Class AnnotationsWebService.PriorCurationValueObject

java.lang.Object

ubic.gemma.rest.AnnotationsWebService.PriorCurationValueObject

Enclosing class:

AnnotationsWebService

public static class AnnotationsWebService.PriorCurationValueObject
extends Object

A term prior curators chose for the query string, with the number of distinct experiments they chose it on.

Deliberately NOT in data, for the same reason ruledOut is not: these terms did not come from the ontology search at all. They are drawn from what curators have recorded, which answers a question lexical search cannot. vehicle, untreated and sham are annotated with reference substance role and reference subject role — terms sharing no word with the string, so no label or synonym search returns them under any ranking. Likewise EAE matches episodic angioedema with eosinophilia lexically and means experimental autoimmune encephalomyelitis, which is what the corpus records.

A client reading data[0] without knowing this field must not silently pick up a term that was never returned by the search, which is why it lives in its own section.

⚠️ These are counts of what curators did, not of what is right. A string mis-tagged for years arrives with a large and confident-looking count. Read an entry as evidence with a denominator attached.

Author:

tesarst

Constructor Summary

Constructors

Constructor	Description
PriorCurationValueObject(String valueUri, String value, long experimentCount, double agreement)

Method Summary

Modifier and Type	Method	Description
double	getAgreement()	This term's share of every annotation made from this string, in [0, 1].
long	getExperimentCount()	Distinct experiments on which this string was annotated with this term.
String	getValue()	The label stored alongside the term in the corpus.
String	getValueUri()	The term prior curators chose.

Methods inherited from class Object

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

Constructor Details

PriorCurationValueObject

public PriorCurationValueObject(String valueUri,
 @Nullable
 String value,
 long experimentCount,
 double agreement)

Method Details

getAgreement

public double getAgreement()

This term's share of every annotation made from this string, in [0, 1].

Read this before the count. The count says how much evidence there is; agreement says whether curators agreed, and the two come apart. wild type resolves to wild type genotype with agreement ≈ 1.0 — settled convention. sham splits 187/28 between reference subject role and reference substance role — still a large count, but a contested one, and a consumer should not treat the two the same way.

This is the field that makes the section safe to use. A convention that was consistently wrong is indistinguishable from one that was consistently right, but a convention that was never settled is visible here, and that is the case where a downstream stage should stop and reason rather than adopt.

getValueUri

public String getValueUri()

The term prior curators chose.

getValue

@Nullable
public String getValue()

The label stored alongside the term in the corpus. Read from curation records rather than resolved from the ontology, so it stays populated for ontologies that are not loaded.

getExperimentCount

public long getExperimentCount()

Distinct experiments on which this string was annotated with this term.