def add_language(self, language, force=False):
"""
Add a language handled by the LGR to the metadata list.
Ensure the language is a valid RFC 5646 language tag.
3.3.3. The language Element
The value of the "language"
element MUST be a valid language tag as described in [RFC5646].
:param str language: A new language of the LGR.
:param bool force: If True, add the language even if format is invalid.
:raises LGRFormatException: if the language parameter
has an invalid format.
"""
try:
if not rfc5646.check(language):
logger.log(logging.WARNING if force else logging.ERROR,
"Invalid language: '%s'", language)
if not force:
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_LANGUAGE_TAG
)
self.languages.append(language)
except UnicodeEncodeError:
# Can't skip this one
logger.error("Invalid non-ASCII language tag '%s'", language)
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_LANGUAGE_TAG)
def validate(self, parents, rules_lookup, classes_lookup):
"""
Ensure a class has a valid definition.
:param parents: List of parent objects of this instance.
:param rules_lookup: Dictionary of defined rules in the LGR. Not used.
:param classes_lookup: Dictionary of defined classes in the LGR to use
for by-ref classes.
"""
logger.debug('Validate %s', self)
if self.by_ref is not None:
if self.by_ref not in classes_lookup:
# From RFC7940, section 6.2.1. Declaring and Invoking Named Classes
# It is an error to reference a named class for which the
# definition has not been seen.
logger.error("Class cannot reference inexisting class '%s'",
self.by_ref)
raise LGRFormatException(LGRFormatException.LGRFormatReason.INVALID_BY_REF)
is_top_level = len(parents) == 0
if (is_top_level and self.name is None) or \
(not is_top_level and self.name is not None):
# From RFC7940, section 6.2.1. Declaring and Invoking Named Classes
# The "name" attribute MUST be present if and only if the class
# is a direct child element of the "rules" element.
logger.error("'name' attribute MUST be present only and only if "
"class is a direct child of the 'rules' element")
raise LGRFormatException(LGRFormatException.LGRFormatReason.INVALID_TOP_LEVEL_NAME)
if self.from_tag is not None:
tag_classname = TAG_CLASSNAME_PREFIX + self.from_tag
if tag_classname not in classes_lookup:
logger.warning("Undefined tag '%s' in current LGR",
self.from_tag)
def __init__(self,
name=None,
comment=None,
ref=None,
from_tag=None,
unicode_property=None,
codepoints=None,
by_ref=None):
"""
Create a class.
:param name: Name of the class.
:param comment: Optional comment associated to the class.
:param ref: Optional list of references.
:param from_tag: Define a tag-based class.
:param unicode_property: Define a Unicode property-based class.
:param codepoints: Initial sequence of code points.
:param by_ref: Name of the referenced class.
"""
self.name = name
self.comment = comment
self.ref = ref
self.from_tag = from_tag
self.unicode_property = unicode_property
self.codepoints = set(codepoints or [])
self.by_ref = by_ref
if by_ref is not None:
if name is not None:
logger.error("Cannot create a class with "
"both a 'by-ref' and a 'name'")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
elif from_tag is not None:
logger.error("Cannot create a class with "
"both a 'by-ref' and a 'from-tag'")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
elif unicode_property is not None:
logger.error("Cannot create a class with "
"both a 'by-ref' and a 'property'")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
elif ref is not None:
logger.error("Cannot create a class with "
"both a 'by-ref' and a 'ref'")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
def set_unicode_version(self, unicode_version, force=False):
"""
Set the unicode-version of the LGR.
Ensure the unicode_version is a valid x.y.z string.
3.3.7. The unicode-version Element
the version number used in creating the LGR
MUST be listed in the form x.y.z, where x, y, and z are positive,
decimal integers (see [Unicode-Versions]).
:param str unicode_version: The Unicode version.
:param bool force: If True, set the date even if format is invalid.
:raises LGRFormatException: If the unicode_version parameter
has an invalid format.
"""
if re.match(r'\d{1,}\.\d{1,}\.\d{1,}', unicode_version) is None:
logger.log(logging.WARNING if force else logging.ERROR,
"Invalid Unicode version: '%s'", unicode_version)
if not force:
raise LGRFormatException(LGRFormatException.LGRFormatReason.
INVALID_UNICODE_VERSION_TAG)
self.unicode_version = unicode_version
def set_languages(self, languages, force=False):
"""
Convenience function to update the languages in the metadata.
:param iterable languages: a collection of language tags as described in [RFC5646].
:param bool force: If True, add the languages even if format is invalid.
:raises LGRFormatException: if the language parameter
has an invalid format.
"""
# check all languages
found_error = False
for language in languages:
try:
if not rfc5646.check(language):
logger.log(logging.WARNING if force else logging.ERROR,
"Invalid language: '%s'", language)
found_error = True
except UnicodeEncodeError:
# Can't skip this one
logger.error("Invalid non-ASCII language tag '%s'", language)
languages.remove(language)
if found_error and not force:
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_LANGUAGE_TAG)
else:
self.languages = list(languages)
def _validate_date(date, force):
"""
Ensure the date is a valid ISO 8601 "full-date" string.
:param str date: Date to validate, as a string.
:param force: If True, do not raise exception on error.
:return: date input.
:raises LGRFormatException: If the date parameter
has an invalid format.
>>> _validate_date('2015-06-25', False) == '2015-06-25'
True
>>> _validate_date('2015-13-26', False) # doctest: +IGNORE_EXCEPTION_DETAIL
Traceback (most recent call last):
...
LGRFormatException
>>> _validate_date('2015', False) # doctest: +IGNORE_EXCEPTION_DETAIL
Traceback (most recent call last):
...
LGRFormatException
"""
# Date shoud be date-fullyear "-" date-month "-" date-mday
date_elements = date.split('-')
if len(date_elements) == 3:
try:
date_elements = [int(d) for d in date_elements]
date_object = datetime.date(date_elements[0], date_elements[1],
date_elements[2])
if date_object.isoformat() == date:
return date
except ValueError:
logger.log(logging.WARNING if force else logging.ERROR,
"Invalid date format: '%s'", date)
if force:
return date
else:
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_DATE_TAG)
logger.log(logging.WARNING if force else logging.ERROR,
"Invalid date format: '%s'", date)
if force:
return date
else:
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_DATE_TAG)
def validate(self, parents, rules_lookup, classes_lookup):
super(ComplementClass, self).validate(parents,
rules_lookup, classes_lookup)
if len(self._children) != 1:
logger.error("'Complement' class MUST contain one element")
raise LGRFormatException(LGRFormatException.LGRFormatReason.INVALID_CHILDREN_NUMBER)
logger.debug('%s is valid', self)
def validate(self, parents, rules_lookup, classes_lookup):
super(DifferenceClass, self).validate(parents,
rules_lookup, classes_lookup)
if len(self._children) != 2:
logger.error("'Difference' class MUST contain two elements")
raise LGRFormatException(LGRFormatException.LGRFormatReason.INVALID_CHILDREN_NUMBER)
logger.debug('%s is valid', self)
def add_child(self, cls):
if self.MAX_CHILDREN > 0 and len(self._children) >= self.MAX_CHILDREN:
logger.error("Cannot have more than %s children",
self.MAX_CHILDREN)
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_CHILDREN_NUMBER)
self._children.append(cls)
def add_codepoint(self, cp):
"""
Add (a) codepoint(s) to the set of codepoints.
:param cp: Code point(s) to add. An integer or a sequence of code points.
"""
if self.by_ref is not None:
logger.error("Cannot add code point to a 'by-ref' class")
raise LGRFormatException(LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
if isinstance(cp, int):
cp = [cp]
self.codepoints.update(cp)
def del_codepoint(self, cp):
"""
Delete (a) codepoint(s) from the set of codepoints.
:param cp: Code point(s) to delete. An integer or a sequence of code points.
"""
if self.by_ref is not None:
logger.error("Cannot delete code point from a 'by-ref' class")
raise LGRFormatException(LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
if isinstance(cp, int):
cp = [cp]
for c in cp:
self.codepoints.discard(c)
def validate(self, parents, rules_lookup, classes_lookup):
"""
Ensure a rule has a valid definition.
:param parents: List of parent objects of this instance.
:param rules_lookup: Dictionary of defined rules in the LGR. Not used.
:param classes_lookup: Dictionary of defined classes in the LGR to use
for by-ref classes.
"""
logger.debug('Validate %s', self)
if self.by_ref is not None:
if self.by_ref not in rules_lookup:
# From RFC7940, section 6.3.4. The "name" and "by-ref" Attributes
# It is an error to reference a rule or class for which
# the definition has not been seen.
logger.error("Rule cannot reference inexisting rule '%s'",
self.by_ref)
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_BY_REF)
is_top_level = len(parents) == 0
if (is_top_level and self.name is None) or \
(not is_top_level and self.name is not None):
# From RFC7940, section section 6.3.4. The name and by-ref Attributes
# rules declared as immediate child elements of the "rules" element
# MUST be named using a unique "name" attribute,
# and all other instances MUST NOT be named.
logger.error("'name' attribute MUST be present only and only if "
"rule is a direct child of the 'rules' element")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.INVALID_TOP_LEVEL_NAME)
logger.debug('Validate %s children', self)
for child in self.children:
child.validate(parents + [self], rules_lookup, classes_lookup)
logger.debug('%s is valid', self)
def __init__(self, name=None, comment=None, by_ref=None):
"""
Create a new rule.
:param name: Name of the rule.
:param comment: Optional comment.
:param by_ref: Name of the referenced rule.
:raises LGRFormatException: If the rule has conflictual parameters.
"""
self.name = name
self.comment = comment
self.by_ref = by_ref
self.children = []
if name is not None and by_ref is not None:
logger.error(
"Cannot create a rule with both a 'name' and a 'by-ref'")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
def __init__(self,
disp,
comment=None,
ref=None,
match=None,
not_match=None,
any_variant=None,
all_variants=None,
only_variants=None):
"""
Create the action element.
:param disp: Disposition of the action.
:param comment: Optional comment.
:param ref: Optional list of references.
:param match: Name of a rule that must be matched.
:param not_match: Name of a rule that must not be matched.
:param any_variant: Sequence of disposition to match to trigger action.
:param all_variants: Sequence of disposition to match to trigger action.
:param only_variants: Sequence of disposition to match to trigger action.
"""
self.disp = disp
self.comment = comment
self.references = ref or []
self.match = match
self.not_match = not_match
self.any_variant = frozenset(any_variant) if any_variant else None
self.all_variants = frozenset(all_variants) if all_variants else None
self.only_variants = frozenset(
only_variants) if only_variants else None
if match is not None and not_match is not None:
# From RFC7940, section 7.1. The "match" and "not-match" Attributes
# An action MUST NOT contain both a "match" and a "not-match" attribute
logger.error("Action contains both 'match' and 'not-match' "
"attributes")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.MATCH_NOT_MATCH)
def del_variant(self, *args, **kwargs):
# From draft-davies-idntables-09, section 4. Code Points and Variants
# A "range" element has no child elements.
logger.error("%r: Range has no variant", self)
raise LGRFormatException(LGRFormatException.
LGRFormatReason.RANGE_NO_CHILD)
def get_script(lgr):
try:
return lgr.metadata.languages[0]
except IndexError:
raise LGRFormatException(
reason=LGRFormatException.LGRFormatReason.INVALID_LANGUAGE_TAG)
def add_child(self, child):
if self.by_ref is not None:
logger.error("Cannot add child to a 'by-ref' rule")
raise LGRFormatException(
LGRFormatException.LGRFormatReason.BY_REF_AND_OTHER)
self.children.append(child)
def del_variant(self, *args, **kwargs):
# From RFC7940, section 5. Code Points and Variants
# A "range" element has no child elements.
logger.error("%r: Range has no variant", self)
raise LGRFormatException(
LGRFormatException.LGRFormatReason.RANGE_NO_CHILD)