mirror of
https://github.com/twitter/the-algorithm.git
synced 2024-06-27 13:36:03 +02:00
ef4c5eb65e
Please note we have force-pushed a new initial commit in order to remove some publicly-available Twitter user information. Note that this process may be required in the future.
84 lines
3.9 KiB
Scala
84 lines
3.9 KiB
Scala
package com.twitter.home_mixer.functional_component.feature_hydrator
|
|
|
|
import com.twitter.finagle.stats.StatsReceiver
|
|
import com.twitter.home_mixer.product.scored_tweets.param.ScoredTweetsParam
|
|
import com.twitter.ml.api.DataRecord
|
|
import com.twitter.product_mixer.component_library.model.candidate.TweetCandidate
|
|
import com.twitter.product_mixer.core.feature.Feature
|
|
import com.twitter.product_mixer.core.feature.FeatureWithDefaultOnFailure
|
|
import com.twitter.product_mixer.core.feature.datarecord.DataRecordInAFeature
|
|
import com.twitter.product_mixer.core.feature.featuremap.FeatureMap
|
|
import com.twitter.product_mixer.core.feature.featuremap.FeatureMapBuilder
|
|
import com.twitter.product_mixer.core.functional_component.feature_hydrator.BulkCandidateFeatureHydrator
|
|
import com.twitter.product_mixer.core.model.common.CandidateWithFeatures
|
|
import com.twitter.product_mixer.core.model.common.Conditionally
|
|
import com.twitter.product_mixer.core.model.common.identifier.FeatureHydratorIdentifier
|
|
import com.twitter.product_mixer.core.pipeline.PipelineQuery
|
|
import com.twitter.stitch.Stitch
|
|
import com.twitter.timelines.clients.strato.twistly.SimClustersRecentEngagementSimilarityClient
|
|
import com.twitter.timelines.configapi.decider.BooleanDeciderParam
|
|
import com.twitter.timelines.prediction.adapters.twistly.SimClustersRecentEngagementSimilarityFeaturesAdapter
|
|
import javax.inject.Inject
|
|
import javax.inject.Singleton
|
|
|
|
object SimClustersEngagementSimilarityFeature
|
|
extends DataRecordInAFeature[PipelineQuery]
|
|
with FeatureWithDefaultOnFailure[PipelineQuery, DataRecord] {
|
|
override def defaultValue: DataRecord = new DataRecord()
|
|
}
|
|
|
|
@Singleton
|
|
class SimClustersEngagementSimilarityFeatureHydrator @Inject() (
|
|
simClustersEngagementSimilarityClient: SimClustersRecentEngagementSimilarityClient,
|
|
statsReceiver: StatsReceiver)
|
|
extends BulkCandidateFeatureHydrator[PipelineQuery, TweetCandidate]
|
|
with Conditionally[PipelineQuery] {
|
|
|
|
override val identifier: FeatureHydratorIdentifier =
|
|
FeatureHydratorIdentifier("SimClustersEngagementSimilarity")
|
|
|
|
override val features: Set[Feature[_, _]] = Set(SimClustersEngagementSimilarityFeature)
|
|
|
|
private val scopedStatsReceiver = statsReceiver.scope(identifier.toString)
|
|
|
|
private val hydratedCandidatesSizeStat = scopedStatsReceiver.stat("hydratedCandidatesSize")
|
|
|
|
private val simClustersRecentEngagementSimilarityFeaturesAdapter =
|
|
new SimClustersRecentEngagementSimilarityFeaturesAdapter
|
|
|
|
override def onlyIf(query: PipelineQuery): Boolean = {
|
|
val param: BooleanDeciderParam =
|
|
ScoredTweetsParam.EnableSimClustersSimilarityFeatureHydrationDeciderParam
|
|
query.params.apply(param)
|
|
}
|
|
|
|
override def apply(
|
|
query: PipelineQuery,
|
|
candidates: Seq[CandidateWithFeatures[TweetCandidate]]
|
|
): Stitch[Seq[FeatureMap]] = {
|
|
val tweetToCandidates = candidates.map(candidate => candidate.candidate.id -> candidate).toMap
|
|
val tweetIds = tweetToCandidates.keySet.toSeq
|
|
val userId = query.getRequiredUserId
|
|
val userTweetEdges = tweetIds.map(tweetId => (userId, tweetId))
|
|
val resultFuture = simClustersEngagementSimilarityClient
|
|
.getSimClustersRecentEngagementSimilarityScores(userTweetEdges).map {
|
|
simClustersRecentEngagementSimilarityScoresMap =>
|
|
hydratedCandidatesSizeStat.add(simClustersRecentEngagementSimilarityScoresMap.size)
|
|
candidates.map { candidate =>
|
|
val similarityFeatureOpt = simClustersRecentEngagementSimilarityScoresMap
|
|
.get(userId -> candidate.candidate.id).flatten
|
|
val dataRecordOpt = similarityFeatureOpt.map { similarityFeature =>
|
|
simClustersRecentEngagementSimilarityFeaturesAdapter
|
|
.adaptToDataRecords(similarityFeature)
|
|
.get(0)
|
|
}
|
|
FeatureMapBuilder()
|
|
.add(SimClustersEngagementSimilarityFeature, dataRecordOpt.getOrElse(new DataRecord))
|
|
.build()
|
|
}
|
|
}
|
|
Stitch.callFuture(resultFuture)
|
|
}
|
|
|
|
}
|