Mercurial > repos > rliterman > csp2

annotate CSP2/CSP2_env/env-d9b9114564458d9d-741b3de822f2aaca6c6caa4325c4afce/opt/bbmap-39.01-1/current/sketch/SSUMap.java @ 68:5028fdace37b

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 2e9511a184a1ca667c7be0c6321a36dc4e3d116d
author jpayne
date Tue, 18 Mar 2025 16:23:26 -0400
parents
children
rev   line source
jpayne@68 1 package sketch;
jpayne@68 2
jpayne@68 3 import java.io.PrintStream;
jpayne@68 4 import java.util.HashMap;
jpayne@68 5
jpayne@68 6 import fileIO.FileFormat;
jpayne@68 7 import fileIO.ReadWrite;
jpayne@68 8 import shared.Shared;
jpayne@68 9 import stream.ConcurrentReadInputStream;
jpayne@68 10 import stream.FASTQ;
jpayne@68 11 import stream.Read;
jpayne@68 12 import structures.ListNum;
jpayne@68 13 import tax.GiToTaxid;
jpayne@68 14 import tax.TaxTree;
jpayne@68 15
jpayne@68 16 public class SSUMap {
jpayne@68 17
jpayne@68 18 public static synchronized void load(PrintStream outstream){
jpayne@68 19 if(r16SFile!=null && r16SMap==null){
jpayne@68 20 r16SMap=load(r16SFile, TaxTree.default16SFile(), outstream);
jpayne@68 21 }
jpayne@68 22 if(r18SFile!=null && r18SMap==null){
jpayne@68 23 r18SMap=load(r18SFile, TaxTree.default18SFile(), outstream);
jpayne@68 24 }
jpayne@68 25 }
jpayne@68 26
jpayne@68 27 private static synchronized HashMap<Integer, byte[]> load(String ssuFile, String defaultFile, PrintStream outstream){
jpayne@68 28 HashMap<Integer, byte[]> map=null;
jpayne@68 29 if(ssuFile!=null){
jpayne@68 30 final boolean oldAminoIn=Shared.AMINO_IN;
jpayne@68 31 final boolean oldInterleaved=FASTQ.FORCE_INTERLEAVED;
jpayne@68 32 Shared.AMINO_IN=false;//SSUs are nucleotide, which can cause a crash, esp. with IUPAC symbols
jpayne@68 33 FASTQ.FORCE_INTERLEAVED=false;
jpayne@68 34
jpayne@68 35 String fname=ssuFile;
jpayne@68 36 if("auto".equalsIgnoreCase(fname)){fname=defaultFile;}
jpayne@68 37 final FileFormat ffssu=FileFormat.testInput(fname, FileFormat.FA, null, true, false);
jpayne@68 38 map=loadSSU(ffssu, outstream);
jpayne@68 39
jpayne@68 40 Shared.AMINO_IN=oldAminoIn;
jpayne@68 41 FASTQ.FORCE_INTERLEAVED=oldInterleaved;
jpayne@68 42 }
jpayne@68 43 return map;
jpayne@68 44 }
jpayne@68 45
jpayne@68 46 private static HashMap<Integer, byte[]> loadSSU(FileFormat ff, PrintStream outstream){
jpayne@68 47 ConcurrentReadInputStream cris=makeCris(ff, outstream);
jpayne@68 48 HashMap<Integer, byte[]> map=new HashMap<Integer, byte[]>(1000000);
jpayne@68 49
jpayne@68 50 //Grab the first ListNum of reads
jpayne@68 51 ListNum<Read> ln=cris.nextList();
jpayne@68 52
jpayne@68 53 //Check to ensure pairing is as expected
jpayne@68 54 if(ln!=null && !ln.isEmpty()){
jpayne@68 55 // if(verbose){outstream.println("Fetched "+ln.size()+" reads.");}
jpayne@68 56 Read r=ln.get(0);
jpayne@68 57 assert(ff.samOrBam() || (r.mate!=null)==cris.paired());
jpayne@68 58 }
jpayne@68 59
jpayne@68 60 //As long as there is a nonempty read list...
jpayne@68 61 while(ln!=null && ln.size()>0){
jpayne@68 62 if(verbose){outstream.println("Fetched "+ln.size()+" reads.");}
jpayne@68 63 for(Read r : ln){
jpayne@68 64 final int tid=GiToTaxid.getID(r.id);
jpayne@68 65 if(tid>=0 && r.length()>1000){
jpayne@68 66 byte[] old=map.get(tid);
jpayne@68 67 if(old==null || old.length<r.length()){map.put(tid, r.bases);}
jpayne@68 68 }
jpayne@68 69 }
jpayne@68 70 cris.returnList(ln.id, ln.list==null || ln.list.isEmpty());
jpayne@68 71
jpayne@68 72 //Fetch a new list
jpayne@68 73 ln=cris.nextList();
jpayne@68 74 }
jpayne@68 75
jpayne@68 76 //Notify the input stream that the final list was used
jpayne@68 77 if(ln!=null){
jpayne@68 78 cris.returnList(ln.id, ln.list==null || ln.list.isEmpty());
jpayne@68 79 }
jpayne@68 80 errorState|=ReadWrite.closeStream(cris);
jpayne@68 81
jpayne@68 82 return map;
jpayne@68 83 }
jpayne@68 84
jpayne@68 85 private static ConcurrentReadInputStream makeCris(FileFormat ff, PrintStream outstream){
jpayne@68 86 if(verbose){outstream.println("makeCris");}
jpayne@68 87 ConcurrentReadInputStream cris=ConcurrentReadInputStream.getReadInputStream(-1, false, ff, null);
jpayne@68 88 cris.start(); //Start the stream
jpayne@68 89 if(verbose){outstream.println("Loading "+ff.name());}
jpayne@68 90 boolean paired=cris.paired();
jpayne@68 91 assert(!paired);
jpayne@68 92 // if(!ffin1.samOrBam()){outstream.println("Input is being processed as "+(paired ? "paired" : "unpaired"));}
jpayne@68 93 return cris;
jpayne@68 94 }
jpayne@68 95
jpayne@68 96 public static boolean hasMap(){return r16SMap!=null || r18SMap!=null;}
jpayne@68 97 public static int r16SCount(){return r16SMap==null ? 0 : r16SMap.size();}
jpayne@68 98 public static int r18SCount(){return r18SMap==null ? 0 : r18SMap.size();}
jpayne@68 99
jpayne@68 100 public static String r16SFile=null;
jpayne@68 101 public static String r18SFile=null;
jpayne@68 102 public static HashMap<Integer, byte[]> r16SMap=null;
jpayne@68 103 public static HashMap<Integer, byte[]> r18SMap=null;
jpayne@68 104 static boolean verbose=false;
jpayne@68 105 static boolean errorState=false;
jpayne@68 106
jpayne@68 107 }