def _input_string_transform_factory(alg):
"""
Given a set of natsort algorithms, return the function to operate
on the pre-split input string according to the user's request.
"""
# Shortcuts.
lowfirst = alg & ns.LOWERCASEFIRST
dumb = alg & ns._DUMB
# Build the chain of functions to execute in order.
function_chain = []
if (dumb and not lowfirst) or (lowfirst and not dumb):
function_chain.append(methodcaller('swapcase'))
if alg & ns.IGNORECASE:
if NEWPY:
function_chain.append(methodcaller('casefold'))
else:
function_chain.append(methodcaller('lower'))
if alg & ns.LOCALENUM:
# Create a regular expression that will remove thousands separators.
strip_thousands = r'''
(?<=[0-9]{{1}}) # At least 1 number
(?<![0-9]{{4}}) # No more than 3 numbers
{nodecimal} # Cannot follow decimal
{thou} # The thousands separator
(?=[0-9]{{3}} # Three numbers must follow
([^0-9]|$) # But a non-number after that
)
'''
nodecimal = r''
if alg & ns.FLOAT:
# Make a regular expression component that will ensure no
# separators are removed after a decimal point.
d = get_decimal_point()
d = r'\.' if d == r'.' else d
nodecimal += r'(?<!' + d + r'[0-9])'
nodecimal += r'(?<!' + d + r'[0-9]{2})'
nodecimal += r'(?<!' + d + r'[0-9]{3})'
strip_thousands = strip_thousands.format(thou=get_thousands_sep(),
nodecimal=nodecimal)
strip_thousands = re.compile(strip_thousands, flags=re.VERBOSE)
function_chain.append(partial(strip_thousands.sub, ''))
# Create a regular expression that will change the decimal point to
# a period if not already a period.
decimal = get_decimal_point()
if alg & ns.FLOAT and decimal != '.':
switch_decimal = r'(?<=[0-9]){decimal}|{decimal}(?=[0-9])'
switch_decimal = switch_decimal.format(decimal=decimal)
switch_decimal = re.compile(switch_decimal)
function_chain.append(partial(switch_decimal.sub, '.'))
# Return the chained functions.
return chain_functions(function_chain)
def input_string_transform_factory(alg):
"""
Create a function to transform a string.
Parameters
----------
alg : ns enum
Indicate how to format the *str*.
Returns
-------
func : callable
A function to be used as the *input_transform* argument to
*parse_string_factory*.
See Also
--------
parse_string_factory
"""
# Shortcuts.
lowfirst = alg & ns.LOWERCASEFIRST
dumb = alg & NS_DUMB
# Build the chain of functions to execute in order.
function_chain = []
if (dumb and not lowfirst) or (lowfirst and not dumb):
function_chain.append(methodcaller("swapcase"))
if alg & ns.IGNORECASE:
function_chain.append(methodcaller("casefold"))
if alg & ns.LOCALENUM:
# Create a regular expression that will remove thousands separators.
strip_thousands = r"""
(?<=[0-9]{{1}}) # At least 1 number
(?<![0-9]{{4}}) # No more than 3 numbers
{nodecimal} # Cannot follow decimal
{thou} # The thousands separator
(?=[0-9]{{3}} # Three numbers must follow
([^0-9]|$) # But a non-number after that
)
"""
nodecimal = r""
if alg & ns.FLOAT:
# Make a regular expression component that will ensure no
# separators are removed after a decimal point.
d = re.escape(get_decimal_point())
nodecimal += r"(?<!" + d + r"[0-9])"
nodecimal += r"(?<!" + d + r"[0-9]{2})"
nodecimal += r"(?<!" + d + r"[0-9]{3})"
strip_thousands = strip_thousands.format(
thou=re.escape(get_thousands_sep()), nodecimal=nodecimal
)
strip_thousands = re.compile(strip_thousands, flags=re.VERBOSE)
function_chain.append(partial(strip_thousands.sub, ""))
# Create a regular expression that will change the decimal point to
# a period if not already a period.
decimal = get_decimal_point()
if alg & ns.FLOAT and decimal != ".":
switch_decimal = r"(?<=[0-9]){decimal}|{decimal}(?=[0-9])"
switch_decimal = switch_decimal.format(decimal=re.escape(decimal))
switch_decimal = re.compile(switch_decimal)
function_chain.append(partial(switch_decimal.sub, "."))
# Return the chained functions.
return chain_functions(function_chain)
def input_string_transform_factory(alg):
"""
Create a function to transform a string.
Parameters
----------
alg : ns enum
Indicate how to format the *str*.
Returns
-------
func : callable
A function to be used as the *input_transform* argument to
*parse_string_factory*.
See Also
--------
parse_string_factory
"""
# Shortcuts.
lowfirst = alg & ns.LOWERCASEFIRST
dumb = alg & NS_DUMB
# Build the chain of functions to execute in order.
function_chain = []
if (dumb and not lowfirst) or (lowfirst and not dumb):
function_chain.append(methodcaller("swapcase"))
if alg & ns.IGNORECASE:
if NEWPY:
function_chain.append(methodcaller("casefold"))
else:
function_chain.append(methodcaller("lower"))
if alg & ns.LOCALENUM:
# Create a regular expression that will remove thousands separators.
strip_thousands = r"""
(?<=[0-9]{{1}}) # At least 1 number
(?<![0-9]{{4}}) # No more than 3 numbers
{nodecimal} # Cannot follow decimal
{thou} # The thousands separator
(?=[0-9]{{3}} # Three numbers must follow
([^0-9]|$) # But a non-number after that
)
"""
nodecimal = r""
if alg & ns.FLOAT:
# Make a regular expression component that will ensure no
# separators are removed after a decimal point.
d = get_decimal_point()
d = r"\." if d == r"." else d
nodecimal += r"(?<!" + d + r"[0-9])"
nodecimal += r"(?<!" + d + r"[0-9]{2})"
nodecimal += r"(?<!" + d + r"[0-9]{3})"
strip_thousands = strip_thousands.format(
thou=get_thousands_sep(), nodecimal=nodecimal
)
strip_thousands = re.compile(strip_thousands, flags=re.VERBOSE)
function_chain.append(partial(strip_thousands.sub, ""))
# Create a regular expression that will change the decimal point to
# a period if not already a period.
decimal = get_decimal_point()
if alg & ns.FLOAT and decimal != ".":
switch_decimal = r"(?<=[0-9]){decimal}|{decimal}(?=[0-9])"
switch_decimal = switch_decimal.format(decimal=decimal)
switch_decimal = re.compile(switch_decimal)
function_chain.append(partial(switch_decimal.sub, "."))
# Return the chained functions.
return chain_functions(function_chain)