def l_trim(cls, haystack, left=''):
if is_empty(haystack):
return haystack
elif is_empty(left):
return haystack.lstrip()
else:
return haystack.lstrip(left)
def get_asr_progress(self):
"""
查询音频转写处理进度
:return:
"""
url = getattr(self, 'asr_progress_url', None)
data = dict(app_id=self.app_id,
signa=self.signa,
ts=self.ts,
task_id=self.task_id)
res = requests.post(url=url, data=data)
res.encoding = encodes.Unicode.UTF_8.value
if res is None or res.status_code != codes.ok:
return AsrInfo(
ok=codes.failed,
err_no=res.status_code if res is not None else codes.bad,
failed='查询音频转写处理进度请求失败')
progress_info, errors = AsrInfo.AsrInfoSchema().load(res.json())
Assert.is_true(is_empty(errors), errors)
if progress_info.ok == codes.success:
progress_details, errors = AsrProgress.AsrProgressSchema().load(
json.loads(progress_info.data))
Assert.is_true(is_empty(errors), errors)
progress_info.data = progress_details
return progress_info
def lexer(self, text, ne_list=[], url=None):
"""
词法分析
:param url:
:param text:
:param ne_list: 过滤命名实体类型
:return:
"""
url = url if url else getattr(self, 'lexer_url', None)
url = url.format(access_token=self.token)
res = requests.post(url=url,
data=json.dumps(obj={'text': text}),
headers=ContentType.JSON_UTF8.value)
res.encoding = encodes.Unicode.UTF_8.value
if res is None or res.status_code != codes.ok:
raise MyError(code=codes.failed, msg='百度词法分析请求失败.')
lexer_res, errors = LexerRes.LexerResSchema().load(
res.json()) # type: LexerRes
Assert.is_true(is_empty(errors), errors)
if is_not_empty(ne_list):
for lexer_item in lexer_res.items[:]:
if lexer_item.ne not in ne_list:
# 根据实体列表过滤
lexer_res.items.remove(lexer_item)
return lexer_res
def probe(cls, file_path, is_ffprobe=True):
"""
解析音频文件参数
:param is_ffprobe: 是否使用 ffprobe 解析文件格式
:param file_path:
:return:
:rtype: Audio
"""
Assert.is_true(os.path.isfile(file_path),
'文件不存在：{0}'.format(file_path))
_, file_name, format_name = FileUtil.get_path_name_ext(file_path)
probe_json = ffmpeg.probe(file_path)
audio_json = probe_json.get('streams')[0]
audio_json.update(probe_json.get('format'))
audio_json.update({
'file_name':
'{name}.{ext}'.format(name=file_name, ext=format_name)
})
audio, errors = AudioSchema().load(audio_json)
Assert.is_true(is_empty(errors), errors)
if not is_ffprobe:
audio.format_name = format_name
return audio
def init_object_dict():
"""
从配置文件初始化字典
:return:
"""
object_dict = {}
file_paths = FileUtil.get_files_by_suffix(base_static, [
'cfg',
])
if is_empty(file_paths):
return object_dict
for file_path in file_paths:
object_dict.update(ConfigUtils(file_path).get_config_dict())
return object_dict
def _asr_prepare(self, **kwargs):
"""
音频转写预处理
:return:
"""
url = getattr(self, 'asr_prepare_url', None)
kwargs.update(dict(app_id=self.app_id, signa=self.signa, ts=self.ts))
res = requests.post(url=url, data=kwargs)
res.encoding = encodes.Unicode.UTF_8.value
if res is None or res.status_code != codes.ok:
return AsrInfo(
ok=codes.failed,
err_no=res.status_code if res is not None else codes.bad,
failed='音频转写预处理请求失败')
prepare_info, errors = AsrInfo.AsrInfoSchema().load(res.json())
Assert.is_true(is_empty(errors), errors)
return prepare_info
def get_asr_result(self):
"""
查询音频转写结果
:return:
"""
url = getattr(self, 'asr_result_url', None)
data = dict(app_id=self.app_id,
signa=self.signa,
ts=self.ts,
task_id=self.task_id)
res = requests.post(url=url, data=data)
res.encoding = encodes.Unicode.UTF_8.value
if res is None or res.status_code != codes.ok:
return AsrInfo(
ok=codes.failed,
err_no=res.status_code if res is not None else codes.bad,
failed='查询音频转写结果请求失败')
result_info, errors = AsrInfo.AsrInfoSchema().load(res.json())
Assert.is_true(is_empty(errors), errors)
return result_info
def hump_to_pep8(cls, s):
"""
驼峰命名转 pep8
小写和大写紧挨一起的地方,加上分隔符,然后全部转小写
:param s:
:return:
"""
if is_empty(s):
return s
res = s[0]
for i in range(1, len(s)):
# s[i] 直接copy 或 先加'_'再copy
if s[i].isupper() and not s[i - 1].isupper(): # 加'_',当前为大写，前一个字母为小写
res += UNDER_LINE
res += s[i]
elif s[i].isupper() and s[i - 1].isupper() and s[i + 1].islower():
# 加'_',当前为大写，前一个字母为小写
res += UNDER_LINE
res += s[i]
else:
res += s[i]
return res.lower()
def _asr_upload(self, upload_file_path):
"""
音频转写文件分片上传
:param upload_file_path: 待上传音频文件路径
:return:
"""
url = getattr(self, 'asr_upload_url', None)
sig = SliceIdGenerator()
with open(upload_file_path, 'rb') as f:
while True:
content = f.read(self.__FILE_PIECE_SIZE)
if not content or len(content) == 0:
break
data = dict(app_id=self.app_id,
signa=self.signa,
ts=self.ts,
task_id=self.task_id,
slice_id=sig.get_next_slice_id())
res = requests.post(url=url,
data=data,
files={'content': content})
res.encoding = encodes.Unicode.UTF_8.value
if res is None or res.status_code != codes.ok:
return AsrInfo(ok=codes.failed,
err_no=res.status_code
if res is not None else codes.bad,
failed='音频转写文件分片上传请求失败')
upload_info, errors = AsrInfo.AsrInfoSchema().load(res.json())
Assert.is_true(is_empty(errors), errors)
# 上传分片失败
Assert.is_true(upload_info.ok == codes.success,
upload_info.err_no,
'分片上传失败：{0}'.format(upload_info.failed))