Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
279 changes: 269 additions & 10 deletions src/rfc3987_syntax2/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -2,22 +2,281 @@
# Copyright (c) 2026 Jan Kowalleck - modifications and maintenance
# SPDX-License-Identifier: MIT

from typing import Any
from pathlib import Path
from threading import Lock
from typing import Any, Callable, Optional, TYPE_CHECKING, Literal
from warnings import warn

from . import syntax_helpers as _syntax_helpers
from lark import Lark, ParseTree, exceptions

__all__ = _syntax_helpers.__all__

_EXPORTS_SH = frozenset(_syntax_helpers.__all__)
RFC3987_SYNTAX_PARSER_TYPE: str = "earley"
RFC3987_SYNTAX_GRAMMAR_PATH: Path = Path(__file__).parent / "syntax_rfc3987.lark"
RFC3987_SYNTAX_TERMS: list[str] = [
"iri",
"iri_reference",
"absolute_iri",
"scheme",
"ihier_part",
"irelative_ref",
"irelative_part",
"iauthority",
"iuserinfo",
"ihost",
"ireg_name",
"ipath",
"ipath_abempty",
"ipath_absolute",
"ipath_noscheme",
"ipath_rootless",
"ipath_empty",
"isegment",
"isegment_nz",
"isegment_nz_nc",
"ipchar",
"iquery",
"ifragment",
"iunreserved",
"ucschar",
"iprivate",
"sub_delims",
"ip_literal",
"ipvfuture",
"ipv6address",
"h16",
"ls32",
"ipv4address",
"dec_octet",
"digit",
"non_zero",
"unreserved",
"alpha",
"hexdig",
"port",
"pct_encoded",
]

_T_SYNTAX_PARSER_TERM = Literal["iri", "iri_reference", "absolute_iri"]
_SYNTAX_PARSER_STARTS: list[_T_SYNTAX_PARSER_TERM] = ["iri", "iri_reference", "absolute_iri"]

def parse(term: _T_SYNTAX_PARSER_TERM, value: str) -> 'ParseTree':
return _get_syntax_parser().parse(value, start=term)


def is_valid_syntax(term: _T_SYNTAX_PARSER_TERM, value: str) -> bool:
try:
parse(term=term, value=value)
except exceptions.UnexpectedInput:
# from Lark.parse()
return False
except exceptions.LarkError as err:
# from Lark internals / initialization (non-UnexpectedInput)
warn("Unexpected LarkError (non-UnexpectedInput) "
f"for term={term!r}: {type(err).__name__}: {err}",
RuntimeWarning,
stacklevel=2)
return False
return True

_T_SYNTAX_VALIDATOR = Callable[[str], bool]

def make_syntax_validator(rule_name: str) -> _T_SYNTAX_VALIDATOR:
parser: Optional[Lark] = None
parser_lock = Lock()

def syntax_validator(text: str) -> bool:
nonlocal parser
with parser_lock:
if parser is None:
parser = Lark(_get_grammar(),
start=rule_name,
parser=RFC3987_SYNTAX_PARSER_TYPE)
try:
parser.parse(text)
except exceptions.LarkError:
return False
else:
return True

return syntax_validator


_grammar: Optional[str] = None
_grammar_lock = Lock()

def _get_grammar() -> str:
"""this is private API"""
global _grammar
with _grammar_lock:
if _grammar is None:
_grammar = RFC3987_SYNTAX_GRAMMAR_PATH.read_text(encoding='utf-8')
return _grammar


_syntax_parser: Optional[Lark] = None
_syntax_parser_lock = Lock()

def _get_syntax_parser() -> Lark:
"""this is private API"""
global _syntax_parser
if _syntax_parser is not None:
return _syntax_parser
grammar = _get_grammar()
with _syntax_parser_lock:
if _syntax_parser is None:
_syntax_parser = Lark(grammar,
start=_SYNTAX_PARSER_STARTS,
parser=RFC3987_SYNTAX_PARSER_TYPE)
return _syntax_parser


is_valid_syntax_iri = make_syntax_validator("iri")

is_valid_syntax_iri_reference = make_syntax_validator("iri_reference")

is_valid_syntax_absolute_iri = make_syntax_validator("absolute_iri")

is_valid_syntax_scheme = make_syntax_validator("scheme")

is_valid_syntax_ihier_part = make_syntax_validator("ihier_part")

is_valid_syntax_irelative_ref = make_syntax_validator("irelative_ref")

is_valid_syntax_irelative_part = make_syntax_validator("irelative_part")

is_valid_syntax_iauthority = make_syntax_validator("iauthority")

is_valid_syntax_iuserinfo = make_syntax_validator("iuserinfo")

is_valid_syntax_ihost = make_syntax_validator("ihost")

is_valid_syntax_ireg_name = make_syntax_validator("ireg_name")

is_valid_syntax_ipath = make_syntax_validator("ipath")

is_valid_syntax_ipath_abempty = make_syntax_validator("ipath_abempty")

is_valid_syntax_ipath_absolute = make_syntax_validator("ipath_absolute")

is_valid_syntax_ipath_noscheme = make_syntax_validator("ipath_noscheme")

is_valid_syntax_ipath_rootless = make_syntax_validator("ipath_rootless")

is_valid_syntax_ipath_empty = make_syntax_validator("ipath_empty")

is_valid_syntax_isegment = make_syntax_validator("isegment")

is_valid_syntax_isegment_nz = make_syntax_validator("isegment_nz")

is_valid_syntax_isegment_nz_nc = make_syntax_validator("isegment_nz_nc")

is_valid_syntax_ipchar = make_syntax_validator("ipchar")

is_valid_syntax_iquery = make_syntax_validator("iquery")

is_valid_syntax_ifragment = make_syntax_validator("ifragment")

is_valid_syntax_iunreserved = make_syntax_validator("iunreserved")

is_valid_syntax_ucschar = make_syntax_validator("ucschar")

is_valid_syntax_iprivate = make_syntax_validator("iprivate")

is_valid_syntax_sub_delims = make_syntax_validator("sub_delims")

is_valid_syntax_ip_literal = make_syntax_validator("ip_literal")

is_valid_syntax_ipvfuture = make_syntax_validator("ipvfuture")

is_valid_syntax_ipv6address = make_syntax_validator("ipv6address")

is_valid_syntax_h16 = make_syntax_validator("h16")

is_valid_syntax_ls32 = make_syntax_validator("ls32")

is_valid_syntax_ipv4address = make_syntax_validator("ipv4address")

is_valid_syntax_dec_octet = make_syntax_validator("dec_octet")

is_valid_syntax_digit = make_syntax_validator("digit")

is_valid_syntax_non_zero = make_syntax_validator("non_zero")

is_valid_syntax_unreserved = make_syntax_validator("unreserved")

is_valid_syntax_alpha = make_syntax_validator("alpha")

is_valid_syntax_hexdig = make_syntax_validator("hexdig")

is_valid_syntax_port = make_syntax_validator("port")

is_valid_syntax_pct_encoded = make_syntax_validator("pct_encoded")

RFC3987_SYNTAX_TERM_VALIDATORS: dict[str, _T_SYNTAX_VALIDATOR] = { # frozendict
"iri": is_valid_syntax_iri,
"iri_reference": is_valid_syntax_iri_reference,
"absolute_iri": is_valid_syntax_absolute_iri,
"scheme": is_valid_syntax_scheme,
"ihier_part": is_valid_syntax_ihier_part,
"irelative_ref": is_valid_syntax_irelative_ref,
"irelative_part": is_valid_syntax_irelative_part,
"iauthority": is_valid_syntax_iauthority,
"iuserinfo": is_valid_syntax_iuserinfo,
"ihost": is_valid_syntax_ihost,
"ireg_name": is_valid_syntax_ireg_name,
"ipath": is_valid_syntax_ipath,
"ipath_abempty": is_valid_syntax_ipath_abempty,
"ipath_absolute": is_valid_syntax_ipath_absolute,
"ipath_noscheme": is_valid_syntax_ipath_noscheme,
"ipath_rootless": is_valid_syntax_ipath_rootless,
"ipath_empty": is_valid_syntax_ipath_empty,
"isegment": is_valid_syntax_isegment,
"isegment_nz": is_valid_syntax_isegment_nz,
"isegment_nz_nc": is_valid_syntax_isegment_nz_nc,
"ipchar": is_valid_syntax_ipchar,
"iquery": is_valid_syntax_iquery,
"ifragment": is_valid_syntax_ifragment,
"iunreserved": is_valid_syntax_iunreserved,
"ucschar": is_valid_syntax_ucschar,
"iprivate": is_valid_syntax_iprivate,
"sub_delims": is_valid_syntax_sub_delims,
"ip_literal": is_valid_syntax_ip_literal,
"ipvfuture": is_valid_syntax_ipvfuture,
"ipv6address": is_valid_syntax_ipv6address,
"h16": is_valid_syntax_h16,
"ls32": is_valid_syntax_ls32,
"ipv4address": is_valid_syntax_ipv4address,
"dec_octet": is_valid_syntax_dec_octet,
"digit": is_valid_syntax_digit,
"non_zero": is_valid_syntax_non_zero,
"unreserved": is_valid_syntax_unreserved,
"alpha": is_valid_syntax_alpha,
"hexdig": is_valid_syntax_hexdig,
"port": is_valid_syntax_port,
"pct_encoded": is_valid_syntax_pct_encoded,
}


# region lazy loaded attrs

if TYPE_CHECKING: # types for lazy-loaded symbols
grammar: str
syntax_parser: Lark


def __getattr__(name: str) -> Any:
if name in _EXPORTS_SH:
value = getattr(_syntax_helpers, name)
globals()[name] = value
return value
if name == "grammar":
print('get_grammar')
global grammar
grammar = _get_grammar()
return grammar
if name == "syntax_parser":
global syntax_parser
syntax_parser = _get_syntax_parser()
return syntax_parser
raise AttributeError(f"module {__name__!r} has no attribute {name!r}")


def __dir__() -> list[str]:
return sorted(set(globals().keys()) | set(__all__))
return sorted(set(globals().keys()) | {"grammar", "syntax_parser"})

# endregion lazy loaded attrs
Loading