NERTagger Class Reference

#include <ner_tagger_mod.h>

Inheritance diagram for NERTagger:

Collaboration diagram for NERTagger:

Public Member Functions
	NERTagger (TiCC::LogStream *, TiCC::LogStream *=0)
bool	init (const TiCC::Configuration &)
void	Classify (frog_data &)
void	post_process (frog_data &)
void	post_process (frog_data &, const std::vector< tc_pair > &)
void	add_declaration (folia::Document &, folia::processor *) const
void	add_result (const frog_data &fd, const std::vector< folia::Word * > &wv) const
bool	read_gazets (const std::string &f, const std::string &p)
bool	read_overrides (const std::string &f, const std::string &p)
std::vector< std::string >	create_ner_list (const std::vector< std::string > &s)
std::vector< std::string >	create_override_list (const std::vector< std::string > &s)
bool	Generate (const std::string &)
void	merge_override (std::vector< tc_pair > &, const std::vector< tc_pair > &, bool, const std::vector< std::string > &) const
Public Member Functions inherited from BaseTagger
	BaseTagger (TiCC::LogStream *, TiCC::LogStream *, const std::string &)
virtual	~BaseTagger ()
void	add_provenance (folia::Document &, folia::processor *) const
std::string	getTagset () const
std::string	set_eos_mark (const std::string &)
bool	fill_map (const std::string &, std::map< std::string, std::string > &)
std::vector< Tagger::TagResult >	tagLine (const std::string &)
std::vector< Tagger::TagResult >	tag_entries (const std::vector< tag_entry > &)
std::string	version () const

Additional Inherited Members
Protected Member Functions inherited from BaseTagger
void	extract_words_tags (const std::vector< folia::Word * > &, const std::string &, std::vector< std::string > &, std::vector< std::string > &)
std::vector< Tagger::TagResult >	call_server (const std::vector< tag_entry > &) const
	BaseTagger (const BaseTagger &)
Protected Attributes inherited from BaseTagger
int	debug
std::string	_label
std::string	tagset
std::string	_version
std::string	textclass
TiCC::LogStream *	err_log
TiCC::LogStream *	dbg_log
std::string	base
std::string	_host
std::string	_port
MbtAPI *	tagger
TiCC::UniFilter *	filter
std::vector< std::string >	_words
std::vector< Tagger::TagResult >	_tag_result
std::map< std::string, std::string >	token_tag_map

Constructor & Destructor Documentation

NERTagger()

NERTagger::NERTagger ( TiCC::LogStream *  l, TiCC::LogStream *  d = 0  )

explicit

initialize a NERTagger instance

Parameters

l	a LogStream for errors
d	a LogStream for debugging

Member Function Documentation

add_declaration()

void NERTagger::add_declaration ( folia::Document &  doc, folia::processor *  proc  ) const

virtual

add ENTITY annotation as an AnnotationType to the document

Parameters

doc	the Document the add to
proc	the processor to add

Implements BaseTagger.

add_result()

void NERTagger::add_result	(	const frog_data &	fd,
		const std::vector< folia::Word * > &	wv
	)		const

add the NER tags in 'fd' to the FoLiA list of Word

Parameters

fd	The tagged results
wv	The folia:Word vector

Classify()

void NERTagger::Classify ( frog_data &  swords )

virtual

Tag one sentence, given in frog_data format

Parameters

swords

the frog_data structure to analyze

When tagging succeeds, 'swords' will be extended with the tag results

Reimplemented from BaseTagger.

create_ner_list()

std::vector<std::string> NERTagger::create_ner_list ( const std::vector< std::string > &  s )

inline

create_override_list()

std::vector<std::string> NERTagger::create_override_list ( const std::vector< std::string > &  s )

inline

Generate()

bool NERTagger::Generate

(

const std::string &

opt_line

)

generate a new tagger using opt_line

init()

bool NERTagger::init ( const TiCC::Configuration &  config )

virtual

initalize a NER tagger from 'config'

Parameters

config

the TiCC::Configuration

Returns

true on succes, false otherwise

first BaseTagger::init() is called to set generic values, then the NER specific values for the gazeteer file-names etc. are added and the files are read.

Reimplemented from BaseTagger.

merge_override()

void NERTagger::merge_override	(	std::vector< tc_pair > &	tags,
		const std::vector< tc_pair > &	overrides,
		bool	unconditional,
		const std::vector< std::string > &	POS_tags
	)		const

merge overrides into the tc_pair vector

Parameters

tags	the list of tc_pair to merge into
overrides	the override information (from gazeteers and such)
unconditional	when true, prefer the information from overrides, otherwise only override when there was no tag assigned yet
POS_tags	a list of POS-tags. When this list is non-empty, only override when it contains a N or a SPEC tag.

post_process() [1/2]

void NERTagger::post_process ( frog_data &  )

virtual

This implements BaseTagger::post_process by DISALLOWING it.

Implements BaseTagger.

post_process() [2/2]

void NERTagger::post_process	(	frog_data &	sentence,
		const std::vector< tc_pair > &	ners
	)

finish the NER processs by updating 'sentence'

Parameters

sentence	a frog_data structure to update with NER info
ners	a sequence of NE tags (maybe 'O') with their confidence

This is a specialized version local to NERTagger. Don't use the norma; post_process()

read_gazets()

bool NERTagger::read_gazets ( const std::string &  f, const std::string &  p  )

inline

read_overrides()

bool NERTagger::read_overrides ( const std::string &  f, const std::string &  p  )

inline

---

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

• include/frog/ner_tagger_mod.h
• src/ner_tagger_mod.cxx