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

planemo upload commit 2e9511a184a1ca667c7be0c6321a36dc4e3d116d
author jpayne
date Tue, 18 Mar 2025 16:23:26 -0400
parents
children
comparison
equal deleted inserted replaced
67:0e9998148a16 68:5028fdace37b
1 package sketch;
2
3 import java.io.PrintStream;
4 import java.util.HashMap;
5
6 import fileIO.FileFormat;
7 import fileIO.ReadWrite;
8 import shared.Shared;
9 import stream.ConcurrentReadInputStream;
10 import stream.FASTQ;
11 import stream.Read;
12 import structures.ListNum;
13 import tax.GiToTaxid;
14 import tax.TaxTree;
15
16 public class SSUMap {
17
18 public static synchronized void load(PrintStream outstream){
19 if(r16SFile!=null && r16SMap==null){
20 r16SMap=load(r16SFile, TaxTree.default16SFile(), outstream);
21 }
22 if(r18SFile!=null && r18SMap==null){
23 r18SMap=load(r18SFile, TaxTree.default18SFile(), outstream);
24 }
25 }
26
27 private static synchronized HashMap<Integer, byte[]> load(String ssuFile, String defaultFile, PrintStream outstream){
28 HashMap<Integer, byte[]> map=null;
29 if(ssuFile!=null){
30 final boolean oldAminoIn=Shared.AMINO_IN;
31 final boolean oldInterleaved=FASTQ.FORCE_INTERLEAVED;
32 Shared.AMINO_IN=false;//SSUs are nucleotide, which can cause a crash, esp. with IUPAC symbols
33 FASTQ.FORCE_INTERLEAVED=false;
34
35 String fname=ssuFile;
36 if("auto".equalsIgnoreCase(fname)){fname=defaultFile;}
37 final FileFormat ffssu=FileFormat.testInput(fname, FileFormat.FA, null, true, false);
38 map=loadSSU(ffssu, outstream);
39
40 Shared.AMINO_IN=oldAminoIn;
41 FASTQ.FORCE_INTERLEAVED=oldInterleaved;
42 }
43 return map;
44 }
45
46 private static HashMap<Integer, byte[]> loadSSU(FileFormat ff, PrintStream outstream){
47 ConcurrentReadInputStream cris=makeCris(ff, outstream);
48 HashMap<Integer, byte[]> map=new HashMap<Integer, byte[]>(1000000);
49
50 //Grab the first ListNum of reads
51 ListNum<Read> ln=cris.nextList();
52
53 //Check to ensure pairing is as expected
54 if(ln!=null && !ln.isEmpty()){
55 // if(verbose){outstream.println("Fetched "+ln.size()+" reads.");}
56 Read r=ln.get(0);
57 assert(ff.samOrBam() || (r.mate!=null)==cris.paired());
58 }
59
60 //As long as there is a nonempty read list...
61 while(ln!=null && ln.size()>0){
62 if(verbose){outstream.println("Fetched "+ln.size()+" reads.");}
63 for(Read r : ln){
64 final int tid=GiToTaxid.getID(r.id);
65 if(tid>=0 && r.length()>1000){
66 byte[] old=map.get(tid);
67 if(old==null || old.length<r.length()){map.put(tid, r.bases);}
68 }
69 }
70 cris.returnList(ln.id, ln.list==null || ln.list.isEmpty());
71
72 //Fetch a new list
73 ln=cris.nextList();
74 }
75
76 //Notify the input stream that the final list was used
77 if(ln!=null){
78 cris.returnList(ln.id, ln.list==null || ln.list.isEmpty());
79 }
80 errorState|=ReadWrite.closeStream(cris);
81
82 return map;
83 }
84
85 private static ConcurrentReadInputStream makeCris(FileFormat ff, PrintStream outstream){
86 if(verbose){outstream.println("makeCris");}
87 ConcurrentReadInputStream cris=ConcurrentReadInputStream.getReadInputStream(-1, false, ff, null);
88 cris.start(); //Start the stream
89 if(verbose){outstream.println("Loading "+ff.name());}
90 boolean paired=cris.paired();
91 assert(!paired);
92 // if(!ffin1.samOrBam()){outstream.println("Input is being processed as "+(paired ? "paired" : "unpaired"));}
93 return cris;
94 }
95
96 public static boolean hasMap(){return r16SMap!=null || r18SMap!=null;}
97 public static int r16SCount(){return r16SMap==null ? 0 : r16SMap.size();}
98 public static int r18SCount(){return r18SMap==null ? 0 : r18SMap.size();}
99
100 public static String r16SFile=null;
101 public static String r18SFile=null;
102 public static HashMap<Integer, byte[]> r16SMap=null;
103 public static HashMap<Integer, byte[]> r18SMap=null;
104 static boolean verbose=false;
105 static boolean errorState=false;
106
107 }