Package org.opensextant.extractors.flexpat

Class RegexPatternManager

java.lang.Object

org.opensextant.extractors.flexpat.RegexPatternManager

Direct Known Subclasses:

PatternManager, PatternManager, PoliPatternManager

public abstract class RegexPatternManager
extends Object

This is the culmination of various date/time extraction efforts in python and Java. This API poses no assumptions on input data or on execution. Features of REGEX patterns file:

• DEFINE - a component of a pattern to match
• RULE - a complete pattern to match

This work started in Java 6 and has the limitation of Java 6 Regex, mainly that there are no named groups available in matching.

See XCoord PatternManager for a good example implementation.

Author:

dlutz (lutzdavp), ubaldino

Field Summary

Fields

Modifier and Type	Field	Description
boolean	debug
protected org.slf4j.Logger	log
protected String	patternFile
protected Map<String,RegexPattern>	patterns
protected List<RegexPattern>	patterns_list
List<PatternTestCase>	testcases
boolean	testing

Constructor Summary

Constructors

Constructor	Description
RegexPatternManager(InputStream s, String n)

Method Summary

Modifier and Type	Method	Description
protected abstract RegexPattern	create_pattern(String fam, String rule, String desc)	Implementation must create a RegexPattern given the basic RULE define, #RULE FAMILY RID REGEX PatternManager here adds compiled pattern and DEFINES.
protected abstract PatternTestCase	create_testcase(String id, String fam, String text)	Implementation must create TestCases given the #TEST directive, #TEST RID TID TEXT
void	disableAll()	Enable a family of patterns
abstract void	enable_pattern(RegexPattern p)	enable an instance of a pattern based on the global settings.
void	enable_patterns(String name)	default adapter -- you must override.
void	enableAll()
RegexPattern	get_pattern(String id)	Access the paterns by ID
Collection<RegexPattern>	get_patterns()
String	getConfigurationDebug()	Instead of relying on a logging API, we now throw Exceptionsages for real configuration errors, and capture configuration details in a buffer if debug is on.
Map<String,String>	group_map(RegexPattern p, Matcher matched)	NOTE: We're dealing with Java6's inability to use named groups.
Map<String,TextEntity>	group_matches(RegexPattern p, Matcher matched)	Matched fields as TextEntities
void	initialize(InputStream io)	Initializes the pattern manager implementations.
protected abstract boolean	validate_pattern(RegexPattern pat)	Implementation has the option to check a pattern; For now invalid patterns are only logged.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

log

protected org.slf4j.Logger log

patterns

protected Map<String,RegexPattern> patterns

patterns_list

protected List<RegexPattern> patterns_list

patternFile

protected String patternFile

debug

public boolean debug

testing

public boolean testing

testcases

public List<PatternTestCase> testcases

Constructor Details

RegexPatternManager

public RegexPatternManager(InputStream s,
 String n)
                    throws IOException

Throws:

IOException

Method Details

get_patterns

public Collection<RegexPattern> get_patterns()

Returns:

collection of patterns

get_pattern

public RegexPattern get_pattern(String id)

Access the paterns by ID

Parameters:

id - pattern id

Returns:

found pattern or null

create_pattern

protected abstract RegexPattern create_pattern(String fam,
 String rule,
 String desc)

Implementation must create a RegexPattern given the basic RULE define, #RULE FAMILY RID REGEX PatternManager here adds compiled pattern and DEFINES.

Parameters:

fam - family

rule - rule ID within the family

desc - optional description

Returns:

pattern object

validate_pattern

protected abstract boolean validate_pattern(RegexPattern pat)

Implementation has the option to check a pattern; For now invalid patterns are only logged.

Parameters:

pat - pattern object

Returns:

true if pattern is valid

create_testcase

protected abstract PatternTestCase create_testcase(String id,
 String fam,
 String text)

Implementation must create TestCases given the #TEST directive, #TEST RID TID TEXT

Parameters:

id - pattern id

fam - pattern family

text - text for test case

Returns:

test case object

enable_pattern

public abstract void enable_pattern(RegexPattern p)

enable an instance of a pattern based on the global settings.

Parameters:

p - the pattern obj to enable

enable_patterns

public void enable_patterns(String name)

default adapter -- you must override. This should be abstract, but not all pattern managers are required to support this.

Parameters:

name - pattern name to enable.

disableAll

public void disableAll()

Enable a family of patterns

enableAll

public void enableAll()

initialize

public void initialize(InputStream io)
                throws IOException

Initializes the pattern manager implementations. Reads the DEFINEs and RULEs from the pattern file and does the requisite substitutions. After initialization patterns HashMap will be populated.

Parameters:

io - stream

Throws:

IOException - if patterns file can not be loaded and parsed

getConfigurationDebug

public String getConfigurationDebug()

Instead of relying on a logging API, we now throw Exceptionsages for real configuration errors, and capture configuration details in a buffer if debug is on.

Returns:

the configuration debug

group_map

public Map<String,String> group_map(RegexPattern p,
 Matcher matched)

NOTE: We're dealing with Java6's inability to use named groups. So we have to track FlexPat slots in line with Matcher fields matched. Essentially this comes down to a simple Name:Offset pairing; our limitation here is no nesting.

Parameters:

p - pattern

matched - matcher

Returns:

map containing the matched groups, as deciphered by Flexpat and the definitions in the patterns file

group_matches

public Map<String,TextEntity> group_matches(RegexPattern p,
 Matcher matched)

Matched fields as TextEntities

Parameters:

p - pattern

matched - java RE Matcher

Returns:

keyed TextEntity