Cours/venv/lib/python3.12/site-packages/babel/messages/extract.py

844 lines
33 KiB
Python
Raw Normal View History

2024-09-02 16:55:06 +00:00
"""
babel.messages.extract
~~~~~~~~~~~~~~~~~~~~~~
Basic infrastructure for extracting localizable messages from source files.
This module defines an extensible system for collecting localizable message
strings from a variety of sources. A native extractor for Python source
files is builtin, extractors for other sources can be added using very
simple plugins.
The main entry points into the extraction functionality are the functions
`extract_from_dir` and `extract_from_file`.
:copyright: (c) 2013-2024 by the Babel Team.
:license: BSD, see LICENSE for more details.
"""
from __future__ import annotations
import ast
import io
import os
import sys
import tokenize
from collections.abc import (
Callable,
Collection,
Generator,
Iterable,
Mapping,
MutableSequence,
)
from functools import lru_cache
from os.path import relpath
from textwrap import dedent
from tokenize import COMMENT, NAME, OP, STRING, generate_tokens
from typing import TYPE_CHECKING, Any
from babel.messages._compat import find_entrypoints
from babel.util import parse_encoding, parse_future_flags, pathmatch
if TYPE_CHECKING:
from typing import IO, Final, Protocol
from _typeshed import SupportsItems, SupportsRead, SupportsReadline
from typing_extensions import TypeAlias, TypedDict
class _PyOptions(TypedDict, total=False):
encoding: str
class _JSOptions(TypedDict, total=False):
encoding: str
jsx: bool
template_string: bool
parse_template_string: bool
class _FileObj(SupportsRead[bytes], SupportsReadline[bytes], Protocol):
def seek(self, __offset: int, __whence: int = ...) -> int: ...
def tell(self) -> int: ...
_SimpleKeyword: TypeAlias = tuple[int | tuple[int, int] | tuple[int, str], ...] | None
_Keyword: TypeAlias = dict[int | None, _SimpleKeyword] | _SimpleKeyword
# 5-tuple of (filename, lineno, messages, comments, context)
_FileExtractionResult: TypeAlias = tuple[str, int, str | tuple[str, ...], list[str], str | None]
# 4-tuple of (lineno, message, comments, context)
_ExtractionResult: TypeAlias = tuple[int, str | tuple[str, ...], list[str], str | None]
# Required arguments: fileobj, keywords, comment_tags, options
# Return value: Iterable of (lineno, message, comments, context)
_CallableExtractionMethod: TypeAlias = Callable[
[_FileObj | IO[bytes], Mapping[str, _Keyword], Collection[str], Mapping[str, Any]],
Iterable[_ExtractionResult],
]
_ExtractionMethod: TypeAlias = _CallableExtractionMethod | str
GROUP_NAME: Final[str] = 'babel.extractors'
DEFAULT_KEYWORDS: dict[str, _Keyword] = {
'_': None,
'gettext': None,
'ngettext': (1, 2),
'ugettext': None,
'ungettext': (1, 2),
'dgettext': (2,),
'dngettext': (2, 3),
'N_': None,
'pgettext': ((1, 'c'), 2),
'npgettext': ((1, 'c'), 2, 3),
}
DEFAULT_MAPPING: list[tuple[str, str]] = [('**.py', 'python')]
# New tokens in Python 3.12, or None on older versions
FSTRING_START = getattr(tokenize, "FSTRING_START", None)
FSTRING_MIDDLE = getattr(tokenize, "FSTRING_MIDDLE", None)
FSTRING_END = getattr(tokenize, "FSTRING_END", None)
def _strip_comment_tags(comments: MutableSequence[str], tags: Iterable[str]):
"""Helper function for `extract` that strips comment tags from strings
in a list of comment lines. This functions operates in-place.
"""
def _strip(line: str):
for tag in tags:
if line.startswith(tag):
return line[len(tag):].strip()
return line
comments[:] = map(_strip, comments)
def default_directory_filter(dirpath: str | os.PathLike[str]) -> bool:
subdir = os.path.basename(dirpath)
# Legacy default behavior: ignore dot and underscore directories
return not (subdir.startswith('.') or subdir.startswith('_'))
def extract_from_dir(
dirname: str | os.PathLike[str] | None = None,
method_map: Iterable[tuple[str, str]] = DEFAULT_MAPPING,
options_map: SupportsItems[str, dict[str, Any]] | None = None,
keywords: Mapping[str, _Keyword] = DEFAULT_KEYWORDS,
comment_tags: Collection[str] = (),
callback: Callable[[str, str, dict[str, Any]], object] | None = None,
strip_comment_tags: bool = False,
directory_filter: Callable[[str], bool] | None = None,
) -> Generator[_FileExtractionResult, None, None]:
"""Extract messages from any source files found in the given directory.
This function generates tuples of the form ``(filename, lineno, message,
comments, context)``.
Which extraction method is used per file is determined by the `method_map`
parameter, which maps extended glob patterns to extraction method names.
For example, the following is the default mapping:
>>> method_map = [
... ('**.py', 'python')
... ]
This basically says that files with the filename extension ".py" at any
level inside the directory should be processed by the "python" extraction
method. Files that don't match any of the mapping patterns are ignored. See
the documentation of the `pathmatch` function for details on the pattern
syntax.
The following extended mapping would also use the "genshi" extraction
method on any file in "templates" subdirectory:
>>> method_map = [
... ('**/templates/**.*', 'genshi'),
... ('**.py', 'python')
... ]
The dictionary provided by the optional `options_map` parameter augments
these mappings. It uses extended glob patterns as keys, and the values are
dictionaries mapping options names to option values (both strings).
The glob patterns of the `options_map` do not necessarily need to be the
same as those used in the method mapping. For example, while all files in
the ``templates`` folders in an application may be Genshi applications, the
options for those files may differ based on extension:
>>> options_map = {
... '**/templates/**.txt': {
... 'template_class': 'genshi.template:TextTemplate',
... 'encoding': 'latin-1'
... },
... '**/templates/**.html': {
... 'include_attrs': ''
... }
... }
:param dirname: the path to the directory to extract messages from. If
not given the current working directory is used.
:param method_map: a list of ``(pattern, method)`` tuples that maps of
extraction method names to extended glob patterns
:param options_map: a dictionary of additional options (optional)
:param keywords: a dictionary mapping keywords (i.e. names of functions
that should be recognized as translation functions) to
tuples that specify which of their arguments contain
localizable strings
:param comment_tags: a list of tags of translator comments to search for
and include in the results
:param callback: a function that is called for every file that message are
extracted from, just before the extraction itself is
performed; the function is passed the filename, the name
of the extraction method and and the options dictionary as
positional arguments, in that order
:param strip_comment_tags: a flag that if set to `True` causes all comment
tags to be removed from the collected comments.
:param directory_filter: a callback to determine whether a directory should
be recursed into. Receives the full directory path;
should return True if the directory is valid.
:see: `pathmatch`
"""
if dirname is None:
dirname = os.getcwd()
if options_map is None:
options_map = {}
if directory_filter is None:
directory_filter = default_directory_filter
absname = os.path.abspath(dirname)
for root, dirnames, filenames in os.walk(absname):
dirnames[:] = [
subdir for subdir in dirnames
if directory_filter(os.path.join(root, subdir))
]
dirnames.sort()
filenames.sort()
for filename in filenames:
filepath = os.path.join(root, filename).replace(os.sep, '/')
yield from check_and_call_extract_file(
filepath,
method_map,
options_map,
callback,
keywords,
comment_tags,
strip_comment_tags,
dirpath=absname,
)
def check_and_call_extract_file(
filepath: str | os.PathLike[str],
method_map: Iterable[tuple[str, str]],
options_map: SupportsItems[str, dict[str, Any]],
callback: Callable[[str, str, dict[str, Any]], object] | None,
keywords: Mapping[str, _Keyword],
comment_tags: Collection[str],
strip_comment_tags: bool,
dirpath: str | os.PathLike[str] | None = None,
) -> Generator[_FileExtractionResult, None, None]:
"""Checks if the given file matches an extraction method mapping, and if so, calls extract_from_file.
Note that the extraction method mappings are based relative to dirpath.
So, given an absolute path to a file `filepath`, we want to check using
just the relative path from `dirpath` to `filepath`.
Yields 5-tuples (filename, lineno, messages, comments, context).
:param filepath: An absolute path to a file that exists.
:param method_map: a list of ``(pattern, method)`` tuples that maps of
extraction method names to extended glob patterns
:param options_map: a dictionary of additional options (optional)
:param callback: a function that is called for every file that message are
extracted from, just before the extraction itself is
performed; the function is passed the filename, the name
of the extraction method and and the options dictionary as
positional arguments, in that order
:param keywords: a dictionary mapping keywords (i.e. names of functions
that should be recognized as translation functions) to
tuples that specify which of their arguments contain
localizable strings
:param comment_tags: a list of tags of translator comments to search for
and include in the results
:param strip_comment_tags: a flag that if set to `True` causes all comment
tags to be removed from the collected comments.
:param dirpath: the path to the directory to extract messages from.
:return: iterable of 5-tuples (filename, lineno, messages, comments, context)
:rtype: Iterable[tuple[str, int, str|tuple[str], list[str], str|None]
"""
# filename is the relative path from dirpath to the actual file
filename = relpath(filepath, dirpath)
for pattern, method in method_map:
if not pathmatch(pattern, filename):
continue
options = {}
for opattern, odict in options_map.items():
if pathmatch(opattern, filename):
options = odict
if callback:
callback(filename, method, options)
for message_tuple in extract_from_file(
method, filepath,
keywords=keywords,
comment_tags=comment_tags,
options=options,
strip_comment_tags=strip_comment_tags,
):
yield (filename, *message_tuple)
break
def extract_from_file(
method: _ExtractionMethod,
filename: str | os.PathLike[str],
keywords: Mapping[str, _Keyword] = DEFAULT_KEYWORDS,
comment_tags: Collection[str] = (),
options: Mapping[str, Any] | None = None,
strip_comment_tags: bool = False,
) -> list[_ExtractionResult]:
"""Extract messages from a specific file.
This function returns a list of tuples of the form ``(lineno, message, comments, context)``.
:param filename: the path to the file to extract messages from
:param method: a string specifying the extraction method (.e.g. "python")
:param keywords: a dictionary mapping keywords (i.e. names of functions
that should be recognized as translation functions) to
tuples that specify which of their arguments contain
localizable strings
:param comment_tags: a list of translator tags to search for and include
in the results
:param strip_comment_tags: a flag that if set to `True` causes all comment
tags to be removed from the collected comments.
:param options: a dictionary of additional options (optional)
:returns: list of tuples of the form ``(lineno, message, comments, context)``
:rtype: list[tuple[int, str|tuple[str], list[str], str|None]
"""
if method == 'ignore':
return []
with open(filename, 'rb') as fileobj:
return list(extract(method, fileobj, keywords, comment_tags,
options, strip_comment_tags))
def _match_messages_against_spec(lineno: int, messages: list[str|None], comments: list[str],
fileobj: _FileObj, spec: tuple[int|tuple[int, str], ...]):
translatable = []
context = None
# last_index is 1 based like the keyword spec
last_index = len(messages)
for index in spec:
if isinstance(index, tuple): # (n, 'c')
context = messages[index[0] - 1]
continue
if last_index < index:
# Not enough arguments
return
message = messages[index - 1]
if message is None:
return
translatable.append(message)
# keyword spec indexes are 1 based, therefore '-1'
if isinstance(spec[0], tuple):
# context-aware *gettext method
first_msg_index = spec[1] - 1
else:
first_msg_index = spec[0] - 1
# An empty string msgid isn't valid, emit a warning
if not messages[first_msg_index]:
filename = (getattr(fileobj, "name", None) or "(unknown)")
sys.stderr.write(
f"{filename}:{lineno}: warning: Empty msgid. It is reserved by GNU gettext: gettext(\"\") "
f"returns the header entry with meta information, not the empty string.\n",
)
return
translatable = tuple(translatable)
if len(translatable) == 1:
translatable = translatable[0]
return lineno, translatable, comments, context
@lru_cache(maxsize=None)
def _find_extractor(name: str):
for ep_name, load in find_entrypoints(GROUP_NAME):
if ep_name == name:
return load()
return None
def extract(
method: _ExtractionMethod,
fileobj: _FileObj,
keywords: Mapping[str, _Keyword] = DEFAULT_KEYWORDS,
comment_tags: Collection[str] = (),
options: Mapping[str, Any] | None = None,
strip_comment_tags: bool = False,
) -> Generator[_ExtractionResult, None, None]:
"""Extract messages from the given file-like object using the specified
extraction method.
This function returns tuples of the form ``(lineno, message, comments, context)``.
The implementation dispatches the actual extraction to plugins, based on the
value of the ``method`` parameter.
>>> source = b'''# foo module
... def run(argv):
... print(_('Hello, world!'))
... '''
>>> from io import BytesIO
>>> for message in extract('python', BytesIO(source)):
... print(message)
(3, u'Hello, world!', [], None)
:param method: an extraction method (a callable), or
a string specifying the extraction method (.e.g. "python");
if this is a simple name, the extraction function will be
looked up by entry point; if it is an explicit reference
to a function (of the form ``package.module:funcname`` or
``package.module.funcname``), the corresponding function
will be imported and used
:param fileobj: the file-like object the messages should be extracted from
:param keywords: a dictionary mapping keywords (i.e. names of functions
that should be recognized as translation functions) to
tuples that specify which of their arguments contain
localizable strings
:param comment_tags: a list of translator tags to search for and include
in the results
:param options: a dictionary of additional options (optional)
:param strip_comment_tags: a flag that if set to `True` causes all comment
tags to be removed from the collected comments.
:raise ValueError: if the extraction method is not registered
:returns: iterable of tuples of the form ``(lineno, message, comments, context)``
:rtype: Iterable[tuple[int, str|tuple[str], list[str], str|None]
"""
func = None
if callable(method):
func = method
elif ':' in method or '.' in method:
if ':' not in method:
lastdot = method.rfind('.')
module, attrname = method[:lastdot], method[lastdot + 1:]
else:
module, attrname = method.split(':', 1)
func = getattr(__import__(module, {}, {}, [attrname]), attrname)
else:
func = _find_extractor(method)
if func is None:
# if no named entry point was found,
# we resort to looking up a builtin extractor
func = _BUILTIN_EXTRACTORS.get(method)
if func is None:
raise ValueError(f"Unknown extraction method {method!r}")
results = func(fileobj, keywords.keys(), comment_tags,
options=options or {})
for lineno, funcname, messages, comments in results:
if not isinstance(messages, (list, tuple)):
messages = [messages]
if not messages:
continue
specs = keywords[funcname] or None if funcname else None
# {None: x} may be collapsed into x for backwards compatibility.
if not isinstance(specs, dict):
specs = {None: specs}
if strip_comment_tags:
_strip_comment_tags(comments, comment_tags)
# None matches all arities.
for arity in (None, len(messages)):
try:
spec = specs[arity]
except KeyError:
continue
if spec is None:
spec = (1,)
result = _match_messages_against_spec(lineno, messages, comments, fileobj, spec)
if result is not None:
yield result
def extract_nothing(
fileobj: _FileObj,
keywords: Mapping[str, _Keyword],
comment_tags: Collection[str],
options: Mapping[str, Any],
) -> list[_ExtractionResult]:
"""Pseudo extractor that does not actually extract anything, but simply
returns an empty list.
"""
return []
def extract_python(
fileobj: IO[bytes],
keywords: Mapping[str, _Keyword],
comment_tags: Collection[str],
options: _PyOptions,
) -> Generator[_ExtractionResult, None, None]:
"""Extract messages from Python source code.
It returns an iterator yielding tuples in the following form ``(lineno,
funcname, message, comments)``.
:param fileobj: the seekable, file-like object the messages should be
extracted from
:param keywords: a list of keywords (i.e. function names) that should be
recognized as translation functions
:param comment_tags: a list of translator tags to search for and include
in the results
:param options: a dictionary of additional options (optional)
:rtype: ``iterator``
"""
funcname = lineno = message_lineno = None
call_stack = -1
buf = []
messages = []
translator_comments = []
in_def = in_translator_comments = False
comment_tag = None
encoding = parse_encoding(fileobj) or options.get('encoding', 'UTF-8')
future_flags = parse_future_flags(fileobj, encoding)
next_line = lambda: fileobj.readline().decode(encoding)
tokens = generate_tokens(next_line)
# Current prefix of a Python 3.12 (PEP 701) f-string, or None if we're not
# currently parsing one.
current_fstring_start = None
for tok, value, (lineno, _), _, _ in tokens:
if call_stack == -1 and tok == NAME and value in ('def', 'class'):
in_def = True
elif tok == OP and value == '(':
if in_def:
# Avoid false positives for declarations such as:
# def gettext(arg='message'):
in_def = False
continue
if funcname:
message_lineno = lineno
call_stack += 1
elif in_def and tok == OP and value == ':':
# End of a class definition without parens
in_def = False
continue
elif call_stack == -1 and tok == COMMENT:
# Strip the comment token from the line
value = value[1:].strip()
if in_translator_comments and \
translator_comments[-1][0] == lineno - 1:
# We're already inside a translator comment, continue appending
translator_comments.append((lineno, value))
continue
# If execution reaches this point, let's see if comment line
# starts with one of the comment tags
for comment_tag in comment_tags:
if value.startswith(comment_tag):
in_translator_comments = True
translator_comments.append((lineno, value))
break
elif funcname and call_stack == 0:
nested = (tok == NAME and value in keywords)
if (tok == OP and value == ')') or nested:
if buf:
messages.append(''.join(buf))
del buf[:]
else:
messages.append(None)
messages = tuple(messages) if len(messages) > 1 else messages[0]
# Comments don't apply unless they immediately
# precede the message
if translator_comments and \
translator_comments[-1][0] < message_lineno - 1:
translator_comments = []
yield (message_lineno, funcname, messages,
[comment[1] for comment in translator_comments])
funcname = lineno = message_lineno = None
call_stack = -1
messages = []
translator_comments = []
in_translator_comments = False
if nested:
funcname = value
elif tok == STRING:
val = _parse_python_string(value, encoding, future_flags)
if val is not None:
buf.append(val)
# Python 3.12+, see https://peps.python.org/pep-0701/#new-tokens
elif tok == FSTRING_START:
current_fstring_start = value
elif tok == FSTRING_MIDDLE:
if current_fstring_start is not None:
current_fstring_start += value
elif tok == FSTRING_END:
if current_fstring_start is not None:
fstring = current_fstring_start + value
val = _parse_python_string(fstring, encoding, future_flags)
if val is not None:
buf.append(val)
elif tok == OP and value == ',':
if buf:
messages.append(''.join(buf))
del buf[:]
else:
messages.append(None)
if translator_comments:
# We have translator comments, and since we're on a
# comma(,) user is allowed to break into a new line
# Let's increase the last comment's lineno in order
# for the comment to still be a valid one
old_lineno, old_comment = translator_comments.pop()
translator_comments.append((old_lineno + 1, old_comment))
elif call_stack > 0 and tok == OP and value == ')':
call_stack -= 1
elif funcname and call_stack == -1:
funcname = None
elif tok == NAME and value in keywords:
funcname = value
if (current_fstring_start is not None
and tok not in {FSTRING_START, FSTRING_MIDDLE}
):
# In Python 3.12, tokens other than FSTRING_* mean the
# f-string is dynamic, so we don't wan't to extract it.
# And if it's FSTRING_END, we've already handled it above.
# Let's forget that we're in an f-string.
current_fstring_start = None
def _parse_python_string(value: str, encoding: str, future_flags: int) -> str | None:
# Unwrap quotes in a safe manner, maintaining the string's encoding
# https://sourceforge.net/tracker/?func=detail&atid=355470&aid=617979&group_id=5470
code = compile(
f'# coding={str(encoding)}\n{value}',
'<string>',
'eval',
ast.PyCF_ONLY_AST | future_flags,
)
if isinstance(code, ast.Expression):
body = code.body
if isinstance(body, ast.Constant):
return body.value
if isinstance(body, ast.JoinedStr): # f-string
if all(isinstance(node, ast.Constant) for node in body.values):
return ''.join(node.value for node in body.values)
# TODO: we could raise an error or warning when not all nodes are constants
return None
def extract_javascript(
fileobj: _FileObj,
keywords: Mapping[str, _Keyword],
comment_tags: Collection[str],
options: _JSOptions,
lineno: int = 1,
) -> Generator[_ExtractionResult, None, None]:
"""Extract messages from JavaScript source code.
:param fileobj: the seekable, file-like object the messages should be
extracted from
:param keywords: a list of keywords (i.e. function names) that should be
recognized as translation functions
:param comment_tags: a list of translator tags to search for and include
in the results
:param options: a dictionary of additional options (optional)
Supported options are:
* `jsx` -- set to false to disable JSX/E4X support.
* `template_string` -- if `True`, supports gettext(`key`)
* `parse_template_string` -- if `True` will parse the
contents of javascript
template strings.
:param lineno: line number offset (for parsing embedded fragments)
"""
from babel.messages.jslexer import Token, tokenize, unquote_string
funcname = message_lineno = None
messages = []
last_argument = None
translator_comments = []
concatenate_next = False
encoding = options.get('encoding', 'utf-8')
last_token = None
call_stack = -1
dotted = any('.' in kw for kw in keywords)
for token in tokenize(
fileobj.read().decode(encoding),
jsx=options.get("jsx", True),
template_string=options.get("template_string", True),
dotted=dotted,
lineno=lineno,
):
if ( # Turn keyword`foo` expressions into keyword("foo") calls:
funcname and # have a keyword...
(last_token and last_token.type == 'name') and # we've seen nothing after the keyword...
token.type == 'template_string' # this is a template string
):
message_lineno = token.lineno
messages = [unquote_string(token.value)]
call_stack = 0
token = Token('operator', ')', token.lineno)
if options.get('parse_template_string') and not funcname and token.type == 'template_string':
yield from parse_template_string(token.value, keywords, comment_tags, options, token.lineno)
elif token.type == 'operator' and token.value == '(':
if funcname:
message_lineno = token.lineno
call_stack += 1
elif call_stack == -1 and token.type == 'linecomment':
value = token.value[2:].strip()
if translator_comments and \
translator_comments[-1][0] == token.lineno - 1:
translator_comments.append((token.lineno, value))
continue
for comment_tag in comment_tags:
if value.startswith(comment_tag):
translator_comments.append((token.lineno, value.strip()))
break
elif token.type == 'multilinecomment':
# only one multi-line comment may precede a translation
translator_comments = []
value = token.value[2:-2].strip()
for comment_tag in comment_tags:
if value.startswith(comment_tag):
lines = value.splitlines()
if lines:
lines[0] = lines[0].strip()
lines[1:] = dedent('\n'.join(lines[1:])).splitlines()
for offset, line in enumerate(lines):
translator_comments.append((token.lineno + offset,
line))
break
elif funcname and call_stack == 0:
if token.type == 'operator' and token.value == ')':
if last_argument is not None:
messages.append(last_argument)
if len(messages) > 1:
messages = tuple(messages)
elif messages:
messages = messages[0]
else:
messages = None
# Comments don't apply unless they immediately precede the
# message
if translator_comments and \
translator_comments[-1][0] < message_lineno - 1:
translator_comments = []
if messages is not None:
yield (message_lineno, funcname, messages,
[comment[1] for comment in translator_comments])
funcname = message_lineno = last_argument = None
concatenate_next = False
translator_comments = []
messages = []
call_stack = -1
elif token.type in ('string', 'template_string'):
new_value = unquote_string(token.value)
if concatenate_next:
last_argument = (last_argument or '') + new_value
concatenate_next = False
else:
last_argument = new_value
elif token.type == 'operator':
if token.value == ',':
if last_argument is not None:
messages.append(last_argument)
last_argument = None
else:
messages.append(None)
concatenate_next = False
elif token.value == '+':
concatenate_next = True
elif call_stack > 0 and token.type == 'operator' \
and token.value == ')':
call_stack -= 1
elif funcname and call_stack == -1:
funcname = None
elif call_stack == -1 and token.type == 'name' and \
token.value in keywords and \
(last_token is None or last_token.type != 'name' or
last_token.value != 'function'):
funcname = token.value
last_token = token
def parse_template_string(
template_string: str,
keywords: Mapping[str, _Keyword],
comment_tags: Collection[str],
options: _JSOptions,
lineno: int = 1,
) -> Generator[_ExtractionResult, None, None]:
"""Parse JavaScript template string.
:param template_string: the template string to be parsed
:param keywords: a list of keywords (i.e. function names) that should be
recognized as translation functions
:param comment_tags: a list of translator tags to search for and include
in the results
:param options: a dictionary of additional options (optional)
:param lineno: starting line number (optional)
"""
from babel.messages.jslexer import line_re
prev_character = None
level = 0
inside_str = False
expression_contents = ''
for character in template_string[1:-1]:
if not inside_str and character in ('"', "'", '`'):
inside_str = character
elif inside_str == character and prev_character != r'\\':
inside_str = False
if level:
expression_contents += character
if not inside_str:
if character == '{' and prev_character == '$':
level += 1
elif level and character == '}':
level -= 1
if level == 0 and expression_contents:
expression_contents = expression_contents[0:-1]
fake_file_obj = io.BytesIO(expression_contents.encode())
yield from extract_javascript(fake_file_obj, keywords, comment_tags, options, lineno)
lineno += len(line_re.findall(expression_contents))
expression_contents = ''
prev_character = character
_BUILTIN_EXTRACTORS = {
'ignore': extract_nothing,
'python': extract_python,
'javascript': extract_javascript,
}