Tanl::Text::RegExp::Pattern Class Reference

Regular Expression matching. More...

#include <RegExp.h>

Inheritance diagram for Tanl::Text::RegExp::Pattern:

Public Member Functions
	Pattern (std::string const &expression, int cflags=0)
	Pattern (char const *expression, int cflags=0)
	Pattern (Pattern const &other)
	Copy constructor.
Pattern &	operator= (Pattern const &other)
	Assignement.
bool	test (std::string const &str, int eflags=0) const
	Tests if the pattern matches at given string str.
bool	test (char const *str, size_t len=0, int eflags=0)
	Tests if the pattern matches at given string str, within the given length len.
int	matchSize (std::string const &text, int eflags=0)
	compute the size of the match.
int	match (const char start, const char end, MatchGroups &pos, int eflags=0)
	Matches the text between start and end and returns the matching positions in pos, expressed as byte-offset from start.
int	match (std::string const &text, MatchGroups &pos, int eflags=0)
	Matches the text in text and returns the matching positions in pos, expressed as adjusted character offset (not byte-offset of the UTF8-Stream).
std::vector< std::string >	match (std::string const &str, int eflags=0)
std::string	replace (std::string &text, std::string &rewrite, bool replaceAll=false)
	Replaces the first substring matching the expression within text with the string rewrite.
Static Public Member Functions
static std::string	escape (std::string &str)
	Escapes all meta characters.
static const unsigned char *	setLocale (char const *locale)
	Set the locale for use during matching.
Static Public Attributes
static const unsigned char *	CharTables = Pattern::setLocale(setlocale(LC_CTYPE, 0))
	The current chartable to use for matching.

Detailed Description

Regular Expression matching.

A pattern is compiled from a regular expression and used in matching. Regular expressions are written using the Perl 5 syntax.

A simple use for testing whether a string matches a pattern is:

      Pattern p("a*b");
      bool b = p.test("aaab");

In order to extract the portions of the string that match, MatchGroups can be used:

      Pattern p("(a*)b");
      MatchGroups m(2);
      string s("daaab");
      int n = p.matches(s, m);

n is the number of groups matched: group 0 represents the substring captured by the whole pattern.

Constructor & Destructor Documentation

Tanl::Text::RegExp::Pattern::Pattern	(	std::string const &	expression,
		int	cflags = `0`
	)

Parameters:

	expression	the regular expression
	cflags	a combination of CompileFlags

NOTE. The ISO Latin-15 locale is used by default: ensure that the locale files for LC_CTYPE=en_US.iso885915 are installed in the OS. This can be changed using SetLocale().

Tanl::Text::RegExp::Pattern::Pattern	(	char const *	expression,
		int	cflags = `0`
	)

Parameters:

	expression	the regular expression
	cflags	a combination of CompileFlags

NOTE. The ISO Latin-15 locale is used by default: ensure that the locale files for LC_CTYPE=en_US.iso885915 are installed in the OS. This can be changed using SetLocale().

References CharTables.

Tanl::Text::RegExp::Pattern::Pattern ( Pattern const & other ) [inline]

Copy constructor.

Use pcre_refcount() to avoid freeing twice _pcre.

Member Function Documentation

std::vector<std::string> Tanl::Text::RegExp::Pattern::match	(	std::string const &	str,
		int	eflags = `0`
	)

Parameters:

	str	the text to match.
	eflags	any combinations of EvaluateFlags

Returns:: an vector<string>: [0] substring matched [1 - n] sub expression with '()'

int Tanl::Text::RegExp::Pattern::match	(	std::string const &	text,
		MatchGroups &	pos,
		int	eflags = `0`
	)

Matches the text in text and returns the matching positions in pos, expressed as adjusted character offset (not byte-offset of the UTF8-Stream).

Parameters:

	text	the string to match.
	pos	the identified matching positions.
	eflags	any combinations of EvaluateFlags

Returns:: 0 if not matching, otherwise the count of matched expressions.

int Tanl::Text::RegExp::Pattern::match	(	const char *	start,
		const char *	end,
		MatchGroups &	pos,
		int	eflags = `0`
	)

Matches the text between start and end and returns the matching positions in pos, expressed as byte-offset from start.

Parameters:

	start	start of the text to match.
	end	end of the text to match.
	pos	the identified matching positions.
	eflags	any combinations of EvaluateFlags

Returns:: 0 if not matching, otherwise the count of matched expressions.

References Tanl::Text::RegExp::MatchGroups::size().

Referenced by Tanl::SST::TokenCategorizer::analyze(), Tanl::NER::TokenCategorizer::analyze(), Tanl::TokenSentenceReader::MoveNext(), Tanl::ConllXSentenceReader::MoveNext(), and Tanl::SentenceReader::MoveNext().

int Tanl::Text::RegExp::Pattern::matchSize	(	std::string const &	text,
		int	eflags = `0`
	)

compute the size of the match.

Parameters:

	text	the text to match.
	eflags	any combinations of EvaluateFlags.

Returns:: 0 if not matching, otherwise the size of the match

Pattern& Tanl::Text::RegExp::Pattern::operator= ( Pattern const & other ) [inline]

Assignement.

Use pcre_refcount() to avoid freeing twice _pcre.

Reimplemented in IXE::Pattern2.

std::string Tanl::Text::RegExp::Pattern::replace	(	std::string &	text,
		std::string &	rewrite,
		bool	replaceAll = `false`
	)

Replaces the first substring matching the expression within text with the string rewrite.

If replaceAll is true, all occurrences are replaced. Within rewrite, backslash-escaped digits ( to ) can be used to insert text matching corresponding parenthesized group from the pattern. in "rewrite" refers to the entire matching text. E.g.,

	string s = "yabba dabba doo";
	RegExp::Pattern("b+").replace(s, "d");

returns "yada dabba doo". The result is undefined if rewrite contains wrong pattern references.

Referenced by Tanl::SST::SstFeatureExtractor::extract(), and Tanl::NER::NerFeatureExtractor::extract().

const unsigned char * Tanl::Text::RegExp::Pattern::setLocale ( char const * locale ) [static]

Set the locale for use during matching.

Use "en_US.iso885915" or similar for recognizing ISO Latin-15 letters.

References CharTables.

bool Tanl::Text::RegExp::Pattern::test	(	char const *	str,
		size_t	len = `0`,
		int	eflags = `0`
	)

Tests if the pattern matches at given string str, within the given length len.

Parameters:

	str	the string to match.
	len	the length of the string to match.
	eflags	any combinations of EvaluateFlags

Returns:: true if matches

bool Tanl::Text::RegExp::Pattern::test	(	std::string const &	str,
		int	eflags = `0`
	)			const

Tests if the pattern matches at given string str.

Parameters:

	str	the string to match.
	eflags	any combinations of EvaluateFlags

Returns:: true if matches

Referenced by Tanl::NER::NerFeatureExtractor::extract(), and Parser::ParseState::transition().

The documentation for this class was generated from the following files:

text/RegExp.h
text/RegExp.cpp

Tanl::Text::RegExp::Pattern Class Reference

Public Member Functions

Static Public Member Functions

Static Public Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation