
nlp-core-service.21.9.6.source-code.ModelCoreService.kt Maven / Gradle / Ivy
/*
* Copyright (C) 2017/2021 e-voyageurs technologies
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package ai.tock.nlp.core.service
import ai.tock.nlp.core.Application
import ai.tock.nlp.core.BuildContext
import ai.tock.nlp.core.CallContext
import ai.tock.nlp.core.EntityRecognition
import ai.tock.nlp.core.EntityType
import ai.tock.nlp.core.Intent
import ai.tock.nlp.core.ModelCore
import ai.tock.nlp.core.NlpEngineType
import ai.tock.nlp.core.configuration.NlpApplicationConfiguration
import ai.tock.nlp.core.quality.EntityMatchError
import ai.tock.nlp.core.quality.IntentMatchError
import ai.tock.nlp.core.quality.TestContext
import ai.tock.nlp.core.quality.TestModelReport
import ai.tock.nlp.core.sample.SampleEntity
import ai.tock.nlp.core.sample.SampleExpression
import ai.tock.nlp.model.EntityBuildContext
import ai.tock.nlp.model.EntityBuildContextForIntent
import ai.tock.nlp.model.EntityBuildContextForSubEntities
import ai.tock.nlp.model.EntityCallContextForIntent
import ai.tock.nlp.model.IntentContext
import ai.tock.nlp.model.NlpClassifier
import ai.tock.shared.error
import ai.tock.shared.injector
import ai.tock.shared.normalize
import com.github.salomonbrys.kodein.instance
import java.time.Duration
import java.time.Instant
import mu.KotlinLogging
/**
*
*/
internal object ModelCoreService : ModelCore {
private val logger = KotlinLogging.logger {}
private val nlpClassifier: NlpClassifier by injector.instance()
override fun warmupModels(context: BuildContext) {
nlpClassifier.warmupIntentModel(IntentContext(context))
context
.application
.intents
.filter { it.entities.isNotEmpty() }
.forEach {
nlpClassifier.warmupEntityModel(
EntityCallContextForIntent(
CallContext(context.application, context.language, context.engineType),
it
)
)
}
}
override fun updateIntentModel(context: BuildContext, expressions: List) {
val nlpContext = IntentContext(context)
if (!context.onlyIfNotExists || !nlpClassifier.isIntentModelExist(nlpContext)) {
nlpClassifier.buildAndSaveIntentModel(nlpContext, context.formatExpressions(expressions))
}
}
override fun updateEntityModelForIntent(
context: BuildContext,
intent: Intent,
expressions: List
) {
val nlpContext = EntityBuildContextForIntent(context, intent)
updateEntityModel(context, nlpContext, expressions)
}
override fun updateEntityModelForEntityType(
context: BuildContext,
entityType: EntityType,
expressions: List
) {
val nlpContext = EntityBuildContextForSubEntities(context, entityType)
updateEntityModel(context, nlpContext, expressions)
}
private fun updateEntityModel(
context: BuildContext,
nlpContext: EntityBuildContext,
expressions: List
) {
if (!context.onlyIfNotExists ||
!nlpClassifier.isEntityModelExist(nlpContext)
) {
nlpClassifier.buildAndSaveEntityModel(nlpContext, context.formatExpressions(expressions))
}
}
override fun deleteOrphans(applicationsAndIntents: Map>, entityTypes: List) {
nlpClassifier.deleteOrphans(applicationsAndIntents, entityTypes)
}
override fun testModel(context: TestContext, expressions: List): TestModelReport {
if (expressions.size < 100) {
error("at least 100 expressions needed")
}
val shuffle = expressions.toMutableList()
shuffle.shuffle()
val limit = (expressions.size * context.threshold).toInt()
val modelExpressions = shuffle.subList(0, limit)
val testedExpressions = shuffle.subList(limit, shuffle.size)
val startDate = Instant.now()
val intentContext = IntentContext(context)
val intentModel = nlpClassifier.buildIntentModel(intentContext, context.formatExpressions(modelExpressions))
val entityModels = modelExpressions
.groupBy { it.intent }
.mapNotNull { (intent, expressions)
->
try {
intent to nlpClassifier.buildEntityModel(
EntityBuildContextForIntent(context, intent),
context.formatExpressions(expressions)
)
} catch (e: Exception) {
logger.error { "entity model build fail for $intent " }
logger.error(e)
null
}
}
.toMap()
val buildDuration = Duration.between(startDate, Instant.now())
val intentErrors = mutableListOf()
val entityErrors = mutableListOf()
testedExpressions.forEach {
val parseResult = NlpCoreService.parse(
context,
it.text,
intentModel,
entityModels
)
if (parseResult.intent != it.intent.name) {
intentErrors.add(IntentMatchError(it, parseResult.intent, parseResult.intentProbability))
} else if (hasNotSameEntities(it.entities, parseResult.entities)) {
entityErrors.add(EntityMatchError(it, parseResult.entities))
}
}
val testDuration = Duration.between(startDate.plus(buildDuration), Instant.now())
return TestModelReport(
expressions,
testedExpressions,
intentErrors,
entityErrors,
buildDuration,
testDuration,
startDate
)
}
private fun hasNotSameEntities(
expectedEntities: List,
entities: List
): Boolean {
return expectedEntities.any { e ->
entities.none {
it.role == e.definition.role && it.entityType == e.definition.entityType && it.isSameRange(
e
)
}
} ||
entities.any {
expectedEntities.none { e ->
it.role == e.definition.role && it.entityType == e.definition.entityType && it.isSameRange(
e
)
}
}
}
override fun getCurrentModelConfiguration(
applicationName: String,
nlpEngineType: NlpEngineType
): NlpApplicationConfiguration =
nlpClassifier.getCurrentModelConfiguration(applicationName, nlpEngineType)
override fun updateModelConfiguration(
applicationName: String,
engineType: NlpEngineType,
configuration: NlpApplicationConfiguration
) = nlpClassifier.updateModelConfiguration(applicationName, engineType, configuration)
private fun BuildContext.formatExpressions(expressions: List): List {
return if (application.normalizeText)
expressions.map { e -> e.copy(text = e.text.normalize(language)) }
else expressions
}
private fun TestContext.formatExpressions(expressions: List): List {
return if (callContext.application.normalizeText) {
expressions.map { e -> e.copy(text = e.text.normalize(callContext.language)) }
} else expressions
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy