Mercurial > repos > rliterman > csp2

annotate CSP2/CSP2_env/env-d9b9114564458d9d-741b3de822f2aaca6c6caa4325c4afce/opt/bbmap-39.01-1/current/gff/VcfToGff.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 gff;
jpayne@68 2
jpayne@68 3 import java.io.PrintStream;
jpayne@68 4
jpayne@68 5 import fileIO.ByteFile;
jpayne@68 6 import fileIO.ByteStreamWriter;
jpayne@68 7 import fileIO.FileFormat;
jpayne@68 8 import shared.Parser;
jpayne@68 9 import shared.PreParser;
jpayne@68 10 import shared.ReadStats;
jpayne@68 11 import shared.Timer;
jpayne@68 12 import shared.Tools;
jpayne@68 13 import structures.ByteBuilder;
jpayne@68 14 import var2.VCFLine;
jpayne@68 15
jpayne@68 16 /**
jpayne@68 17 * Stripped out of GffLine into independent class.
jpayne@68 18 * @author Brian Bushnell
jpayne@68 19 * @date Sep 12, 2018
jpayne@68 20 *
jpayne@68 21 */
jpayne@68 22 public class VcfToGff {
jpayne@68 23
jpayne@68 24 /** Translates VCF to GFF */
jpayne@68 25 public static void main(String[] args){
jpayne@68 26 Timer t=new Timer();
jpayne@68 27 PrintStream outstream=System.err;
jpayne@68 28 {//Preparse block for help, config files, and outstream
jpayne@68 29 PreParser pp=new PreParser(args, new Object() { }.getClass().getEnclosingClass(), false);
jpayne@68 30 args=pp.args;
jpayne@68 31 outstream=pp.outstream;
jpayne@68 32 t.outstream=outstream;
jpayne@68 33 }
jpayne@68 34
jpayne@68 35 Parser parser=new Parser();
jpayne@68 36 String in=null;
jpayne@68 37 String out=null;
jpayne@68 38 boolean overwrite=true, append=false;
jpayne@68 39
jpayne@68 40 //Parse each argument
jpayne@68 41 for(int i=0; i<args.length; i++){
jpayne@68 42 String arg=args[i];
jpayne@68 43
jpayne@68 44 //Break arguments into their constituent parts, in the form of "a=b"
jpayne@68 45 String[] split=arg.split("=");
jpayne@68 46 String a=split[0].toLowerCase();
jpayne@68 47 String b=split.length>1 ? split[1] : null;
jpayne@68 48
jpayne@68 49 if(a.equals("in") || a.equals("vcf")){
jpayne@68 50 in=b;
jpayne@68 51 }else if(a.equals("out") || a.equals("gff")){
jpayne@68 52 out=b;
jpayne@68 53 }else if(parser.parse(arg, a, b)){
jpayne@68 54 //do nothing
jpayne@68 55 }else if(in==null && b==null && i==0 && Tools.canRead(arg)){
jpayne@68 56 in=arg;
jpayne@68 57 }else if(in==null && b==null && i==1){
jpayne@68 58 out=arg;
jpayne@68 59 }else{
jpayne@68 60 outstream.println("Unknown parameter "+args[i]);
jpayne@68 61 assert(false) : "Unknown parameter "+args[i];
jpayne@68 62 }
jpayne@68 63 }
jpayne@68 64
jpayne@68 65 {//Process parser fields
jpayne@68 66 overwrite=ReadStats.overwrite=parser.overwrite;
jpayne@68 67 append=ReadStats.append=parser.append;
jpayne@68 68 }
jpayne@68 69
jpayne@68 70 //Ensure output files can be written
jpayne@68 71 if(!Tools.testOutputFiles(overwrite, append, false, out)){
jpayne@68 72 outstream.println((out==null)+", "+out);
jpayne@68 73 throw new RuntimeException("\n\noverwrite="+overwrite+"; Can't write to output files "+out+"\n");
jpayne@68 74 }
jpayne@68 75
jpayne@68 76 //Ensure input files can be read
jpayne@68 77 if(!Tools.testInputFiles(false, true, in)){
jpayne@68 78 throw new RuntimeException("\nCan't read some input files.\n");
jpayne@68 79 }
jpayne@68 80
jpayne@68 81 //Ensure that no file was specified multiple times
jpayne@68 82 if(!Tools.testForDuplicateFiles(true, in, out)){
jpayne@68 83 throw new RuntimeException("\nSome file names were specified multiple times.\n");
jpayne@68 84 }
jpayne@68 85
jpayne@68 86 translate(in, out, overwrite, append);
jpayne@68 87 t.stop("Time: \t");
jpayne@68 88 }
jpayne@68 89
jpayne@68 90 /** Translates VCF to GFF */
jpayne@68 91 private static void translate(String in, String out, boolean overwrite, boolean append){
jpayne@68 92 //Create output FileFormat objects
jpayne@68 93 FileFormat ffout=FileFormat.testOutput(out, FileFormat.GFF, "gff", true, overwrite, append, false);
jpayne@68 94
jpayne@68 95 //Create input FileFormat objects
jpayne@68 96 FileFormat ffin=FileFormat.testInput(in, FileFormat.VCF, "vcf", true, true);
jpayne@68 97
jpayne@68 98 ByteFile bf=ByteFile.makeByteFile(ffin);
jpayne@68 99 ByteStreamWriter bsw=null;
jpayne@68 100 if(ffout!=null){
jpayne@68 101 bsw=new ByteStreamWriter(ffout);
jpayne@68 102 bsw.start();
jpayne@68 103 }
jpayne@68 104
jpayne@68 105 ByteBuilder bb=new ByteBuilder(17000);
jpayne@68 106 bb.append("##gff-version 3\n");
jpayne@68 107 String header="#seqid source type start end score strand phase attributes";
jpayne@68 108 for(byte[] line=bf.nextLine(); line!=null; line=bf.nextLine()){
jpayne@68 109 if(line.length>1){
jpayne@68 110 if(line[0]=='#'){
jpayne@68 111 if(Tools.startsWith(line, "##fileformat") || Tools.startsWith(line, "##FORMAT") ||
jpayne@68 112 Tools.startsWith(line, "##INFO") || Tools.startsWith(line, "#CHROM POS")){
jpayne@68 113 //skip
jpayne@68 114 }else{
jpayne@68 115 int i=1;
jpayne@68 116 while(i<line.length && line[i]=='#'){i++;}
jpayne@68 117 i--;
jpayne@68 118 bb.append(line, i, line.length-i);
jpayne@68 119 bb.nl();
jpayne@68 120 }
jpayne@68 121 }else{
jpayne@68 122 if(header!=null){
jpayne@68 123 bb.append(header).append('\n');
jpayne@68 124 header=null;
jpayne@68 125 }
jpayne@68 126 VCFLine vline=new VCFLine(line);
jpayne@68 127 GffLine gline=new GffLine(vline);
jpayne@68 128 gline.appendTo(bb);
jpayne@68 129 bb.nl();
jpayne@68 130 }
jpayne@68 131 }
jpayne@68 132 if(bb.length()>=16384){
jpayne@68 133 if(bsw!=null){
jpayne@68 134 bsw.print(bb);
jpayne@68 135 }
jpayne@68 136 bb.clear();
jpayne@68 137 }
jpayne@68 138 }
jpayne@68 139 if(bb.length()>0){
jpayne@68 140 if(bsw!=null){
jpayne@68 141 bsw.print(bb);
jpayne@68 142 }
jpayne@68 143 bb.clear();
jpayne@68 144 }
jpayne@68 145 bf.close();
jpayne@68 146 if(bsw!=null){bsw.poisonAndWait();}
jpayne@68 147 }
jpayne@68 148
jpayne@68 149 }