[med-svn] [Git][med-team/bbmap][upstream] New upstream version 38.90+dfsg
Nilesh Patra
gitlab at salsa.debian.org
Fri Feb 5 10:37:29 GMT 2021
Nilesh Patra pushed to branch upstream at Debian Med / bbmap
Commits:
ca9d9345 by Nilesh Patra at 2021-02-05T15:48:14+05:30
New upstream version 38.90+dfsg
- - - - -
18 changed files:
- README.md
- current/assemble/BubblePopper.java
- current/clump/KmerReduce.java
- current/consensus/BaseGraph.java
- current/dna/AminoAcid.java
- current/dna/MotifProbsN.java
- current/jgi/AdjustHomopolymers.java
- current/jgi/RQCFilter2.java
- current/shared/Shared.java
- current/sketch/Sketch.java
- current/stream/FASTQ.java
- current/stream/Read.java
- current/structures/ByteBuilder.java
- current/structures/Quantizer.java
- docs/changelog.txt
- kmutate.sh
- rqcfilter2.sh
- tadpole.sh
Changes:
=====================================
README.md
=====================================
@@ -1,6 +1,6 @@
# BBTools bioinformatics tools, including BBMap.
-# Author: Brian Bushnell, Jon Rood, Shijie Yao
+# Author: Brian Bushnell, Jon Rood, Shijie Yao, Jasper Toscani Field
# Language: Java, Bash
# Information about documentation is in /docs/readme.txt.
-# Version 38.89
+# Version 38.90
=====================================
current/assemble/BubblePopper.java
=====================================
@@ -636,9 +636,13 @@ public class BubblePopper {
}
}
}
- if(verbose){System.err.println("Mid nodes concur.");}
- assert(leftDest<0 || leftDest==center.id); //TODO: This triggered once nondeterministially
+ if(leftDest>=0 && leftDest!=center.id){return false;}//workaround for actual assertion failure
+ assert(leftDest<0 || leftDest==center.id) :
+ leftDest+", "+center.id; //TODO: This triggered once nondeterministially; determine why
+ if(rightDest>=0 && rightDest!=center.id){return false;}//workaround for potential assertion failure
assert(rightDest<0 || rightDest==dest.id);
+
+ if(verbose){System.err.println("Mid nodes concur.");}
return leftDest>=0 && rightDest>=0;
}
=====================================
current/clump/KmerReduce.java
=====================================
@@ -12,6 +12,7 @@ import fileIO.FileFormat;
import fileIO.ReadWrite;
import jgi.BBMerge;
import kmer.KmerTableSet;
+import shared.KillSwitch;
import shared.Parse;
import shared.Parser;
import shared.PreParser;
@@ -351,7 +352,7 @@ public class KmerReduce {
/*--------------------------------------------------------------*/
public byte[] toBytes(final long kmer){
- byte[] dest=new byte[k];
+ byte[] dest=KillSwitch.allocByte1D(k);
fill(kmer, dest, 0);
return dest;
}
=====================================
current/consensus/BaseGraph.java
=====================================
@@ -8,6 +8,7 @@ import aligner.AlignmentResult;
import aligner.FlatAligner2;
import dna.AminoAcid;
import prok.GeneCaller;
+import shared.KillSwitch;
import shared.Tools;
import shared.TrimRead;
import stream.FASTQ;
@@ -606,7 +607,7 @@ public class BaseGraph extends ConsensusObject implements Serializable {
insCount=0;
// System.err.println("rpos\tdw\trw\tiw");
- final byte[] temp=new byte[2];
+ final byte[] temp=KillSwitch.allocByte1D(2);
for(int i=0; i<ref.length; i++) {
final BaseNode dnode=del[i];
=====================================
current/dna/AminoAcid.java
=====================================
@@ -3,6 +3,7 @@ import java.util.Arrays;
import java.util.HashMap;
import align2.QualityTools;
+import shared.KillSwitch;
import shared.Tools;
import structures.ByteBuilder;
@@ -433,7 +434,7 @@ public final class AminoAcid {
public static final byte[] numberToBases(int code, int n){
- byte[] bytes=new byte[n];
+ byte[] bytes=KillSwitch.allocByte1D(n);
for(int i=n-1; i>=0; i--){
int temp=code&3;
@@ -642,7 +643,7 @@ public final class AminoAcid {
final int stop=frame+blen;
final int alen=blen/3;
- byte[] out=new byte[alen];
+ byte[] out=KillSwitch.allocByte1D(alen);
for(int i=2+frame, j=0; i<stop; i+=3, j++){
byte a=toByte(bases[i-2], bases[i-1], bases[i]);
out[j]=a;
@@ -656,7 +657,7 @@ public final class AminoAcid {
final int blen=stop-start;
final int alen=blen/3;
- byte[] out=new byte[alen];
+ byte[] out=KillSwitch.allocByte1D(alen);
for(int i=2+start, j=0; i<stop; i+=3, j++){
byte a=toByte(bases[i-2], bases[i-1], bases[i]);
out[j]=a;
@@ -672,7 +673,7 @@ public final class AminoAcid {
final int stop=frame+blen;
final int alen=blen/3;
- byte[] out=new byte[alen];
+ byte[] out=KillSwitch.allocByte1D(alen);
for(int i=2+frame, j=0; i<stop; i+=3, j++){
byte qa=quals[i-2], qb=quals[i-1], qc=quals[i];
float pa=QualityTools.PROB_CORRECT[qa], pb=QualityTools.PROB_CORRECT[qb], pc=QualityTools.PROB_CORRECT[qc];
@@ -694,7 +695,7 @@ public final class AminoAcid {
final int alen=aminos.length;
final int blen=alen*3;
- final byte[] out=new byte[blen];
+ final byte[] out=KillSwitch.allocByte1D(blen);
for(int i=0, j=0; i<alen; i++, j+=3){
int code=aminoToCode[aminos[i]];
out[j+2]=numberToBase[(code&3)];
=====================================
current/dna/MotifProbsN.java
=====================================
@@ -2,6 +2,8 @@ package dna;
import java.util.Locale;
+import shared.KillSwitch;
+
public class MotifProbsN extends Motif {
public static void main(String args[]){
@@ -65,7 +67,7 @@ public class MotifProbsN extends Motif {
super(name_, p.length, cen);
N=n;
- chunk=new byte[N];
+ chunk=KillSwitch.allocByte1D(N);
baseProb=Motif.baseProbN[N];
probs=p;
=====================================
current/jgi/AdjustHomopolymers.java
=====================================
@@ -9,6 +9,7 @@ import dna.AminoAcid;
import fileIO.ByteFile;
import fileIO.FileFormat;
import fileIO.ReadWrite;
+import shared.KillSwitch;
import shared.Parse;
import shared.Parser;
import shared.PreParser;
@@ -440,7 +441,7 @@ public class AdjustHomopolymers {
private byte[] fakeQuality(int minlen){
if(fakeQuality.length<minlen){
- fakeQuality=new byte[minlen+10];
+ fakeQuality=KillSwitch.allocByte1D(minlen+10);
Arrays.fill(fakeQuality, Shared.FAKE_QUAL);
}
return fakeQuality;
=====================================
current/jgi/RQCFilter2.java
=====================================
@@ -2455,6 +2455,7 @@ public class RQCFilter2 {
final boolean useBBSplit=mouseCatDogHuman;
{
+ if(deterministic){argList.add("deterministic");}
argList.add("ordered="+(ordered && false)); //Ordered is too slow here
argList.add("k="+map_k);
argList.add("idtag=t");
@@ -2742,6 +2743,7 @@ public class RQCFilter2 {
ArrayList<String> argList=new ArrayList<String>();
{
+ if(deterministic){argList.add("deterministic");}
argList.add("k="+map_k);
argList.add("idtag=t");
argList.add("printunmappedcount");
@@ -2866,6 +2868,7 @@ public class RQCFilter2 {
ArrayList<String> argList=new ArrayList<String>();
{
+ if(deterministic){argList.add("deterministic");}
if(ordered){argList.add("ordered");}
argList.add("quickmatch");
argList.add("k="+map_k);
@@ -2999,6 +3002,7 @@ public class RQCFilter2 {
final String outPre=(outPrefix==null ? outDir : (tmpDir==null ? outDir : tmpDir)+outPrefix);
{
+ if(deterministic){argList.add("deterministic");}
if(ordered){argList.add("ordered");}
argList.add("minratio=.9");
argList.add("maxindel=3");
@@ -3873,9 +3877,11 @@ public class RQCFilter2 {
private String out1=null;
/** Secondary output reads file */
private String out2=null;
-
+
private boolean deleteTemp=true;
+ private boolean deterministic=true;
+
private boolean ordered=false;
private boolean dryrun=false;
=====================================
current/shared/Shared.java
=====================================
@@ -124,8 +124,8 @@ public class Shared {
public static final int GAPCOST=Tools.max(1, GAPLEN/64);
public static final byte GAPC='-';
- public static String BBMAP_VERSION_STRING="38.89";
- public static String BBMAP_VERSION_NAME="Kmer Sets";
+ public static String BBMAP_VERSION_STRING="38.90";
+ public static String BBMAP_VERSION_NAME="Byte Safety";
public static boolean TRIM_READ_COMMENTS=false;
public static boolean TRIM_RNAME=false; //For mapped sam reads
=====================================
current/sketch/Sketch.java
=====================================
@@ -9,6 +9,7 @@ import java.util.concurrent.atomic.AtomicInteger;
import dna.AminoAcid;
import fileIO.ReadWrite;
+import shared.KillSwitch;
import shared.Tools;
import structures.AbstractBitSet;
import structures.ByteBuilder;
@@ -699,7 +700,7 @@ public class Sketch extends SketchObject implements Comparable<Sketch>, Cloneabl
toHeader(bb);
bb.append("\n");
byte[] temp=null;
- if(CODING==A48){temp=new byte[12];}
+ if(CODING==A48){temp=KillSwitch.allocByte1D(12);}
for(int i=0; i<keys.length; i++){
long key=keys[i];
int count=(keyCounts==null ? 1 : keyCounts[i]);
@@ -729,7 +730,7 @@ public class Sketch extends SketchObject implements Comparable<Sketch>, Cloneabl
long prev=0;
toHeader(bb);
bb.append("\n");
- final byte[] temp=new byte[12];
+ final byte[] temp=KillSwitch.allocByte1D(12);
if(keyCounts==null){
for(int i=0; i<keys.length; i++){
=====================================
current/stream/FASTQ.java
=====================================
@@ -905,7 +905,7 @@ public class FASTQ {
}
public static String qualToString(byte[] quals){
- byte[] q2=new byte[quals.length];
+ byte[] q2=KillSwitch.allocByte1D(quals.length);
for(int i=0; i<quals.length; i++){
q2[i]=(byte)(quals[i]+ASCII_OFFSET);
}
=====================================
current/stream/Read.java
=====================================
@@ -3875,14 +3875,14 @@ public final class Read implements Comparable<Read>, Cloneable, Serializable{
public static byte[] getFakeQuality(int len){
if(len>=QUALCACHE.length){
- byte[] r=new byte[len];
+ byte[] r=KillSwitch.allocByte1D(len);
Arrays.fill(r, (byte)30);
return r;
}
if(QUALCACHE[len]==null){
synchronized(QUALCACHE){
if(QUALCACHE[len]==null){
- QUALCACHE[len]=new byte[len];
+ QUALCACHE[len]=KillSwitch.allocByte1D(len);
Arrays.fill(QUALCACHE[len], (byte)30);
}
}
=====================================
current/structures/ByteBuilder.java
=====================================
@@ -20,17 +20,17 @@ public final class ByteBuilder implements Serializable, CharSequence {
private static final long serialVersionUID = -4786450129730831665L;
public ByteBuilder(){
- array=new byte[32];
+ array=KillSwitch.allocByte1D(32);
}
public ByteBuilder(int initial){
assert(initial>=1);
- array=new byte[initial];
+ array=KillSwitch.allocByte1D(initial);
}
public ByteBuilder(Object o){
String s=o.toString();
- array=new byte[s.length()+1];
+ array=KillSwitch.allocByte1D(s.length()+1);
append(s);
}
@@ -560,7 +560,7 @@ public final class ByteBuilder implements Serializable, CharSequence {
if(array.length<=maxLen){return;}
// assert(length<=maxLen) : length+", "+array.length+", "+maxLen;
if(length<1){
- array=new byte[maxLen];
+ array=KillSwitch.allocByte1D(maxLen);
}else{
array=KillSwitch.copyOf(array, maxLen);
length=Tools.min(length, maxLen);
@@ -645,7 +645,7 @@ public final class ByteBuilder implements Serializable, CharSequence {
public byte[] array;
public int length=0;
- private final byte[] numbuffer=new byte[40];
+ private final byte[] numbuffer=KillSwitch.allocByte1D(40);
public static final byte[] numbers=new byte[] {'0', '1', '2', '3', '4', '5', '6', '7', '8', '9'};
public static final byte[] nullBytes="null".getBytes();
=====================================
current/structures/Quantizer.java
=====================================
@@ -1,5 +1,6 @@
package structures;
+import shared.KillSwitch;
import shared.Parse;
import shared.Tools;
import stream.Read;
@@ -64,7 +65,7 @@ public class Quantizer {
}
private static final byte[] makeQualityRemapArray(byte[] quantizeArray) {
- byte[] array=new byte[128];
+ byte[] array=KillSwitch.allocByte1D(128);
for(int i=0; i<array.length; i++){
byte q=0;
for(byte x : quantizeArray){
=====================================
docs/changelog.txt
=====================================
@@ -890,12 +890,16 @@ Wrote KExpand and kmutate.sh for creating sets of mutant kmers.
Split Parse class off from Tools.
38.89
Added support for degenerate amino acid symbols B, J, Z.
+38.90
+Added workaround for BubblePopper 640 assertion failure.
+RQCFilter deterministic mode added for mapping phases.
+Wrapped instances of byte array instantiation in allocByte1D.
TODO: Variant whitelist for callvariants
TODO: filtervcf multiple position interval support.
TODO: refnames flag in seal should clearly indicate it outputs on file per ref file. Also kmers should be unified per reference.
-TODO: BubblePopper 640 assertion fired (shijie).
+TODO: Understand how BubblePopper 640 assertion can occur.
TODO: ApplyVariants fails silently if the header mismatches in certain ways - e.g. "NC123" versus "NC123 virus".
TODO: make universal flag document including e.g. cq (changequality).
TODO: bzip2 fails on JGI cloud, though lbzip2 works. Test bzip2 on Cori.
=====================================
kmutate.sh
=====================================
@@ -5,10 +5,10 @@ echo "
Written by Brian Bushnell
Last modified January 26, 2021
-Description: Generates mutants of kmers in input data, allowing specified
-numbers of substitutions, insertions, and deletions. Intended for
-pattern-matching with BBDuk or Seal, analyzing barcodes and other short
-synthetic oligos, and so forth. Input may be fasta or fastq, compressed or raw.
+Description: Given a reference, generates a kmer spectrum including a
+specified number of substitutions, insertions, and deletions. The output is
+useful for analyzing barcodes or other short oligos, and filtering using
+BBDuk or Seal. Input may be fasta or fastq, compressed or raw.
See also kcompress, kmercountexact, and bbkmerset.
Usage: kmutate.sh in=<file> out=<file> k=<kmer length> edist=<edit distance>
=====================================
rqcfilter2.sh
=====================================
@@ -3,7 +3,7 @@
usage(){
echo "
Written by Brian Bushnell
-Last modified January 26, 2021
+Last modified February 3, 2021
Description: RQCFilter2 is a revised version of RQCFilter that uses a common path for all dependencies.
The dependencies are available at http://portal.nersc.gov/dna/microbial/assembly/bushnell/RQCFilterData.tar
=====================================
tadpole.sh
=====================================
@@ -3,7 +3,7 @@
usage(){
echo "
Written by Brian Bushnell
-Last modified May 28, 2020
+Last modified February 3, 2021
Description: Uses kmer counts to assemble contigs, extend sequences,
or error-correct reads. Tadpole has no upper bound for kmer length,
View it on GitLab: https://salsa.debian.org/med-team/bbmap/-/commit/ca9d9345ae63881d0a4aacd4774e9e9d0589ad28
--
View it on GitLab: https://salsa.debian.org/med-team/bbmap/-/commit/ca9d9345ae63881d0a4aacd4774e9e9d0589ad28
You're receiving this email because of your account on salsa.debian.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/debian-med-commit/attachments/20210205/0a7c8c39/attachment-0001.html>
More information about the debian-med-commit
mailing list