from datetime import datetime
from io import StringIO
import re
import sys
import time
from typing import Optional, TextIO, Union
from .util import itemize
if sys.version_info[:2] >= (3,9):
from collections.abc import Iterable, Mapping
from re import Match
Tuple = tuple
else:
from typing import Iterable, Mapping, Match, Tuple
[docs]def dump(
props: Union[Mapping[str,str], Iterable[Tuple[str,str]]],
fp: TextIO,
separator: str = '=',
comments: Optional[str] = None,
timestamp: Union[None, bool, float, datetime] = True,
sort_keys: bool = False,
ensure_ascii: bool = True,
ensure_ascii_comments: Optional[bool] = None,
) -> None:
"""
Write a series of key-value pairs to a file in simple line-oriented
``.properties`` format.
.. versionchanged:: 0.6.0
``ensure_ascii`` and ``ensure_ascii_comments`` parameters added
:param props: A mapping or iterable of ``(key, value)`` pairs to write to
``fp``. All keys and values in ``props`` must be `str` values. If
``sort_keys`` is `False`, the entries are output in iteration order.
:param TextIO fp: A file-like object to write the values of ``props`` to.
It must have been opened as a text file.
:param str separator: The string to use for separating keys & values. Only
``" "``, ``"="``, and ``":"`` (possibly with added whitespace) should
ever be used as the separator.
:param Optional[str] comments: if non-`None`, ``comments`` will be written
to ``fp`` as a comment before any other content
:param timestamp: If neither `None` nor `False`, a timestamp in the form of
``Mon Sep 02 14:00:54 EDT 2016`` is written as a comment to ``fp``
after ``comments`` (if any) and before the key-value pairs. If
``timestamp`` is `True`, the current date & time is used. If it is a
number, it is converted from seconds since the epoch to local time. If
it is a `datetime.datetime` object, its value is used directly, with
naïve objects assumed to be in the local timezone.
:type timestamp: `None`, `bool`, number, or `datetime.datetime`
:param bool sort_keys: if true, the elements of ``props`` are sorted
lexicographically by key in the output
:param bool ensure_ascii: if true, all non-ASCII characters will be
replaced with ``\\uXXXX`` escape sequences in the output; if false,
non-ASCII characters will be passed through as-is
:param Optional[bool] ensure_ascii_comments: if true, all non-ASCII
characters in ``comments`` will be replaced with ``\\uXXXX`` escape
sequences in the output; if `None`, only non-Latin-1 characters will be
escaped; if false, no characters will be escaped
:return: `None`
"""
if comments is not None:
print(to_comment(comments, ensure_ascii=ensure_ascii_comments), file=fp)
if timestamp is not None and timestamp is not False:
print(to_comment(java_timestamp(timestamp)), file=fp)
for k,v in itemize(props, sort_keys=sort_keys):
print(
join_key_value(k, v, separator, ensure_ascii=ensure_ascii),
file=fp,
)
[docs]def dumps(
props: Union[Mapping[str,str], Iterable[Tuple[str,str]]],
separator: str = '=',
comments: Optional[str] = None,
timestamp: Union[None, bool, float, datetime] = True,
sort_keys: bool = False,
ensure_ascii: bool = True,
ensure_ascii_comments: Optional[bool] = None,
) -> str:
"""
Convert a series of key-value pairs to a `str` in simple line-oriented
``.properties`` format.
.. versionchanged:: 0.6.0
``ensure_ascii`` and ``ensure_ascii_comments`` parameters added
:param props: A mapping or iterable of ``(key, value)`` pairs to serialize.
All keys and values in ``props`` must be `str` values. If
``sort_keys`` is `False`, the entries are output in iteration order.
:param str separator: The string to use for separating keys & values. Only
``" "``, ``"="``, and ``":"`` (possibly with added whitespace) should
ever be used as the separator.
:param Optional[str] comments: if non-`None`, ``comments`` will be output
as a comment before any other content
:param timestamp: If neither `None` nor `False`, a timestamp in the form of
``Mon Sep 02 14:00:54 EDT 2016`` is output as a comment after
``comments`` (if any) and before the key-value pairs. If ``timestamp``
is `True`, the current date & time is used. If it is a number, it is
converted from seconds since the epoch to local time. If it is a
`datetime.datetime` object, its value is used directly, with naïve
objects assumed to be in the local timezone.
:type timestamp: `None`, `bool`, number, or `datetime.datetime`
:param bool sort_keys: if true, the elements of ``props`` are sorted
lexicographically by key in the output
:param bool ensure_ascii: if true, all non-ASCII characters will be
replaced with ``\\uXXXX`` escape sequences in the output; if false,
non-ASCII characters will be passed through as-is
:param Optional[bool] ensure_ascii_comments: if true, all non-ASCII
characters in ``comments`` will be replaced with ``\\uXXXX`` escape
sequences in the output; if `None`, only non-Latin-1 characters will be
escaped; if false, no characters will be escaped
:rtype: text string
"""
s = StringIO()
dump(props, s, separator=separator, comments=comments, timestamp=timestamp,
sort_keys=sort_keys, ensure_ascii=ensure_ascii,
ensure_ascii_comments=ensure_ascii_comments)
return s.getvalue()
NON_ASCII_RGX = re.compile(r'[^\x00-\x7F]')
NON_LATIN1_RGX = re.compile(r'[^\x00-\xFF]')
NEWLINE_OLD_COMMENT_RGX = re.compile(r'\n(?![#!])')
NON_N_EOL_RGX = re.compile(r'\r\n?')
[docs]def join_key_value(
key: str,
value: str,
separator: str = '=',
ensure_ascii: bool = True,
) -> str:
r"""
Join a key and value together into a single line suitable for adding to a
simple line-oriented ``.properties`` file. No trailing newline is added.
>>> join_key_value('possible separators', '= : space')
'possible\\ separators=\\= \\: space'
.. versionchanged:: 0.6.0
``ensure_ascii`` parameter added
:param str key: the key
:param str value: the value
:param str separator: the string to use for separating the key & value.
Only ``" "``, ``"="``, and ``":"`` (possibly with added whitespace)
should ever be used as the separator.
:param bool ensure_ascii: if true, all non-ASCII characters will be
replaced with ``\\uXXXX`` escape sequences in the output; if false,
non-ASCII characters will be passed through as-is
:rtype: str
"""
# Escapes `key` and `value` the same way as java.util.Properties.store()
value = _base_escape(value, ensure_ascii=ensure_ascii)
if value.startswith(' '):
value = '\\' + value
return escape(key, ensure_ascii=ensure_ascii) + separator + value
_escapes = {
'\t': r'\t',
'\n': r'\n',
'\f': r'\f',
'\r': r'\r',
'!': r'\!',
'#': r'\#',
':': r'\:',
'=': r'\=',
'\\': r'\\',
}
def _esc(m: Match[str]) -> str:
c = m.group()
try:
return _escapes[c]
except KeyError:
return _to_u_escape(c)
def _to_u_escape(c: str) -> str:
co = ord(c)
if co > 0xFFFF:
# Does Python really not have a decent builtin way to calculate
# surrogate pairs?
assert co <= 0x10FFFF
co -= 0x10000
return '\\u{0:04x}\\u{1:04x}'.format(
0xD800 + (co >> 10),
0xDC00 + (co & 0x3FF)
)
else:
return f'\\u{co:04x}'
NEEDS_ESCAPE_ASCII_RGX = re.compile(r'[^\x20-\x7E]|[\\#!=:]')
NEEDS_ESCAPE_UNICODE_RGX = re.compile(r'[\x00-\x1F\x7F]|[\\#!=:]')
def _base_escape(field: str, ensure_ascii: bool = True) -> str:
rgx = NEEDS_ESCAPE_ASCII_RGX if ensure_ascii else NEEDS_ESCAPE_UNICODE_RGX
return rgx.sub(_esc, field)
[docs]def escape(field: str, ensure_ascii: bool = True) -> str:
"""
Escape a string so that it can be safely used as either a key or value in a
``.properties`` file. All non-ASCII characters, all nonprintable or space
characters, and the characters ``\\ # ! = :`` are all escaped using either
the single-character escapes recognized by `unescape` (when they exist) or
``\\uXXXX`` escapes (after converting non-BMP characters to surrogate
pairs).
.. versionchanged:: 0.6.0
``ensure_ascii`` parameter added
:param str field: the string to escape
:param bool ensure_ascii: if true, all non-ASCII characters will be
replaced with ``\\uXXXX`` escape sequences in the output; if false,
non-ASCII characters will be passed through as-is
:rtype: str
"""
return _base_escape(field, ensure_ascii=ensure_ascii).replace(' ', r'\ ')
DAYS_OF_WEEK = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
MONTHS = [
'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec',
]
[docs]def java_timestamp(timestamp: Union[None, bool, float, datetime] = True) -> str:
"""
.. versionadded:: 0.2.0
Returns a timestamp in the format produced by |date_tostring|_, e.g.::
Mon Sep 02 14:00:54 EDT 2016
If ``timestamp`` is `True` (the default), the current date & time is
returned.
If ``timestamp`` is `None` or `False`, an empty string is returned.
If ``timestamp`` is a number, it is converted from seconds since the epoch
to local time.
If ``timestamp`` is a `datetime.datetime` object, its value is used
directly, with naïve objects assumed to be in the local timezone.
The timestamp is always constructed using the C locale.
:param timestamp: the date & time to display
:type timestamp: `None`, `bool`, number, or `datetime.datetime`
:rtype: str
.. |date_tostring| replace:: Java 8's ``Date.toString()``
.. _date_tostring: https://docs.oracle.com/javase/8/docs/api/java/util/Date.html#toString--
"""
if timestamp is None or timestamp is False:
return ''
if isinstance(timestamp, datetime) and timestamp.tzinfo is not None:
timebits = timestamp.timetuple()
# Assumes `timestamp.tzinfo.tzname()` is meaningful/useful
tzname = timestamp.tzname()
else:
### TODO: Reimplement this using datetime.astimezone() to convert
### everything to an aware datetime?
ts: Optional[float]
if timestamp is True:
ts = None
elif isinstance(timestamp, datetime):
# Use `datetime.timestamp()`, as it (unlike `datetime.timetuple()`)
# takes `fold` into account for naïve datetimes.
ts = timestamp.timestamp()
else:
# If it's not a number, it's localtime()'s problem now.
ts = timestamp
timebits = time.localtime(ts)
tzname = timebits.tm_zone
assert 1 <= timebits.tm_mon <= 12, 'invalid month'
assert 0 <= timebits.tm_wday <= 6, 'invalid day of week'
return '{wday} {mon} {t.tm_mday:02d}' \
' {t.tm_hour:02d}:{t.tm_min:02d}:{t.tm_sec:02d}' \
' {tz} {t.tm_year:04d}'.format(
t=timebits,
tz=tzname,
mon=MONTHS[timebits.tm_mon-1],
wday=DAYS_OF_WEEK[timebits.tm_wday]
)
[docs]def javapropertiesreplace_errors(e: Union[UnicodeError]) -> Tuple[str, int]:
"""
.. versionadded:: 0.6.0
Implements the ``'javapropertiesreplace'`` error handling (for text
encodings only): unencodable characters are replaced by ``\\uXXXX`` escape
sequences (with non-BMP characters converted to surrogate pairs first)
"""
if isinstance(e, UnicodeEncodeError):
return (''.join(map(_to_u_escape, e.object[e.start:e.end])), e.end)
else:
raise e # pragma: no cover