def get_output(self, output: AutoOutput = AUTO) -> Optional[Class]:
if Auto.is_defined(output) and not isinstance(output, LoggingLevel):
return output
if hasattr(self, 'get_logger'):
logger = self.get_logger()
if Auto.is_defined(logger):
return logger
elif Auto.is_auto(output):
return print
def get_err_msg(cls,
scale: Union[str, DynamicEnum] = '{}',
available_scales=AUTO):
list_available_scales = Auto.delayed_acquire(available_scales,
cls.get_enum_items)
if Auto.is_defined(list_available_scales):
str_available_scales = ' or '.join(map(str, list_available_scales))
else:
str_available_scales = '{}'
return 'Expected time-scale {}, got {}'.format(str_available_scales,
scale)
def get_year_and_week_from_week_abs(week_abs: int,
min_year: Union[int,
Auto] = AUTO) -> tuple:
min_year = Auto.acquire(min_year, _min_year)
delta_year = int(week_abs / WEEKS_IN_YEAR)
year = min_year + delta_year
week = week_abs - delta_year * WEEKS_IN_YEAR
return year, week
def get_week_abs_from_year_and_week(
year: int,
week: int,
min_year: Union[int, Auto] = AUTO,
) -> int:
min_year = Auto.acquire(min_year, get_min_year())
week_abs = (year - min_year) * WEEKS_IN_YEAR + week
return week_abs
def get_str_count(self, default: Optional[str] = '(iter)') -> Optional[str]:
if hasattr(self, 'get_count'):
count = self.get_count()
else:
count = None
if Auto.is_defined(count):
return str(count)
else:
return default
def get_year_start_monday(year: Union[int, Date],
as_iso_date: AutoBool = True) -> Date:
as_iso_date = Auto.delayed_acquire(as_iso_date, is_iso_date, year)
if isinstance(year, int):
year_start_date = PyDate(year=year, month=1, day=1)
else:
year_start_date = get_year_first_date(year, as_iso_date=as_iso_date)
year_start_monday = year_start_date + timedelta(
days=-year_start_date.weekday())
return to_date(year_start_monday, as_iso_date)
def get_month_first_date(d: Date, as_iso_date: AutoBool = AUTO) -> Date:
as_iso_date = Auto.delayed_acquire(as_iso_date, is_iso_date, d)
if as_iso_date:
if not is_iso_date(d):
d = to_date(as_iso_date=True)
return d[:8] + '01'
else:
if not is_py_date(d):
d = get_py_date(d)
return PyDate(d.year, d.month, 1)
def get_date_from_day_abs(
day_abs: int,
min_date: Union[Date, Auto] = AUTO,
as_iso_date: bool = True,
) -> Date:
min_date = Auto.delayed_acquire(min_date,
get_year_start_monday,
get_min_year(),
as_iso_date=as_iso_date)
cur_date = get_shifted_date(min_date, days=day_abs)
return cur_date
def get_rounded_date(d: Date,
scale: DateScale,
as_iso_date: AutoBool = AUTO) -> Date:
scale = DateScale.convert(scale)
as_iso_date = Auto.delayed_acquire(as_iso_date, is_iso_date, d)
if scale == DateScale.Day:
return to_date(d, as_iso_date=as_iso_date)
elif scale == DateScale.Week:
return get_monday_date(d, as_iso_date=as_iso_date)
elif scale == DateScale.Month:
return get_month_first_date(d, as_iso_date=as_iso_date)
elif scale == DateScale.Year:
return get_year_first_date(d, as_iso_date=as_iso_date)
else:
raise ValueError(DateScale.get_err_msg(scale))
def get_year_first_date(d: [Date, int], as_iso_date: AutoBool = True) -> Date:
as_iso_date = Auto.delayed_acquire(as_iso_date, is_iso_date, d)
if isinstance(d, int):
if d > 1900:
return get_date_from_year(d, as_iso_date=as_iso_date)
else:
return get_date_from_year_and_month(year=d,
month=1,
as_iso_date=as_iso_date)
else: # isinstance(d, Date)
year_no = get_year_from_date(d, decimal=False)
iso_date = '{}-01-01'.format(year_no)
if as_iso_date:
return iso_date
else:
return get_date(iso_date)
def get_one_line_repr(
self,
str_meta: Union[str, Auto, None] = AUTO,
max_len: int = JUPYTER_LINE_LEN,
crop: str = CROP_SUFFIX,
) -> str:
template = '{cls}({meta})'
class_name = self.__class__.__name__
str_meta = Auto.delayed_acquire(str_meta, self.get_str_meta)
one_line_repr = template.format(cls=class_name, meta=str_meta)
full_line_len = len(one_line_repr)
if full_line_len > max_len:
exceeded_len = full_line_len + len(crop) - max_len
str_meta = str_meta[:-exceeded_len]
one_line_repr = template.format(cls=class_name, meta=str_meta + crop)
return one_line_repr
def get_date_from_month_abs(
month_abs: int,
min_date: Union[Date, Auto] = AUTO,
as_iso_date: bool = True,
) -> Date:
if Auto.is_defined(min_date):
min_year = get_year_from_date(min_date)
else:
min_year = get_min_year()
year_delta = int((month_abs - 1) / MONTHS_IN_YEAR)
year_no = min_year + year_delta
month_no = month_abs - year_delta * MONTHS_IN_YEAR
cur_date = get_date_from_year_and_month(year=year_no,
month=month_no,
as_iso_date=as_iso_date)
return cur_date
def get_meta_records(self) -> Generator:
init_defaults = self._get_init_defaults()
for key, value in self.get_meta_items():
actual_type = type(value).__name__
expected_type = self._get_init_types().get(key)
if hasattr(expected_type, '__name__'):
expected_type = expected_type.__name__
else:
expected_type = str(expected_type).replace('typing.', '')
default = init_defaults.get(key)
yield dict(
key=key,
value=self._get_value_repr(value),
default=self._get_value_repr(default),
actual_type=actual_type,
expected_type=expected_type or '-',
defined='-' if value == default else '+' if Auto.is_defined(value) else 'x',
)
def output_line(self, line: str, output: AutoOutput = AUTO) -> None:
if isinstance(output, LoggingLevel):
logger_kwargs = dict(level=output)
output = AUTO
else:
logger_kwargs = dict()
if Auto.is_auto(output):
if hasattr(self, 'log'):
return self.log(msg=line, **logger_kwargs)
else:
output = self.get_output()
if isinstance(output, Callable):
return output(line)
elif output:
if hasattr(output, 'output_line'):
return output.output_line(line)
elif hasattr(output, 'log'):
return output.log(msg=line, **logger_kwargs)
else:
raise TypeError('Expected Output, Logger or Auto, got {}'.format(output))
def get_data_description(
self,
count: int = DEFAULT_ROWS_COUNT,
title: Optional[str] = 'Data:',
max_len: int = JUPYTER_LINE_LEN,
) -> Generator:
if title:
yield title
if hasattr(self, 'get_data_caption'):
yield self.get_data_caption()
if hasattr(self, 'get_data'):
data = self.get_data()
if data:
shape_repr = self.get_shape_repr()
if shape_repr:
yield 'First {count} data items from {shape}:'.format(count=count, shape=shape_repr)
if isinstance(data, dict):
records = map(
lambda i: dict(key=i[0], value=i[1], defined='+' if Auto.is_defined(i[1]) else '-'),
data.items(),
)
yield from self._get_columnar_lines(
records, columns=DICT_DESCRIPTION_COLUMNS, count=count, max_len=max_len,
)
elif isinstance(data, Iterable):
for n, item in enumerate(data):
if n >= count:
break
line = ' - ' + str(item)
yield line[:max_len]
elif isinstance(data, DescribeMixin) or hasattr(data, 'get_meta_description'):
for line in data.get_meta_description():
yield line
else:
line = str(data)
yield line[:max_len]
else:
yield '(data attribute is empty)'
else:
yield '(data attribute not found)'
def describe(
self,
show_header: bool = True,
count: AutoCount = AUTO,
comment: Optional[str] = None,
depth: int = 1,
output: AutoOutput = AUTO,
as_dataframe: bool = Auto,
**kwargs
):
as_dataframe = Auto.acquire(as_dataframe, hasattr(self, 'show') or hasattr(self, 'show_example'))
show_meta = show_header or not self.has_data()
if show_header:
for line in self.get_str_headers():
self.output_line(line, output=output)
if comment:
self.output_line(comment, output=output)
if show_meta:
for line in self.get_meta_description():
self.output_line(line, output=output)
if self.has_data():
if not as_dataframe:
self.output_blank_line(output=output)
for line in self.get_data_description(count=count, **kwargs):
self.output_line(line, output=output)
elif depth > 0:
for attribute, value in self.get_meta_items():
if isinstance(value, DescribeMixin) or hasattr(value, 'describe'):
self.output_blank_line(output=output)
self.output_line('{attribute}:'.format(attribute=attribute), output=output)
value.describe(show_header=False, depth=depth - 1, output=output)
if self.has_data() and as_dataframe:
if hasattr(self, 'show_example'):
return self.show_example(count=count, **kwargs)
elif hasattr(self, 'show'):
return self.show(count=count, **kwargs)
else:
raise AttributeError('{} does not support dataframe'.format(self))
def _get_columnar_lines(
cls,
records: Iterable,
columns: Array,
count: AutoCount = None,
with_title: bool = True,
prefix: str = PREFIX_VALUE,
delimiter: str = ' ',
max_len: int = JUPYTER_LINE_LEN,
) -> Generator:
count = Auto.acquire(count, DEFAULT_ROWS_COUNT)
formatter = cls._get_formatter(columns=columns, delimiter=delimiter)
if with_title:
column_names = cls._get_column_names(columns, ex=PREFIX_FIELD)
title_record = cls._get_cropped_record(column_names, columns=columns, max_len=max_len, ex=PREFIX_FIELD)
yield formatter.format(**{k: v.upper() for k, v in title_record.items()})
for n, r in enumerate(records):
if count is not None and n >= count:
break
if prefix and PREFIX_FIELD not in r:
r[PREFIX_FIELD] = prefix
r = cls._get_cropped_record(r, columns=columns, max_len=max_len)
yield formatter.format(**r)
def get_monday_date(d: Date, as_iso_date: AutoBool = AUTO) -> Date:
cur_date = get_date(d)
if not Auto.is_defined(as_iso_date):
as_iso_date = is_iso_date(d)
monday_date = cur_date + timedelta(days=-cur_date.weekday())
return to_date(monday_date, as_iso_date)
def get_year_from_week_abs(week_abs: int,
min_year: Union[int, Auto] = AUTO) -> int:
min_year = Auto.acquire(min_year, _min_year)
delta_year = int(week_abs / WEEKS_IN_YEAR)
return min_year + delta_year
def get_count_repr(self, default: str = '<iter>') -> str:
count = self.get_str_count(default=default)
if not Auto.is_defined(count):
count = default
return '{} items'.format(count)
def get_day_abs_from_date(d: Date, min_date: Union[Date, Auto] = AUTO) -> int:
min_date = Auto.delayed_acquire(min_date, get_year_start_monday,
get_min_year())
return get_days_between(min_date, d)