||||
try:
|
||||
print(xml_content)
|
||||
root = ET.fromstring(xml_content)
|
||||
@@ -266,7 +282,7 @@ class XunFeiAudioEvaluator_cn:
|
||||
return descriptions.get(dp_message, f"未知({dp_message})")
|
||||
|
||||
def run_evaluation(self):
|
||||
"""运行评测"""
|
||||
"""运行评测（修复SSL配置）"""
|
||||
start_time = datetime.now()
|
||||
websocket.enableTrace(False)
|
||||
|
||||
@@ -279,7 +295,12 @@ class XunFeiAudioEvaluator_cn:
|
||||
on_open=self.on_open
|
||||
)
|
||||
|
||||
ws_entity.run_forever(sslopt={"cert_reqs": ssl.CERT_NONE})
|
||||
# 修复SSL配置，启用证书验证
|
||||
import certifi
|
||||
ws_entity.run_forever(sslopt={
|
||||
"cert_reqs": ssl.CERT_REQUIRED,
|
||||
"ca_certs": certifi.where()
|
||||
})
|
||||
|
||||
end_time = datetime.now()
|
||||
evaluation_time = end_time - start_time
|
||||
@@ -290,15 +311,15 @@ class XunFeiAudioEvaluator_cn:
|
||||
|
||||
# 使用示例
|
||||
if __name__ == '__main__':
|
||||
# 配置参数
|
||||
# appid = "5b83f8d6"
|
||||
# api_secret = "604fa6cb9c5ab664a0d153fe0ccc6802"
|
||||
# api_key = "5beb887923204000bfcb402046bb05a6"
|
||||
# 中文使用示例
|
||||
appid = XF_APPID
|
||||
api_secret = XF_APISECRET
|
||||
api_key = XF_APIKEY
|
||||
# audio_file = "./1.mp3"
|
||||
audio_file = r'D:\dsWork\dsProject\dsLightRag\static\audio\audio_afc0a96e382c428cba2f00e3f71e4e8f.mp3'
|
||||
audio_file = r'D:\dsWork\dsProject\dsLightRag\static\audio\test_cn.wav' # 中文测试音频
|
||||
txt = "大家好。很高兴认识你们。今天天气很好。我正在使用这个工具练习中文发音。"
|
||||
evaluator = XunFeiAudioEvaluator_cn(appid, api_key, api_secret, audio_file, txt)
|
||||
results, eval_time = evaluator.run_evaluation()
|
||||
print(evaluator.get_evaluation_summary())
|
||||
# 创建评测器实例
|
||||
txt="Hello everyone! Nice to meet you. Today is a beautiful day. I am learning English pronunciation with this tool."
|
||||
evaluator = XunFeiAudioEvaluator_cn(appid, api_key, api_secret, audio_file, "english", txt)
|
||||
|
||||
Reference in New Issue
Block a user