Hass/xiaozhi-esp32-server
1
0
Fork 0
mirror of https://github.com/xinnan-tech/xiaozhi-esp32-server.git synced 2026-07-30 19:13:55 +08:00
Code Issues Packages Projects Releases Wiki Activity

修复test_page.html没有hello消息，导致缺失audio_format的bug (#1151)

Browse Source
This commit is contained in:
hrz
2025-05-08 15:08:11 +08:00
committed by GitHub
parent af3d00662e
commit 98bfe863fe
9 changed files with 116 additions and 64 deletions
Download Patch File Download Diff File Expand all files Collapse all files
main/xiaozhi-server/core/providers/asr/tencent.py
+46 -28
View File
@@ -17,12 +17,14 @@ from config.logger import setup_logging
TAG = __name__
logger = setup_logging()
class ASRProvider(ASRProviderBase):
API_URL = "https://asr.tencentcloudapi.com"
API_VERSION = "2019-06-14"
FORMAT = "pcm" # 支持的音频格式：pcm, wav, mp3
def __init__(self, config: dict, delete_audio_file: bool = True):
super().__init__()
self.secret_id = config.get("secret_id")
self.secret_key = config.get("secret_key")
self.output_dir = config.get("output_dir")
@@ -45,7 +47,9 @@ class ASRProvider(ASRProviderBase):
return file_path
async def speech_to_text(self, opus_data: List[bytes], session_id: str) -> Tuple[Optional[str], Optional[str]]:
async def speech_to_text(
self, opus_data: List[bytes], session_id: str
) -> Tuple[Optional[str], Optional[str]]:
"""将语音数据转换为文本"""
if not opus_data:
logger.bind(tag=TAG).warn("音频数据为空！")
@@ -72,7 +76,7 @@ class ASRProvider(ASRProviderBase):
self.save_audio_to_file(pcm_data, session_id)
# 将音频数据转换为Base64编码
base64_audio = base64.b64encode(combined_pcm_data).decode('utf-8')
base64_audio = base64.b64encode(combined_pcm_data).decode("utf-8")
# 构建请求体
request_body = self._build_request_body(base64_audio)
@@ -85,7 +89,9 @@ class ASRProvider(ASRProviderBase):
result = self._send_request(request_body, timestamp, authorization)
if result:
logger.bind(tag=TAG).debug(f"腾讯云语音识别耗时: {time.time() - start_time:.3f}s | 结果: {result}")
logger.bind(tag=TAG).debug(
f"腾讯云语音识别耗时: {time.time() - start_time:.3f}s | 结果: {result}"
)
return result, file_path
@@ -102,7 +108,7 @@ class ASRProvider(ASRProviderBase):
"SourceType": 1, # 音频数据来源为语音文件
"VoiceFormat": self.FORMAT, # 音频格式
"Data": base64_audio, # Base64编码的音频数据
"DataLen": len(base64_audio) # 数据长度
"DataLen": len(base64_audio), # 数据长度
}
return json.dumps(request_map)
@@ -135,9 +141,11 @@ class ASRProvider(ASRProviderBase):
action = "SentenceRecognition" # 接口名称
# 构建规范头部信息，注意顺序和格式
canonical_headers = f"content-type:{content_type.lower()}\n" + \
f"host:{host.lower()}\n" + \
f"x-tc-action:{action.lower()}\n"
canonical_headers = (
f"content-type:{content_type.lower()}\n"
+ f"host:{host.lower()}\n"
+ f"x-tc-action:{action.lower()}\n"
)
signed_headers = "content-type;host;x-tc-action"
@@ -145,21 +153,25 @@ class ASRProvider(ASRProviderBase):
payload_hash = self._sha256_hex(request_body)
# 构建规范请求字符串
canonical_request = f"{http_request_method}\n" + \
f"{canonical_uri}\n" + \
f"{canonical_query_string}\n" + \
f"{canonical_headers}\n" + \
f"{signed_headers}\n" + \
f"{payload_hash}"
canonical_request = (
f"{http_request_method}\n"
+ f"{canonical_uri}\n"
+ f"{canonical_query_string}\n"
+ f"{canonical_headers}\n"
+ f"{signed_headers}\n"
+ f"{payload_hash}"
)
# 计算规范请求的哈希值
hashed_canonical_request = self._sha256_hex(canonical_request)
# 构建待签名字符串
string_to_sign = f"{algorithm}\n" + \
f"{timestamp}\n" + \
f"{credential_scope}\n" + \
f"{hashed_canonical_request}"
string_to_sign = (
f"{algorithm}\n"
+ f"{timestamp}\n"
+ f"{credential_scope}\n"
+ f"{hashed_canonical_request}"
)
# 计算签名密钥
secret_date = self._hmac_sha256(f"TC3{self.secret_key}", date)
@@ -167,13 +179,17 @@ class ASRProvider(ASRProviderBase):
secret_signing = self._hmac_sha256(secret_service, "tc3_request")
# 计算签名
signature = self._bytes_to_hex(self._hmac_sha256(secret_signing, string_to_sign))
signature = self._bytes_to_hex(
self._hmac_sha256(secret_signing, string_to_sign)
)
# 构建授权头
authorization = f"{algorithm} " + \
f"Credential={self.secret_id}/{credential_scope}, " + \
f"SignedHeaders={signed_headers}, " + \
f"Signature={signature}"
authorization = (
f"{algorithm} "
+ f"Credential={self.secret_id}/{credential_scope}, "
+ f"SignedHeaders={signed_headers}, "
+ f"Signature={signature}"
)
return timestamp, authorization
@@ -181,7 +197,9 @@ class ASRProvider(ASRProviderBase):
logger.bind(tag=TAG).error(f"生成认证头失败: {e}", exc_info=True)
raise RuntimeError(f"生成认证头失败: {e}")
def _send_request(self, request_body: str, timestamp: str, authorization: str) -> Optional[str]:
def _send_request(
self, request_body: str, timestamp: str, authorization: str
) -> Optional[str]:
"""发送请求到腾讯云API"""
headers = {
"Content-Type": "application/json; charset=utf-8",
@@ -190,7 +208,7 @@ class ASRProvider(ASRProviderBase):
"X-TC-Action": "SentenceRecognition",
"X-TC-Version": self.API_VERSION,
"X-TC-Timestamp": timestamp,
"X-TC-Region": "ap-shanghai"
"X-TC-Region": "ap-shanghai",
}
try:
@@ -221,16 +239,16 @@ class ASRProvider(ASRProviderBase):
def _sha256_hex(self, data: str) -> str:
"""计算字符串的SHA256哈希值"""
digest = hashlib.sha256(data.encode('utf-8')).digest()
digest = hashlib.sha256(data.encode("utf-8")).digest()
return self._bytes_to_hex(digest)
def _hmac_sha256(self, key, data: str) -> bytes:
"""计算HMAC-SHA256"""
if isinstance(key, str):
key = key.encode('utf-8')
key = key.encode("utf-8")
return hmac.new(key, data.encode('utf-8'), hashlib.sha256).digest()
return hmac.new(key, data.encode("utf-8"), hashlib.sha256).digest()
def _bytes_to_hex(self, bytes_data: bytes) -> str:
"""字节数组转十六进制字符串"""
return ''.join(f"{b:02x}" for b in bytes_data)
return "".join(f"{b:02x}" for b in bytes_data)