Package org.opensextant.extractors.geo.rules

Class MajorPlaceRule

java.lang.Object

org.opensextant.extractors.geo.rules.GeocodeRule

org.opensextant.extractors.geo.rules.MajorPlaceRule

public class MajorPlaceRule
extends GeocodeRule

Major Place rule -- fire this rule after Country rule. Try to find all countries in scope first, then major places. If you try to infer country from major places first you get a lot of false positives. Country name space is smaller and more reliable. LOTS of caveats: these rules enforce the notion that country names are drivers here, and major places amplify. IF we see a National Capital we can infer a country, provided no countries have been seen in document IF we see a major place, add that evidence weighting it higher if the country of that major place is also mentioned in document.

Author:

ubaldino

Field Summary

Fields

Modifier and Type	Field	Description
static final String	ADMIN
static final String	CAPITAL
static final String	MENTIONED_COUNTRY
static final String	POP

Fields inherited from class org.opensextant.extractors.geo.rules.GeocodeRule

AVG_WORD_LEN, boundaryObserver, coordObserver, countryObserver, defaultMethod, LEX1, LEX2, locationOnly, log, LOWERCASE, NAME, textCase, UPPERCASE, weight

Constructor Summary

Constructors

Constructor	Description
MajorPlaceRule(Map<String,Integer> populationStats)	Major Place assigns a score to places that are national capitals, provinces, or cities with sizable population.

Method Summary

Modifier and Type	Method	Description
void	evaluate(List<PlaceCandidate> names)
void	evaluate(PlaceCandidate name, org.opensextant.data.Place geo)	attach either a Capital or Admin region ID, giving it some weight based on various properties or context.
static boolean	isRuleFor(PlaceCandidate pc)	Determine if this rule was applied to the candidate.
void	reset()	no-op, unless overriden.

Methods inherited from class org.opensextant.extractors.geo.rules.GeocodeRule

filterByNameOnly, filterOutByFrequency, internalPlaceID, isRelevant, isShort, logMsg, sameBoundary, sameCountry, sameCountry, sameLexicalName, setBoundaryObserver, setCountryObserver, setDefaultMethod, setGeohash, setLocationObserver, setTextCase, textCase

Methods inherited from class java.lang.Object

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

Field Details

CAPITAL

public static final String CAPITAL

See Also:

• Constant Field Values

ADMIN

public static final String ADMIN

See Also:

• Constant Field Values

POP

public static final String POP

See Also:

• Constant Field Values

MENTIONED_COUNTRY

public static final String MENTIONED_COUNTRY

See Also:

• Constant Field Values

Constructor Details

MajorPlaceRule

public MajorPlaceRule(Map<String,Integer> populationStats)

Major Place assigns a score to places that are national capitals, provinces, or cities with sizable population. Log(population) adds up to one point to place weight. Population data is indexed by location/grid using geohash. Source:geonames.org Population stats are deterministic -- they do not change during the processing and they are not context specific. So we only assess population per location ONCE not per mention.

Parameters:

populationStats - optional population stats.

Method Details

reset

public void reset()

Description copied from class: GeocodeRule

no-op, unless overriden.

Overrides:

reset in class GeocodeRule

evaluate

public void evaluate(List<PlaceCandidate> names)

Overrides:

evaluate in class GeocodeRule

Parameters:

names - list of found place names

isRuleFor

public static boolean isRuleFor(PlaceCandidate pc)

Determine if this rule was applied to the candidate.

Parameters:

pc -

Returns:

evaluate

public void evaluate(PlaceCandidate name,
 org.opensextant.data.Place geo)

attach either a Capital or Admin region ID, giving it some weight based on various properties or context.

Specified by:

evaluate in class GeocodeRule

Parameters:

name - matched name in text

geo - gazetteer entry or location