the-algorithm/home-mixer/server/src/main/scala/com/twitter/home_mixer/functional_component/filter/PreviouslySeenTweetsFilter.scala
twitter-team ef4c5eb65e Twitter Recommendation Algorithm
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.
2023-03-31 17:36:31 -05:00

38 lines
1.5 KiB
Scala

package com.twitter.home_mixer.functional_component.filter
import com.twitter.home_mixer.util.CandidatesUtil
import com.twitter.home_mixer.util.TweetImpressionsHelper
import com.twitter.product_mixer.component_library.model.candidate.TweetCandidate
import com.twitter.product_mixer.core.functional_component.filter.Filter
import com.twitter.product_mixer.core.functional_component.filter.FilterResult
import com.twitter.product_mixer.core.model.common.CandidateWithFeatures
import com.twitter.product_mixer.core.model.common.identifier.FilterIdentifier
import com.twitter.product_mixer.core.pipeline.PipelineQuery
import com.twitter.stitch.Stitch
/**
* Filter out users' previously seen tweets from 2 sources:
* 1. Heron Topology Impression Store in Memcache;
* 2. Manhattan Impression Store;
*/
object PreviouslySeenTweetsFilter extends Filter[PipelineQuery, TweetCandidate] {
override val identifier: FilterIdentifier = FilterIdentifier("PreviouslySeenTweets")
override def apply(
query: PipelineQuery,
candidates: Seq[CandidateWithFeatures[TweetCandidate]]
): Stitch[FilterResult[TweetCandidate]] = {
val seenTweetIds =
query.features.map(TweetImpressionsHelper.tweetImpressions).getOrElse(Set.empty)
val (removed, kept) = candidates.partition { candidate =>
val tweetIdAndSourceId = CandidatesUtil.getTweetIdAndSourceId(candidate)
tweetIdAndSourceId.exists(seenTweetIds.contains)
}
Stitch.value(FilterResult(kept = kept.map(_.candidate), removed = removed.map(_.candidate)))
}
}