feat: fallback to fast string splitting

fulcrumgenomics · Jan 3, 2024 · f806f57 · f806f57
1 parent 9d5a422
commit f806f57
Show file tree

Hide file tree

Showing 5 changed files with 6 additions and 14 deletions.
diff --git a/src/main/scala/com/fulcrumgenomics/umi/CollectDuplexSeqMetrics.scala b/src/main/scala/com/fulcrumgenomics/umi/CollectDuplexSeqMetrics.scala
@@ -31,7 +31,6 @@ import com.fulcrumgenomics.cmdline.{ClpGroups, FgBioTool}
 import com.fulcrumgenomics.commons.collection.BetterBufferedIterator
 import com.fulcrumgenomics.commons.util.{LazyLogging, NumericCounter, SimpleCounter}
 import com.fulcrumgenomics.sopt.{arg, clp}
-import com.fulcrumgenomics.umi.Umis.UmiSeparatorPattern
 import com.fulcrumgenomics.util.Metric.{Count, Proportion}
 import com.fulcrumgenomics.util._
 import htsjdk.samtools.util.{Interval, IntervalList, Murmur3, OverlapDetector}
@@ -390,8 +389,8 @@ class CollectDuplexSeqMetrics
     val umi1s = IndexedSeq.newBuilder[String] // ab UMI 1 (and ba UMI 2) sequences
     val umi2s = IndexedSeq.newBuilder[String] // ab UMI 2 (and ba UMI 1) sequences
 
-    ab.iterator.map(r => UmiSeparatorPattern.split(r[String](this.umiTag), -1)).foreach { case Array(u1, u2) => umi1s += u1; umi2s += u2 }
-    ba.iterator.map(r => UmiSeparatorPattern.split(r[String](this.umiTag), -1)).foreach { case Array(u1, u2) => umi1s += u2; umi2s += u1 }
+    ab.iterator.map(r => r[String](this.umiTag).split("-", -1)).foreach { case Array(u1, u2) => umi1s += u1; umi2s += u2 }
+    ba.iterator.map(r => r[String](this.umiTag).split("-", -1)).foreach { case Array(u1, u2) => umi1s += u2; umi2s += u1 }
 
     val Seq(abConsensusUmi, baConsensusUmi) = Seq(umi1s, umi2s).map(_.result()).map{ umis =>
       val consensus = this.consensusBuilder.callConsensus(umis)
@@ -531,7 +530,7 @@ class CollectDuplexSeqMetrics
     val uniqueTotal      = metrics.map(_.unique_observations).sum.toDouble
 
     metrics.foreach { m =>
-      val Array(umi1, umi2)                   = UmiSeparatorPattern.split(m.umi, -1)
+      val Array(umi1, umi2)                   = m.umi.split("-", -1)
       m.fraction_raw_observations             = m.raw_observations / rawTotal
       m.fraction_unique_observations          = m.unique_observations / uniqueTotal
       m.fraction_unique_observations_expected = singleUmiMetrics(umi1).fraction_unique_observations * singleUmiMetrics(umi2).fraction_unique_observations

diff --git a/src/main/scala/com/fulcrumgenomics/umi/CorrectUmis.scala b/src/main/scala/com/fulcrumgenomics/umi/CorrectUmis.scala
@@ -185,7 +185,7 @@ class CorrectUmis
           missingUmisRecords += 1
           rejectOut.foreach(w => w += rec)
         case Some(umi: String) =>
-          val sequences = umi.split('-')
+          val sequences = umi.split("-", -1)
           if (sequences.exists(_.length != umiLength)) {
             if (wrongLengthRecords == 0) {
               logger.warning(s"Read (${rec.name}) detected with unexpected length UMI(s): ${sequences.mkString(" ")}.")

diff --git a/src/main/scala/com/fulcrumgenomics/umi/DuplexConsensusCaller.scala b/src/main/scala/com/fulcrumgenomics/umi/DuplexConsensusCaller.scala
@@ -31,7 +31,6 @@ import com.fulcrumgenomics.commons.util.LazyLogging
 import com.fulcrumgenomics.umi.DuplexConsensusCaller._
 import com.fulcrumgenomics.umi.UmiConsensusCaller.ReadType.{ReadType, _}
 import com.fulcrumgenomics.umi.UmiConsensusCaller.{SimpleRead, SourceRead}
-import com.fulcrumgenomics.umi.Umis.UmiSeparatorPattern
 import com.fulcrumgenomics.util.NumericTypes.PhredScore
 
 /**
@@ -324,7 +323,7 @@ class DuplexConsensusCaller(override val readNamePrefix: String,
     // UMI bases are present, `None` otherwise.
     reads.flatMap(_.sam).flatMap { rec =>
       rec.get[String](ConsensusTags.UmiBases).map { umi =>
-        if (rec.firstOfPair == firstOfPair) umi else UmiSeparatorPattern.split(umi, -1).reverse.mkString("-")
+        if (rec.firstOfPair == firstOfPair) umi else umi.split("-", -1).reverse.mkString("-")
       }
     }
   }

diff --git a/src/main/scala/com/fulcrumgenomics/umi/GroupReadsByUmi.scala b/src/main/scala/com/fulcrumgenomics/umi/GroupReadsByUmi.scala
@@ -34,7 +34,6 @@ import com.fulcrumgenomics.cmdline.{ClpGroups, FgBioTool}
 import com.fulcrumgenomics.commons.util.{LazyLogging, NumericCounter, SimpleCounter}
 import com.fulcrumgenomics.sopt.{arg, clp}
 import com.fulcrumgenomics.umi.GroupReadsByUmi._
-import com.fulcrumgenomics.umi.Umis.UmiSeparatorPattern
 import com.fulcrumgenomics.util.Metric.{Count, Proportion}
 import com.fulcrumgenomics.util.Sequences.countMismatches
 import com.fulcrumgenomics.util._
@@ -832,7 +831,7 @@ class GroupReadsByUmi
         val pos1 = if (r1.positiveStrand) r1.unclippedStart else r1.unclippedEnd
         val pos2 = if (r2.positiveStrand) r2.unclippedStart else r2.unclippedEnd
         val r1Lower = r1.refIndex < r2.refIndex || (r1.refIndex == r2.refIndex && (pos1 < pos2 || (pos1 == pos2 && r1.positiveStrand)))
-        val umis = UmiSeparatorPattern.split(umi, -1) // Split and ensure we return empty strings for missing UMIs.
+        val umis = umi.split("-", -1) // Split and ensure we return empty strings for missing UMIs.
         require(umis.length == 2, s"Paired strategy used but umi did not contain 2 segments delimited by a '-': $umi")
 
         if (r1Lower) paired.lowerReadUmiPrefix  + ":" + umis(0) + "-" + paired.higherReadUmiPrefix + ":" + umis(1)

diff --git a/src/main/scala/com/fulcrumgenomics/umi/Umis.scala b/src/main/scala/com/fulcrumgenomics/umi/Umis.scala
@@ -27,13 +27,8 @@ package com.fulcrumgenomics.umi
 
 import com.fulcrumgenomics.bam.api.SamRecord
 
-import java.util.regex.Pattern
-
 object Umis {
 
-  /** The separator pattern for concatenated UMIs. For example "ACGT-GTAA". */
-  val UmiSeparatorPattern: Pattern = Pattern.compile("-")
-
   /** Copies the UMI sequence from the read name.
     *
     * The read name is split by the given name delimiter, and the last field is assumed to be the UMI sequence.  The UMI