cs-lex 20250914

Lexical analysis functions, tokenisers, transcribers: an arbitrary assortment of lexical and tokenisation functions useful for writing recursive descent parsers, of which I have several. There are also some transcription functions for producing text from various objects, such as `hexify` and `unctrl`.

Lexical analysis functions, tokenisers, transcribers: an arbitrary assortment of lexical and tokenisation functions useful for writing recursive descent parsers, of which I have several. There are also some transcription functions for producing text from various objects, such as hexify and unctrl.

Latest release 20250914:

• New single_space(s[,sep=]) function to split a string and rejoin, with a single space by default.
• Obsolete the @has_format_attributes decorator, how covered by FormatableMixin automatically.
• New FormatMapping(obj,base_format_mapping[,missing]) class presenting a mapping where some entries are callable, used by FormatableMixin.format_as.
• FStr: rename lc to lc_ to reflect the use of the lc_() function.
• format_as: new optional missing(mapping,key) parameter to supply values for keys not in format_mapping.
• FormatAsError: annotate the error with .key, .format_s, .format_mapping.
• Many small internal changes.

Generally the get_* functions accept a source string and an offset (usually optional, default 0) and return a token and the new offset, raising ValueError on failed tokenisation.

Short summary:

• as_lines: Generator yielding complete lines from arbitrary pieces of text from the iterable of str chunks.
• BaseToken: A mixin for token dataclasses.
• camelcase: Convert a snake cased string snakecased into camel case.
• common_prefix: Return the common prefix of the strings strs.
• common_suffix: Return the common suffix of the strings strs.
• CoreTokens: A mixin for token dataclasses whose subclasses include Identifier, 'NumericValueandQuotedString`.
• cropped: If the length of s exceeds max_length (default 32), replace enough of the tail with ellipsis and the last roffset (default 1) characters of s to fit in max_length characters.
• cropped_repr: Compute a cropped repr() of obj.
• cutprefix: Remove prefix from the front of s if present. Return the suffix if s.startswith(prefix), else s. As with str.startswith, prefix may be a str or a tuple of str. If a tuple, the first matching prefix from the tuple will be removed.
• cutsuffix: Remove suffix from the end of s if present. Return the prefix if s.endswith(suffix), else s. As with str.endswith, suffix may be a str or a tuple of str. If a tuple, the first matching suffix from the tuple will be removed.
• FFloat: Formattable float.
• FInt: Formattable int.
• FNumericMixin: A FormatableMixin subclass.
• format_as: Format the string format_s using Formatter.vformat, return the formatted result. This is a wrapper for str.format_map which raises a more informative FormatAsError exception on failure.
• format_attribute: A decorator to mark a method as available as a format method. This sets method.is_format_attribute=True.
• format_escape: Escape {} characters in a string to protect them from str.format.
• format_recover: Decorator for __format__ methods which replaces failed formats with {self:format_spec}.
• FormatableFormatter: A string.Formatter subclass interacting with objects which inherit from FormatableMixin.
• FormatableMixin: A subclass of FormatableFormatter which provides 2 main features: - a __format__ method which parses the format_spec string into multiple colon separated terms whose results chain - a format_as method which formats a format string using str.format_map with a suitable mapping derived from the instance via its format_kwargs method (whose default is to return the instance itself).
• FormatAsError: Subclass of LookupError for use by format_as.
• FormatMapping: A Mapping subclass based on an object and a mapping intended for use by the FormatableMixin.format_as method. The mapping maps field names to values, where the values may be callables accepting an object. Fetching a value from the mapping will call value(obj) if the value is callable. Some additonal extra field names are provided if not already present in the mapping: - self: the object.
• FStr: A str subclass with the FormatableMixin methods, particularly its __format__ method which uses str method names as valid formats.
• get_chars: Scan the string s for characters in gochars starting at offset. Return (match,new_offset).
• get_decimal: Scan the string s for decimal characters starting at offset (default 0). Return (dec_string,new_offset).
• get_decimal_or_float_value: Fetch a decimal or basic float (nnn.nnn) value from the str s at offset (default 0). Return (value,new_offset).
• get_decimal_value: Scan the string s for a decimal value starting at offset (default 0). Return (value,new_offset).
• get_delimited: Collect text from the string s from position offset up to the first occurence of delimiter delim; return the text excluding the delimiter and the offset after the delimiter.
• get_dotted_identifier: Scan the string s for a dotted identifier (by default an ASCII letter or underscore followed by letters, digits or underscores) with optional trailing dot and another dotted identifier, starting at offset (default 0). Return (match,new_offset).
• get_envvar: Parse a simple environment variable reference to $varname or $x where "x" is a special character.
• get_hexadecimal: Scan the string s for hexadecimal characters starting at offset (default 0). Return (hex_string,new_offset).
• get_hexadecimal_value: Scan the string s for a hexadecimal value starting at offset (default 0). Return (value,new_offset).
• get_identifier: Scan the string s for an identifier (by default an ASCII letter or underscore followed by letters, digits or underscores) starting at offset (default 0). Return (match,new_offset).
• get_ini_clause_entryname: Parse a [clausename]entryname string from s at offset (default 0). Return (clausename,entryname,new_offset).
• get_ini_clausename: Parse a [clausename] string from s at offset (default 0). Return (clausename,new_offset).
• get_nonwhite: Scan the string s for characters not in string.whitespace starting at offset (default 0). Return (match,new_offset).
• get_other_chars: Scan the string s for characters not in stopchars starting at offset (default 0). Return (match,new_offset).
• get_prefix_n: Strip a leading prefix and numeric value n from the string s starting at offset (default 0). Return the matched prefix, the numeric value and the new offset. Returns (None,None,offset) on no match.
• get_qstr: Get quoted text with slosh escapes and optional environment substitution.
• get_qstr_or_identifier: Parse a double quoted string or an identifier.
• get_sloshed_text: Collect slosh escaped text from the string s from position offset (default 0) and return the decoded unicode string and the offset of the completed parse.
• get_suffix_part: Strip a trailing "part N" suffix from the string s. Return the matched suffix and the number part number. Retrn (None,None) on no match.
• get_tokens: Parse the string s from position offset using the supplied tokeniser functions getters. Return the list of tokens matched and the final offset.
• get_uc_identifier: Scan the string s for an identifier as for get_identifier, but require the letters to be uppercase.
• get_white: Scan the string s for characters in string.whitespace starting at offset (default 0). Return (match,new_offset).
• has_format_attributes: OBSOLETE version of has_format_attributes, suggestion: @has_format_attributes is no longer needed.
• hexify: A flavour of binascii.hexlify returning a str.
• htmlify: Convert a string for safe transcription in HTML.
• htmlquote: Quote a string for use in HTML.
• Identifier: A dotted identifier.
• indent: Return the paragraph indented by line_indent (default " ").
• is_dotted_identifier: Test if the string s is an identifier from position offset onward.
• is_identifier: Test if the string s is an identifier from position offset (default 0) onward.
• is_uc_identifier: Test if the string s is an uppercase identifier from position offset (default 0) onward.
• isUC_: Check that a string matches the regular expression ^[A-Z][A-Z_0-9]*$.
• jsquote: Quote a string for use in JavaScript.
• lc_: Return value.lower() with '-' translated into '_' and ' ' translated into '-'.
• match_tokens: Wrapper for get_tokens which catches ValueError exceptions and returns (None,offset).
• NumericValue: An int or float literal.
• parseUC_sAttr: Take an attribute name attr and return (key,is_plural).
• phpquote: Quote a string for use in PHP code.
• printt: A wrapper for tabulate() to print the results. Each positional argument is a table row.
• QuotedString: A double quoted string.
• r: Like typed_str but using repr instead of str. This is available as both typed_repr and r.
• s: Return "type(obj).name:str(obj)" for some object obj. This is available as both typed_str and s.
• single_space: Return the string s stripped and with internal whitespace replaced by sep (default " ").
• skipwhite: Convenience routine for skipping past whitespace; returns the offset of the next nonwhitespace character.
• slosh_mapper: Return a string to replace backslash-c, or None.
• slosh_quote: Quote a string raw_s with quote character q.
• snakecase: Convert a camel cased string camelcased into snake case.
• split_remote_path: OBSOLETE version of split_remote_path, suggestion: cs.fs.RemotePath.from_str.
• stripped_dedent: Slightly smarter dedent which ignores a string's opening indent.
• strlist: Convert an iterable to strings and join with sep (default ', ').
• tabpadding: Compute some spaces to use a tab padding at an offfset.
• tabulate: A generator yielding lines of values from rows aligned in columns.
• texthexify: Transcribe the bytes bs to text using compact text runs for some common text values.
• titleify_lc: Translate '-' into ' ' and '_' translated into '-', then titlecased.
• typed_repr: Like typed_str but using repr instead of str. This is available as both typed_repr and r.
• typed_str: Return "type(obj).name:str(obj)" for some object obj. This is available as both typed_str and s.
• unctrl: Return the string s with TABs expanded and control characters replaced with printable representations.
• untexthexify: Decode a textual representation of binary data into binary data.

Module contents:

• as_lines(chunks, partials=None): Generator yielding complete lines from arbitrary pieces of text from the iterable of str chunks.

  After completion, any remaining newline-free chunks remain in the partials list; they will be unavailable to the caller unless the list is presupplied.

• class BaseToken(cs.deco.Promotable): A mixin for token dataclasses.

  Presently I use this in cs.app.tagger.rules and cs.app.pilfer.parse.

BaseToken.__post_init__(self): An omitted offset,end_offset means the token is the whole source_text.

BaseToken.from_str(text: str) -> 'BaseToken': Parse test as a token of type cls, return the token. Raises SyntaxError on a parse failure. This is a wrapper for the parse class method.

BaseToken.matched_text: The text from self.source_text which matches this token.

BaseToken.parse(text: str, offset: int = 0, *, skip=False) -> Tuple[ForwardRef('BaseToken'), int]: Parse a token from test at offset (default 0). Return a BaseToken subclass instance. Raise SyntaxError if no subclass parses it. Raise EOFError if at the end of the text, checked after any whitespace if skip is true. The returned token's .end_offset is the next parse point.

This base class method attempts the .parse method of all the public subclasses.

Parameters:

• text: the text being parsed
• offset: the offset within the text of the the parse cursor
• skip: if true (default False), skip any leading whitespace before matching

BaseToken.scan(text: str, offset: int = 0, *, skip=True) -> Iterable[ForwardRef('BaseToken')]: Scan text, parsing tokens using BaseToken.parse and yielding them. Parameters are as for BaseToken.parse except as follows:

• encountering end of text end the iteration instead of raising EOFError
• skip defaults to True to allow whitespace between tokens

BaseToken.token_classes(): Return the baseToken subclasses to consider when parsing a token stream.

• camelcase(snakecased: str, first_letter_only: bool = False) -> str: Convert a snake cased string snakecased into camel case.

  Parameters:

  • snakecased: the snake case string to convert
  • first_letter_only: optional flag (default False); if true then just ensure that the first character of a word is uppercased, otherwise use str.title

  Example:

  >>> camelcase('abc_def')
  'abcDef'
  >>> camelcase('ABc_def')
  'abcDef'
  >>> camelcase('abc_dEf')
  'abcDef'
  >>> camelcase('abc_dEf', first_letter_only=True)
  'abcDEf'
  

• common_prefix(*strs): Return the common prefix of the strings strs.

  Examples:

  >>> common_prefix('abc', 'def')
  ''
  >>> common_prefix('abc', 'abd')
  'ab'
  >>> common_prefix('abc', 'abcdef')
  'abc'
  >>> common_prefix('abc', 'abcdef', 'abz')
  'ab'
  >>> # contrast with cs.fileutils.common_path_prefix
  >>> common_prefix('abc/def', 'abc/def1', 'abc/def2')
  'abc/def'
  

• common_suffix(*strs): Return the common suffix of the strings strs.

• class CoreTokens(BaseToken): A mixin for token dataclasses whose subclasses include Identifier, 'NumericValueandQuotedString`.

• cropped(s: str, max_length: int = 32, roffset: int = 1, ellipsis: str = '...'): If the length of s exceeds max_length (default 32), replace enough of the tail with ellipsis and the last roffset (default 1) characters of s to fit in max_length characters.

• cropped_repr(obj, roffset=1, max_length=32, inner_max_length=None): Compute a cropped repr() of obj.

  Parameters:

  • obj: the object to represent
  • max_length: the maximum length of the representation, default 32
  • inner_max_length: the maximum length of the representations of members of obj, default max_length//2
  • roffset: the number of trailing characters to preserve, default 1

• cutprefix(s, prefix): Remove prefix from the front of s if present. Return the suffix if s.startswith(prefix), else s. As with str.startswith, prefix may be a str or a tuple of str. If a tuple, the first matching prefix from the tuple will be removed.

  Example:

  >>> abc_def = 'abc.def'
  >>> cutprefix(abc_def, 'abc.')
  'def'
  >>> cutprefix(abc_def, 'zzz.')
  'abc.def'
  >>> cutprefix(abc_def, '.zzz') is abc_def
  True
  >>> cutprefix('this_that', ('this', 'thusly'))
  '_that'
  >>> cutprefix('thusly_that', ('this', 'thusly'))
  '_that'
  

• cutsuffix(s, suffix): Remove suffix from the end of s if present. Return the prefix if s.endswith(suffix), else s. As with str.endswith, suffix may be a str or a tuple of str. If a tuple, the first matching suffix from the tuple will be removed.

  Example:

  >>> abc_def = 'abc.def'
  >>> cutsuffix(abc_def, '.def')
  'abc'
  >>> cutsuffix(abc_def, '.zzz')
  'abc.def'
  >>> cutsuffix(abc_def, '.zzz') is abc_def
  True
  >>> cutsuffix('this_that', ('that', 'tother'))
  'this_'
  >>> cutsuffix('this_tother', ('that', 'tother'))
  'this_'
  

• class FFloat(FNumericMixin, builtins.float): Formattable float.

• class FInt(FNumericMixin, builtins.int): Formattable int.

• class FNumericMixin(FormatableMixin): A FormatableMixin subclass.

FNumericMixin.localtime(self): Treat this as a UNIX timestamp and return a localtime datetime.

FNumericMixin.utctime(self): Treat this as a UNIX timestamp and return a UTC datetime.

• format_as(format_s: str, format_mapping, formatter=None, *, error_sep=None, missing: Optional[Callable[[Mapping, Any], Any]] = None, strict=False): Format the string format_s using Formatter.vformat, return the formatted result. This is a wrapper for str.format_map which raises a more informative FormatAsError exception on failure.

  Parameters:

  • format_s: the format string to use as the template
  • format_mapping: the mapping of available replacement fields
  • formatter: an optional string.Formatter-like instance with a .vformat(format_string,args,kwargs) method, usually a subclass of string.Formatter; if not specified then FormatableFormatter is used
  • error_sep: optional separator for the multipart error message, default from FormatAsError.DEFAULT_SEPARATOR: '; '
  • missing: an optional callable to turn a key missing from format_mapping into a value to interpolate

• format_attribute(method): A decorator to mark a method as available as a format method. This sets method.is_format_attribute=True.

  For example, the FormatableMixin.json method is defined like this:

  @format_attribute
  def json(self):
      return self.FORMAT_JSON_ENCODER.encode(self)
  

  which allows a FormatableMixin subclass instance to be used in a format string like this:

  {instance:json}
  

  to insert a JSON transcription of the instance.

  It is recommended that methods marked with @format_attribute have no side effects and do not modify state, as they are intended for use in ad hoc format strings supplied by an end user.

• format_escape(s): Escape {} characters in a string to protect them from str.format.

• format_recover(*da, **dkw): Decorator for __format__ methods which replaces failed formats with {self:format_spec}.

• class FormatableFormatter(string.Formatter): A string.Formatter subclass interacting with objects which inherit from FormatableMixin.

FormatableFormatter.format_field(value, format_spec: str): Format a value using format_field, returning an FStr (a str subclass with additional format_spec features).

We actually recognise colon separated chains of formats and apply each format to the previously converted value. The final result is promoted to an FStr before return.

FormatableFormatter.get_arg_name(field_name): The default initial arg_name is an identifier.

Returns (prefix,offset), and ('',0) if there is no arg_name.

FormatableFormatter.get_field(self, field_name, args, kwargs): Get the object referenced by the field text field_name. Raises KeyError for an unknown field_name.

FormatableFormatter.get_format_subspecs(format_spec): Parse a format_spec as a sequence of colon separated components, return a list of the components.

FormatableFormatter.get_subfield(value, subfield_text: str): Resolve value against subfield_text, the remaining field text after the term which resolved to value.

For example, a format {name.blah[0]} has the field text name.blah[0]. A get_field implementation might initially resolve name to some value, leaving .blah[0] as the subfield_text. This method supports taking that value and resolving it against the remaining text .blah[0].

For generality, if subfield_text is the empty string value is returned unchanged.

FormatableFormatter.get_value(self, arg_name, args, kwargs): Get the object with index arg_name.

This default implementation returns (kwargs[arg_name],arg_name).

• class FormatableMixin(FormatableFormatter): A subclass of FormatableFormatter which provides 2 main features:

  • a __format__ method which parses the format_spec string into multiple colon separated terms whose results chain
  • a format_as method which formats a format string using str.format_map with a suitable mapping derived from the instance via its format_kwargs method (whose default is to return the instance itself)

  The format_as method is like an inside out str.format or object.__format__ method.

  The str.format method is designed for formatting a string from a variety of other objects supplied in the keyword arguments.

  The object.__format__ method is for filling out a single str.format replacement field from a single object.

  By contrast, format_as is designed to fill out an entire format string from the current object.

  For example, the cs.tagset.TagSet class subclasses FormatableMixin to provide a format_as method whose replacement fields are derived from the tags in the tag set.

  Subclasses wanting to provide additional format_spec terms should:

  • override FormatableFormatter.format_field1 to implement terms with no colons, letting format_field do the split into terms
  • override FormatableFormatter.get_format_subspecs to implement the parse of format_spec into a sequence of terms. This might recognise a special additional syntax and quietly fall back to super().get_format_subspecs if that is not present.

FormatableMixin.__format__(self, format_spec): Format self according to format_spec.

This implementation calls self.format_field. As such, a format_spec is considered a sequence of colon separated terms.

Classes wanting to implement additional format string syntaxes should either:

• override FormatableFormatter.format_field1 to implement terms with no colons, letting format_field1 do the split into terms
• override FormatableFormatter.get_format_subspecs to implement the term parse.

The default implementation of __format1__ just calls super().__format__. Implementations providing specialised formats should implement them in __format1__ with fallback to super().__format1__.

FormatableMixin.__init_subclass__(**kw): Prefill the cls.format_attributes mapping from the superclass and any format attributes of cls.

FormatableMixin.convert_field(self, value, conversion): The default converter for fields calls Formatter.convert_field.

This is a tiny shim to transmute the '' conversion to None which is what Formatter.convert_field expects.

FormatableMixin.convert_via_method_or_attr(self, format_spec) -> Tuple[Any, int]: Apply a method or attribute name based conversion to self where format_spec starts with a method or attribute name. Return (converted,offset) being the converted value and the offset after the method name.

Note that if there is not a leading identifier on format_spec then this method returns (self,0).

The converted value is obtained from getattr(self,name); if this raises an AttributeError a second attempt is made with getattr(FStr(self),attr) if self is not already an FStr (this provides the common utility methods on other types).

If the value is callable but does not have a true .is_format_attribute a TypeError is raised, otherwise the value is called to complete the conversion. (The .is_format_attribute is usually set by decorating a method with the @format_attribute decorator.)

The motivating example was a PurePosixPath, which does not JSON transcribe; this tweak supports both posixpath:basename via the pathlib stuff and posixpath:json via FStr even though a PurePosixPath does not subclass FStr.

FormatableMixin.format_as(self, format_s: str, *, error_sep: Optional[str] = None, missing: Optional[Callable[[Mapping, Any], Any]] = None, strict=True, **format_kwargs_kw): Return the string format_s formatted using the mapping returned by self.format_kwargs(**format_kwargs_kw).

If a class using the mixin has no format_kwargs() method to provide a mapping for str.format_map then the instance itself is used as the mapping.

FormatableMixin.json(self): The value transcribed as compact JSON.

• class FormatAsError(builtins.LookupError): Subclass of LookupError for use by format_as.
• class FormatMapping(collections.abc.Mapping): A Mapping subclass based on an object and a mapping intended for use by the FormatableMixin.format_as method. The mapping maps field names to values, where the values may be callables accepting an object. Fetching a value from the mapping will call value(obj) if the value is callable. Some additonal extra field names are provided if not already present in the mapping:
  • self: the object

FormatMapping.__getitem__(self, field_name: str): Fetch the value for field_name. If the value is callable, call value(self.obj) to get the value.

FormatMapping.items(self): Proxy .items via self.mapping.

FormatMapping.keys(self): Proxy .keys via self.mapping.

• class FStr(FormatableMixin, builtins.str): A str subclass with the FormatableMixin methods, particularly its __format__ method which uses str method names as valid formats.

  It also has a bunch of utility methods which are available as :method in format strings.

FStr.basename(self): Treat as a filesystem path and return the basename.

FStr.dirname(self): Treat as a filesystem path and return the dirname.

FStr.f(self): Parse self as a float.

FStr.i(self, base=10): Parse self as an int.

FStr.lc_(self): Lowercase using lc_().

FStr.path(self): Convert to a native filesystem pathlib.Path.

FStr.posix_path(self): Convert to a Posix filesystem pathlib.Path.

FStr.windows_path(self): Convert to a Windows filesystem pathlib.Path.

• get_chars(s: str, offset: int, gochars: str) -> Tuple[str, int]: Scan the string s for characters in gochars starting at offset. Return (match,new_offset).

  gochars may also be a callable, in which case a character ch is accepted if gochars(ch) is true.

• get_decimal(s: str, offset: int = 0) -> Tuple[str, int]: Scan the string s for decimal characters starting at offset (default 0). Return (dec_string,new_offset).

• get_decimal_or_float_value(s: int, offset: int = 0) -> Tuple[str, int]: Fetch a decimal or basic float (nnn.nnn) value from the str s at offset (default 0). Return (value,new_offset).

• get_decimal_value(s: str, offset: int = 0) -> Tuple[str, int]: Scan the string s for a decimal value starting at offset (default 0). Return (value,new_offset).

• get_delimited(s, offset, delim): Collect text from the string s from position offset up to the first occurence of delimiter delim; return the text excluding the delimiter and the offset after the delimiter.

• get_dotted_identifier(s: str, offset: int = 0, **kw) -> Tuple[str, int]: Scan the string s for a dotted identifier (by default an ASCII letter or underscore followed by letters, digits or underscores) with optional trailing dot and another dotted identifier, starting at offset (default 0). Return (match,new_offset).

  Note: the empty string and an unchanged offset will be returned if there is no leading letter/underscore.

  Keyword arguments are passed to get_identifier (used for each component of the dotted identifier).

• get_envvar(s, offset=0, environ=None, default=None, specials=None): Parse a simple environment variable reference to $varname or $x where "x" is a special character.

  Parameters:

  • s: the string with the variable reference
  • offset: the starting point for the reference
  • default: default value for missing environment variables; if None (the default) a ValueError is raised
  • environ: the environment mapping, default os.environ
  • specials: the mapping of special single character variables

• get_hexadecimal(s: str, offset: int = 0) -> Tuple[str, int]: Scan the string s for hexadecimal characters starting at offset (default 0). Return (hex_string,new_offset).

• get_hexadecimal_value(s: str, offset: int = 0) -> Tuple[str, int]: Scan the string s for a hexadecimal value starting at offset (default 0). Return (value,new_offset).

• get_identifier(s: str, offset: int = 0, *, alpha: str = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ', number: str = '0123456789', extras: str = '_') -> Tuple[str, int]: Scan the string s for an identifier (by default an ASCII letter or underscore followed by letters, digits or underscores) starting at offset (default 0). Return (match,new_offset).

  Note: the empty string and an unchanged offset will be returned if there is no leading letter/underscore.

  Parameters:

  • s: the string to scan
  • offset: the starting offset, default 0.
  • alpha: the characters considered alphabetic, default string.ascii_letters.
  • number: the characters considered numeric, default string.digits.
  • extras: extra characters considered part of an identifier, default '_'.

• get_ini_clause_entryname(s: str, offset: int = 0) -> Tuple[str, str, int]: Parse a [clausename]entryname string from s at offset (default 0). Return (clausename,entryname,new_offset).

• get_ini_clausename(s: str, offset: int = 0) -> Tuple[str, int]: Parse a [clausename] string from s at offset (default 0). Return (clausename,new_offset).

• get_nonwhite(s: str, offset: int = 0) -> Tuple[str, int]: Scan the string s for characters not in string.whitespace starting at offset (default 0). Return (match,new_offset).

• get_other_chars(s: str, offset: int = 0, stopchars: Optional[str] = None) -> Tuple[str, int]: Scan the string s for characters not in stopchars starting at offset (default 0). Return (match,new_offset).

• get_prefix_n(s: str, prefix: str, n: Optional[int] = None, *, offset: int = 0) -> str: Strip a leading prefix and numeric value n from the string s starting at offset (default 0). Return the matched prefix, the numeric value and the new offset. Returns (None,None,offset) on no match.

  Parameters:

  • s: the string to parse
  • prefix: the prefix string which must appear at offset or an object with a match(str,offset) method such as an re.Pattern regexp instance
  • n: optional integer value; if omitted any value will be accepted, otherwise the numeric part must match n

  If prefix is a str, the "matched prefix" return value is prefix. Otherwise the "matched prefix" return value is the result of the prefix.match(s,offset) call. The result must also support a .end() method returning the offset in s beyond the match, used to locate the following numeric portion.

  Examples:

  import re get_prefix_n('s03e01--', 's') ('s', 3, 3) get_prefix_n('s03e01--', 's', 3) ('s', 3, 3) get_prefix_n('s03e01--', 's', 4) (None, None, 0) get_prefix_n('s03e01--', re.compile('[es]',re.I)) (<re.Match object; span=(0, 1), match='s'>, 3, 3) get_prefix_n('s03e01--', re.compile('[es]',re.I), offset=3) (<re.Match object; span=(3, 4), match='e'>, 1, 6)

• get_qstr(s, offset=0, q='"', environ=None, default=None, env_specials=None): Get quoted text with slosh escapes and optional environment substitution.

  Parameters:

  • s: the string containg the quoted text.
  • offset: the starting point, default 0.
  • q: the quote character, default '"'. If q is None, do not expect the string to be delimited by quote marks.
  • environ: if not None, also parse and expand $envvar references.
  • default: passed to get_envvar

• get_qstr_or_identifier(s, offset): Parse a double quoted string or an identifier.

• get_sloshed_text(s, delim, offset=0, slosh='\\', mapper=<function slosh_mapper at 0x107087740>, specials=None): Collect slosh escaped text from the string s from position offset (default 0) and return the decoded unicode string and the offset of the completed parse.

  Parameters:

  • delim: end of string delimiter, such as a single or double quote.
  • offset: starting offset within s, default 0.
  • slosh: escape character, default a slosh ('').
  • mapper: a mapping function which accepts a single character and returns a replacement string or None; this is used the replace things such as '\t' or '\n'. The default is the slosh_mapper function, whose default mapping is SLOSH_CHARMAP.
  • specials: a mapping of other special character sequences and parse functions for gathering them up. When one of the special character sequences is found in the string, the parse function is called to parse at that point. The parse functions accept s and the offset of the special character. They return the decoded string and the offset past the parse.

  The escape character slosh introduces an encoding of some replacement text whose value depends on the following character. If the following character is:

  • the escape character slosh, insert the escape character.
  • the string delimiter delim, insert the delimiter.
  • the character 'x', insert the character with code from the following 2 hexadecimal digits.
  • the character 'u', insert the character with code from the following 4 hexadecimal digits.
  • the character 'U', insert the character with code from the following 8 hexadecimal digits.
  • a character from the keys of mapper

• get_suffix_part(s: str, *, keywords: Iterable[str] = ('part',), numeral_map: Optional[Mapping[str, int]] = None) -> Union[Tuple[str, int], Tuple[NoneType, NoneType]]: Strip a trailing "part N" suffix from the string s. Return the matched suffix and the number part number. Retrn (None,None) on no match.

  Parameters:

  • s: the string
  • keywords: an iterable of str to match, or a single str; default 'part'
  • numeral_map: an optional mapping of numeral names to numeric values; default NUMERAL_NAMES['en'], the English numerals

  Exanmple:

  >>> get_suffix_part('s09e10 - A New World: Part One')
  (': Part One', 1)
  

• get_tokens(s, offset, getters): Parse the string s from position offset using the supplied tokeniser functions getters. Return the list of tokens matched and the final offset.

  Parameters:

  • s: the string to parse.
  • offset: the starting position for the parse.
  • getters: an iterable of tokeniser specifications.

  Each tokeniser specification getter is either:

  • a callable expecting (s,offset) and returning (token,new_offset)
  • a literal string, to be matched exactly
  • a tuple or list with values (func,args,kwargs); call func(s,offset,*args,**kwargs)
  • an object with a .match method such as a regex; call getter.match(s,offset) and return a match object with a .end() method returning the offset of the end of the match

• get_uc_identifier(s: str, offset: int = 0, number: str = '0123456789', extras: str = '_') -> Tuple[str, int]: Scan the string s for an identifier as for get_identifier, but require the letters to be uppercase.

• get_white(s: str, offset: int = 0) -> Tuple[str, int]: Scan the string s for characters in string.whitespace starting at offset (default 0). Return (match,new_offset).

• has_format_attributes(*da, **dkw): OBSOLETE version of has_format_attributes, suggestion: @has_format_attributes is no longer needed

  A obsolete class decorator formerly used to walk the class for @formatmethod methods. This is now done by FormatableMixin.__init_subclass__.

• hexify(bs: bytes) -> str: A flavour of binascii.hexlify returning a str.

• htmlify(s: str, nbsp: bool = False) -> str: Convert a string for safe transcription in HTML.

  Parameters:

  • s: the string
  • nbsp: replaces spaces with " " to prevent word folding, default False.

• htmlquote(s: str) -> str: Quote a string for use in HTML.

• class Identifier(CoreTokens): A dotted identifier.

Identifier.parse(text: str, offset: int = 0, *, skip=False) -> Tuple[str, cs.lex.CoreTokens, int]: Parse a dotted identifier from test.

• indent(paragraph: str, line_indent: str = ' ') -> str: Return the paragraph indented by line_indent (default " ").

• is_dotted_identifier(s: str, offset: int = 0, **kw) -> bool: Test if the string s is an identifier from position offset onward.

• is_identifier(s: str, offset: int = 0, **kw) -> bool: Test if the string s is an identifier from position offset (default 0) onward.

• is_uc_identifier(s: str, offset: int = 0, **kw) -> bool: Test if the string s is an uppercase identifier from position offset (default 0) onward.

• isUC_(s): Check that a string matches the regular expression ^[A-Z][A-Z_0-9]*$.

• jsquote(s: str) -> str: Quote a string for use in JavaScript.

• lc_(value: str) -> str: Return value.lower() with '-' translated into '_' and ' ' translated into '-'.

  I use this to construct lowercase filenames containing a readable transcription of a title string.

  See also titleify_lc(), an imperfect reversal of this.

• match_tokens(s, offset, getters): Wrapper for get_tokens which catches ValueError exceptions and returns (None,offset).

• class NumericValue(_LiteralValue): An int or float literal.

NumericValue.parse(text: str, offset: int = 0, *, skip=False) -> 'NumericValue': Parse a Python style int or float.

• parseUC_sAttr(attr): Take an attribute name attr and return (key,is_plural).

  Examples:

  • 'FOO' returns ('FOO',False).
  • 'FOOs' or 'FOOes' returns ('FOO',True). Otherwise return (None,False).

• phpquote(s: str) -> str: Quote a string for use in PHP code.

• printt(*table, file=None, flush=False, indent='', print_func=None, **tabulate_kw): A wrapper for tabulate() to print the results. Each positional argument is a table row.

  Parameters:

  • file: optional output file, passed to print_func
  • flush: optional flush flag, passed to print_func
  • indent: optional leading indent for the output lines
  • print_func: optional print() function, default builtins.print Other keyword arguments are passed to tabulate().

• class QuotedString(_LiteralValue): A double quoted string.

QuotedString.parse(text: str, offset: int = 0, *, skip=False) -> 'QuotedString': Parse a double quoted string from text.

• r(obj: Any, max_length: Optional[int] = None, *, use_cls: bool = False) -> str: Like typed_str but using repr instead of str. This is available as both typed_repr and r.

• s(obj: Any, use_cls: bool = False, use_repr: bool = False, max_length: int = 32) -> str: Return "type(obj).name:str(obj)" for some object obj. This is available as both typed_str and s.

  Parameters:

  • use_cls: default False; if true, use str(type(obj)) instead of type(obj).__name__
  • use_repr: default False; if true, use repr(obj) instead of str(obj)

  I use this a lot when debugging. Example:

  from cs.lex import typed_str as s
  ......
  X("foo = %s", s(foo))
  

• single_space(s: str, *, sep=' ') -> str: Return the string s stripped and with internal whitespace replaced by sep (default " ").

• skipwhite(s: str, offset: int = 0) -> Tuple[str, int]: Convenience routine for skipping past whitespace; returns the offset of the next nonwhitespace character.

• slosh_mapper(c, charmap=None): Return a string to replace backslash-c, or None.

• slosh_quote(raw_s: str, q: str): Quote a string raw_s with quote character q.

• snakecase(camelcased: str) -> str: Convert a camel cased string camelcased into snake case.

  Parameters:

  • cameelcased: the cameel case string to convert
  • first_letter_only: optional flag (default False); if true then just ensure that the first character of a word is uppercased, otherwise use str.title

  Example:

  >>> snakecase('abcDef')
  'abc_def'
  >>> snakecase('abcDEf')
  'abc_def'
  >>> snakecase('AbcDef')
  'abc_def'
  

• split_remote_path(remotepath: str) -> Tuple[Optional[str], str]: OBSOLETE version of split_remote_path, suggestion: cs.fs.RemotePath.from_str

  Split a path with an optional leading [user@]rhost: prefix into the prefix and the remaining path. None is returned for the prefix is there is none. This is useful for things like rsync targets etc.

  OBSOLETE, use cs.fs.RemotePath.from_str instead.

• stripped_dedent(s: str, post_indent: str = '', sub_indent: str = '') -> str: Slightly smarter dedent which ignores a string's opening indent.

  Algorithm: strip the supplied string s, pull off the leading line, dedent the rest, put back the leading line.

  This is a lot like the inspect.cleandoc() function.

  This supports my preferred docstring layout, where the opening line of text is on the same line as the opening quote.

  The optional post_indent parameter may be used to indent the dedented text before return.

  The optional sub_indent parameter may be used to indent the second and following lines if the dedented text before return.

  Examples:

  >>> def func(s):
  ...   """ Slightly smarter dedent which ignores a string's opening indent.
  ...       Strip the supplied string `s`. Pull off the leading line.
  ...       Dedent the rest. Put back the leading line.
  ...   """
  ...   pass
  ...
  >>> from cs.lex import stripped_dedent
  >>> print(stripped_dedent(func.__doc__))
  Slightly smarter dedent which ignores a string's opening indent.
  Strip the supplied string `s`. Pull off the leading line.
  Dedent the rest. Put back the leading line.
  >>> print(stripped_dedent(func.__doc__, sub_indent='  '))
  Slightly smarter dedent which ignores a string's opening indent.
    Strip the supplied string `s`. Pull off the leading line.
    Dedent the rest. Put back the leading line.
  >>> print(stripped_dedent(func.__doc__, post_indent='  '))
    Slightly smarter dedent which ignores a string's opening indent.
    Strip the supplied string `s`. Pull off the leading line.
    Dedent the rest. Put back the leading line.
  >>> print(stripped_dedent(func.__doc__, post_indent='  ', sub_indent='| '))
    Slightly smarter dedent which ignores a string's opening indent.
    | Strip the supplied string `s`. Pull off the leading line.
    | Dedent the rest. Put back the leading line.
  

• strlist(ary: Iterable, sep: str = ', ') -> str: Convert an iterable to strings and join with sep (default ', ').

• tabpadding(padlen: int, tabsize: int = 8, offset: int = 0) -> str: Compute some spaces to use a tab padding at an offfset.

• tabulate(*rows, sep=' ', ppcls=None): A generator yielding lines of values from rows aligned in columns.

  Each row in rows is a list of strings. Non-str objects are promoted to str via pprint.pformat. If the strings contain newlines they will be split into subrows.

  Example:

  >>> for row in tabulate(
  ...     ['one col'],
  ...     ['three', 'column', 'row'],
  ...     ['row3', 'multi\nline\ntext', 'goes\nhere', 'and\nhere'],
  ...     ['two', 'cols'],
  ... ):
  ...     print(row)
  ...
  one col
  three    column  row
  row3     multi   goes  and
           line    here  here
           text
  two      cols
  >>>
  

• texthexify(bs: bytes, shiftin: str = '[', shiftout: str = ']', whitelist: Optional[str] = None) -> str: Transcribe the bytes bs to text using compact text runs for some common text values.

  This can be reversed with the untexthexify function.

  This is an ad doc format devised to be compact but also to expose "text" embedded within to the eye. The original use case was transcribing a binary directory entry format, where the filename parts would be somewhat visible in the transcription.

  The output is a string of hexadecimal digits for the encoded bytes except for runs of values from the whitelist, which are enclosed in the shiftin and shiftout markers and transcribed as is. The default whitelist is values of the ASCII letters, the decimal digits and the punctuation characters '_-+.,'. The default shiftin and shiftout markers are '[' and ']'.

  String objects converted with either hexify and texthexify output strings may be freely concatenated and decoded with untexthexify.

  Example:

  >>> texthexify(b'&^%&^%abcdefghi)(*)(*')
  '265e25265e25[abcdefghi]29282a29282a'
  

  Parameters:

  • bs: the bytes to transcribe
  • shiftin: Optional. The marker string used to indicate a shift to direct textual transcription of the bytes, default: '['.
  • shiftout: Optional. The marker string used to indicate a shift from text mode back into hexadecimal transcription, default ']'.
  • whitelist: an optional bytes or string object indicating byte values which may be represented directly in text; the default value is the ASCII letters, the decimal digits and the punctuation characters '_-+.,'.

• titleify_lc(value_lc: str) -> str: Translate '-' into ' ' and '_' translated into '-', then titlecased.

  See also lc_(), which this reverses imperfectly.

• typed_repr(obj: Any, max_length: Optional[int] = None, *, use_cls: bool = False) -> str: Like typed_str but using repr instead of str. This is available as both typed_repr and r.

• typed_str(obj: Any, use_cls: bool = False, use_repr: bool = False, max_length: int = 32) -> str: Return "type(obj).name:str(obj)" for some object obj. This is available as both typed_str and s.

  Parameters:

  • use_cls: default False; if true, use str(type(obj)) instead of type(obj).__name__
  • use_repr: default False; if true, use repr(obj) instead of str(obj)

  I use this a lot when debugging. Example:

  from cs.lex import typed_str as s
  ......
  X("foo = %s", s(foo))
  

• unctrl(s: str, tabsize: int = 8) -> str: Return the string s with TABs expanded and control characters replaced with printable representations.

• untexthexify(s: str, shiftin: str = '[', shiftout: str = ']') -> str: Decode a textual representation of binary data into binary data.

  This is the reverse of the texthexify function.

  Outside of the shiftin/shiftout markers the binary data are represented as hexadecimal. Within the markers the bytes have the values of the ordinals of the characters.

  Example:

  >>> untexthexify('265e25265e25[abcdefghi]29282a29282a')
  b'&^%&^%abcdefghi)(*)(*'
  

  Parameters:

  • s: the string containing the text representation.
  • shiftin: Optional. The marker string commencing a sequence of direct text transcription, default '['.
  • shiftout: Optional. The marker string ending a sequence of direct text transcription, default ']'.

Release Log

Release 20250914:

• New single_space(s[,sep=]) function to split a string and rejoin, with a single space by default.
• Obsolete the @has_format_attributes decorator, how covered by FormatableMixin automatically.
• New FormatMapping(obj,base_format_mapping[,missing]) class presenting a mapping where some entries are callable, used by FormatableMixin.format_as.
• FStr: rename lc to lc_ to reflect the use of the lc_() function.
• format_as: new optional missing(mapping,key) parameter to supply values for keys not in format_mapping.
• FormatAsError: annotate the error with .key, .format_s, .format_mapping.
• Many small internal changes.

Release 20250724: tabulate: new optional ppcls argument to supply a custom PrettyPrinter class for formatting, use a date and datetime aware class by default.

Release 20250428:

• cutprefix,cutsuffix: also accept a tuple of str like str.startswith and str.endswith.
• typed_str: use cropped_repr() instead of repr().

Release 20250414: Obsolete split_remote_path(), supplanted by cs.fs.RemotePath.from_str().

Release 20250323:

• tabulate: format nonstr using pformat.
• New printt() wrapper for tabulate() which prints the table.

Release 20250103:

• Move Identifier, NumericValue, QuotedString in from cs.app.tagger.rules.
• BaseToken: expose the parsing subclass selection as a `.token_classes class method.

Release 20241207: tabulate: split cells containing newlines over multiple output rows.

Release 20241122: tabulate: make the default separator two spaces instead of one, immediate return if no rows (avoids max() of empty list).

Release 20241119.1: Add PyPI classifier, in part to test an updated release script.

Release 20241119: stripped_dedent: new optional sub_indent parameter for indenting the second and following lines, handy for usage messages.

Release 20241109:

• stripped_dedent: new optional post_indent parameter to indent the dedented text.
• New tabulate(*rows) generator function yielding lines of padded columns.

Release 20240630: New indent(paragraph,line_indent=" ") function.

Release 20240519: New get_suffix_part() to extract things line ": Part One" from something such as a TV episode name.

Release 20240316: Fixed release upload artifacts.

Release 20240211: New split_remote_path() function to recognise [[user@]host]:path.

Release 20231018: New is_uc_identifier function.

Release 20230401: Import update.

Release 20230217.1: Fix package requirements.

Release 20230217:

• New get_prefix_n function to parse a numeric value preceeded by a prefix.
• Drop strip_prefix_n, get_prefix_n is more general and I had not got around to using strip_prefix_n yet - when I did, I ended up writing get_prefix_n.

Release 20230210:

• @has_format_attributes: new optional inherit parameter to inherit superclass (or other) format attributes, default False.
• New FNumericMixin, FFloat, FInt FormatableMixin subclasses like FStr - they add .localtime and .utctime formattable attributes.

Release 20220918: typed_str(): crop the value part, default max_length=32, bugfix message cropping.

Release 20220626:

• Remove dependency on cs.py3, we've been Python 2 incompatible for a while.
• FormatableFormatter.format_field: promote None to FStr(None).

Release 20220227:

• typed_str,typed_repr: make max_length the first optional positional parameter, make other parameters keyword only.
• New camelcase() and snakecase() functions.

Release 20211208: Docstring updates.

Release 20210913:

• FormatableFormatter.FORMAT_RE_ARG_NAME_s: strings commencing with digits now match \d+(.\d+)[a-z]+, eg "02d".
• Alias typed_str as s and typed_repr as r.
• FormatableFormatter: new .format_mode thread local state object initially with strict=False, used to control whether unknown fields leave a placeholder or raise KeyError.
• FormatableFormatter.format_field: assorted fixes.

Release 20210906: New strip_prefix_n() function to strip a leading prefix and numeric value n from the start of a string.

Release 20210717:

• Many many changes to FormatableMixin, FormatableFormatter and friends around supporting {foo|conv1|con2|...} instead of {foo!conv}. Still in flux.
• New typed_repr like typed_str but using repr.

Release 20210306:

• New cropped() function to crop strings.
• Rework cropped_repr() to do the repr() itself, and to crop the interiors of tuples and lists.
• cropped_repr: new inner_max_length for cropping the members of collections.
• cropped_repr: special case for length=1 tuples.
• New typed_str(o) object returning type(o).name:str(o) in the default case, useful for debugging.

Release 20201228: Minor doc updates.

Release 20200914:

• Hide terribly special purpose lastlinelen() in cs.hier under a private name.
• New common_prefix and common_suffix function to compare strings.

Release 20200718: get_chars: accept a callable for gochars, indicating a per character test function.

Release 20200613: cropped_repr: replace hardwired 29 with computed length

Release 20200517:

• New get_ini_clausename to parse "[clausename]".
• New get_ini_clause_entryname parsing "[clausename]entryname".
• New cropped_repr for returning a shortened repr()+"..." if the length exceeds a threshold.
• New format_escape function to double {} characters to survive str.format.

Release 20200318:

• New lc_() function to lowercase and dash a string, new titleify_lc() to mostly reverse lc_().
• New format_as function, FormatableMixin and related FormatAsError.

Release 20200229: New cutprefix and cutsuffix functions.

Release 20190812: Fix bad slosh escapes in strings.

Release 20190220: New function get_qstr_or_identifier.

Release 20181108: new function get_decimal_or_float_value to read a decimal or basic float

Release 20180815: No semantic changes; update some docstrings and clean some lint, fix a unit test.

Release 20180810:

• New get_decimal_value and get_hexadecimal_value functions.
• New stripped_dedent function, a slightly smarter textwrap.dedent.

Release 20171231: New function get_decimal. Drop unused function dict2js.

Release 20170904: Python 2/3 ports, move rfc2047 into new cs.rfc2047 module.

Release 20160828:

• Use "install_requires" instead of "requires" in DISTINFO.
• Discard str1(), pointless optimisation.
• unrfc2047: map _ to SPACE, improve exception handling.
• Add phpquote: quote a string for use in PHP code; add docstring to jsquote.
• Add is_identifier test.
• Add get_dotted_identifier.
• Add is_dotted_identifier.
• Add get_hexadecimal.
• Add skipwhite, convenince wrapper for get_white returning just the next offset.
• Assorted bugfixes and improvements.

Release 20150120: cs.lex: texthexify: backport to python 2 using cs.py3 bytes type

Release 20150118: metadata updates

Release 20150116: PyPI metadata and slight code cleanup.