def __init__(self, key):
"""
:param key: 秘钥
"""
Assert.is_true(is_not_empty(key), 'AES 秘钥不能为空')
self.key = key
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 __init__(self, api_key=None, secret_key=None, baidu_cloud=None):
"""
百度云图像识别
:param str api_key: 应用的API Key
:param str secret_key: 应用的Secret Key
:param BaiduCloud baidu_cloud:
"""
self.token = None
if is_not_empty(api_key) and is_not_empty(secret_key):
baidu_cloud = BaiduCloud(api_key, secret_key)
baidu_cloud.init_token()
elif is_not_empty(baidu_cloud) and is_not_empty(baidu_cloud.token):
pass
else:
raise MyError('缺失鉴权 Token 参数.')
self.token = baidu_cloud.token
def _format_convert(cls,
input_file,
output_file,
output_ac,
output_ar,
input_ac=None,
input_ar=None,
input_format=None,
input_acodec=None,
output_format=None,
output_acodec=None):
"""
音频文件格式转换
:param input_file:
:param output_file:
:param output_ac:
:param output_ar:
:param input_ac:
:param input_ar:
:param input_format:
:param input_acodec:
:param output_format:
:param output_acodec:
:return:
"""
output_dir, _, _ = FileUtil.get_path_name_ext(output_file,
is_strict=False)
FileUtil.creat_dirs(output_dir)
if cls.is_pcm(input_file):
# 格式为 pcm 格式，需要音频文件参数
Assert.is_true((is_not_empty(input_ac) and is_not_empty(input_ar)
and is_not_empty(input_format)
and is_not_empty(input_acodec)), 'pcm 缺失文件参数')
_, error = (ffmpeg.input(
input_file,
**cls._file_args(
input_format, input_acodec, input_ac, input_ar)).output(
output_file,
**cls._file_args(output_format, output_acodec, output_ac,
output_ar)).run(capture_stdout=False,
overwrite_output=True))
def _file_args(cls, format=None, acodec=None, ac=None, ar=None):
"""
音频文件参数信息
:param str format: 格式
:param str acodec: 编码器
:param int ac: 声道数
:param int ar: 采样率
:return:
"""
file_dict_args = {}
if is_not_empty(format):
file_dict_args['format'] = format
if is_not_empty(acodec):
file_dict_args['acodec'] = acodec
if is_not_empty(ac):
file_dict_args['ac'] = ac
if is_not_empty(ar):
file_dict_args['ar'] = ar
return file_dict_args
def decorated_function(*args, **kwargs):
args_tuple = args[1:]
if is_not_empty(args_tuple):
return func(*args, **kwargs)
try:
kwargs_dict = current_app.get_object_dict(key)
kwargs_dict.update(kwargs)
return func(*args, **kwargs_dict)
except RuntimeError:
# 未创建 flask 上下文，直接返回原参数
return func(*args, **kwargs)
def __init__(self, app_id=None, api_key=None, secret_key=None, baidu_cloud=None, **kwargs):
"""
音频应用
:param app_id:
:param api_key:
:param secret_key:
:param baidu_cloud:
"""
self.app_id = app_id
self.token = None
if is_not_empty(api_key) and is_not_empty(secret_key):
baidu_cloud = BaiduCloud(api_key, secret_key)
baidu_cloud.init_token()
elif is_not_empty(baidu_cloud) and is_not_empty(baidu_cloud.token):
pass
else:
raise MyError('缺失鉴权 Token 参数.')
self.token = baidu_cloud.token
if kwargs:
for key, value in kwargs.items():
setattr(self, key, value)
def __init__(self, app_id=None, api_key=None, sign=None, **kwargs):
"""
讯飞开放平台数字签名
:param str app_id: 讯飞开放平台应用ID
:param str api_key: 讯飞开放平台应用秘钥
:param Sign sign: 讯飞开放平台数字签名对象
"""
if is_not_empty(app_id) and is_not_empty(api_key):
sign = Sign(app_id, api_key)
elif is_not_empty(sign) and is_not_empty(sign.signa):
pass
else:
raise MyError('缺失数字签名参数.')
self.app_id = app_id
self.ts = sign.ts
self.signa = sign.signa
self.task_id = None
if kwargs:
for key, value in kwargs.items():
setattr(self, key, value)
def amount_formatting(s, formats):
"""
金额格式化
:param s:
:param formats:
:return:
"""
try:
if is_not_empty(s):
for format in formats:
s = s.replace(format, '')
return float(s.strip())
return None
except:
return None
def init_token(self):
"""
百度云鉴权接口
:return:
:rtype:Token
"""
res = requests.post(url=self.token_url, 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.server_error,
msg='百度云鉴权请求失败【{0}】'.
format(res.status_code if res is not None else codes.bad)
)
token, _ = Token.TokenSchema().load(res.json())
if is_not_empty(token.error):
raise MyError(
code=codes.server_error,
msg='百度云鉴权失败【{0}】: {1}'.format(token.error, token.error_description)
)
self.token = token.access_token
return token
def digest(input_str, salt, hash_cls, iterations):
"""
对字符串进行散列, 支持md5与sha1算法.
:param str input_str: 待散列字符
:param bytes salt: 盐
:param str hash_cls: 算法实例
:param int iterations: 跌算次数
:return:
"""
from utils.encodes import Unicode
_hash_model = 'hashlib'
hash_instance = create_instance(_hash_model, hash_cls)
if is_not_empty(salt):
hash_instance.update(salt)
hash_instance.update(bytes(input_str, encoding=Unicode.UTF_8.value))
res = hash_instance.digest()
for i in range(1, iterations):
res = create_instance(_hash_model, hash_cls, res).digest()
return res
def decorated_function(*args, **kwargs):
# 根据参数类型实例化序列化实例
cls = schema_cls
if isinstance(cls, type):
cls = cls(*s, **ks)
elif isinstance(cls, basestring):
cls = create_instance(module_name, cls, *s, **ks)
keys, res = func(*args, **kwargs)
if not res:
return res
# 查询转字典
res_add = Dict(dict(zip(keys, res)))
res_dic = res_add.load_nest()
# 字典反序列化
entity, errors = cls.load(res_dic, partial=cls.partial_db())
if is_not_empty(errors):
raise MyError(msg=errors)
return entity
def pdf_to_pic(path,
pic_dir,
format=FileFormat.JPG.value,
loss=True,
gamma=True,
zoom=210,
min_size=1.50,
max_size=15.0):
"""
从pdf中提取图片
:param path: pdf的路径
:param pic_dir: 图片保存的路径
:param format: 图片格式
:param bool loss: 是否压缩
:param bool gamma: 是否 gamma 矫正
:param int zoom: 保存图片分辨率
:param min_size: 文件大小最小阈值
:param max_size: 文件大小最大阈值
:return: {page_num, success_num, fail_num, msg}
"""
Assert.is_true(os.path.isfile(path), '文件不存在, path: {0}'.format(path))
page_num, success_num, fail_num = 0, 0, 0
detail_info = {'images': []}
pdf = None
# 初始化图片压缩
pngquant = PngQuant(min_quality=80,
max_quality=100,
tmp_file=os.path.join(pic_dir, 'quant.tmp.png'))
try:
FileUtil.creat_dirs(pic_dir)
pdf = fitz.Document(path)
page_num = pdf.pageCount
for pg in range(page_num):
pg_zoom = zoom
pm_dict = {'page_code': pg + 1}
try:
page = pdf[pg] # type: Page
trans = fitz.Matrix(pg_zoom / 100.0,
pg_zoom / 100.0).preRotate(0)
pm = page.getPixmap(matrix=trans, alpha=False) # 获得每一页的流对象
page_path = FileUtil.path_join(pic_dir, '{0}.{1}'.format(
(pg + 1), format)) # 图片路径
pm.writeImage(page_path) # 保存图片
file_size = FileUtil.get_file_size(page_path)
if file_size <= min_size:
# 低于最小阀值
pg_zoom = zoom * 2
trans = fitz.Matrix(pg_zoom / 100.0,
pg_zoom / 100.0).preRotate(0)
pm = page.getPixmap(matrix=trans, alpha=False)
pm.writeImage(page_path)
elif file_size >= max_size:
# 大于最大阀值
pg_zoom = zoom * 2 / 3
trans = fitz.Matrix(pg_zoom / 100.0,
pg_zoom / 100.0).preRotate(0)
pm = page.getPixmap(matrix=trans, alpha=False)
pm.writeImage(page_path)
if gamma:
# gamma 矫正
Enhancer().enhance(page_path, False, False, False,
False, True)
if loss:
# 对图片进行近无损压缩
pngquant.quant_image(page_path)
pm_dict['img_path'] = page_path
success_num = success_num + 1
except Exception as e:
page_path = pm_dict.get('img_path', '')
if is_not_empty(page_path):
pm_dict.pop('img_path')
if os.path.isfile(page_path):
# 处理失败，删除失败文件
FileUtil.del_file(page_path)
pm_dict['error_msg'] = repr(e) if repr(e) else 'pdf转图片失败'
fail_num = fail_num + 1
finally:
detail_info.get('images').append(pm_dict)
except Exception as e1:
detail_info['error_msg'] = repr(e1) if repr(e1) else '处理pdf失败'
finally:
if is_not_empty(pdf):
pdf.close()
if success_num + fail_num != page_num:
fail_num = page_num - success_num
return page_num, success_num, fail_num, detail_info
