zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #934 from broadinstitute/ks_cat_variants_extensions 

CatVariants now allows different input / output file types.
This commit is contained in:
kshakir 2015-04-15 11:44:56 -03:00
parent 6adf91d49f 90b579c78e
commit 8329fe46d1
2 changed files with 61 additions and 47 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
public/gatk-tools-public/src/main/java/org/broadinstitute/gatk/tools/CatVariants.java
View File

@ -147,37 +147,30 @@ public class CatVariants extends CommandLineProgram {
INVALID
}
private FileType fileExtensionCheck(File inFile, File outFile) {
private FileType fileExtensionCheck(File inFile, FileType previousFileType) {
final String inFileName = inFile.toString().toLowerCase();
final String outFileName = outFile.toString().toLowerCase();
FileType inFileType = FileType.INVALID;
if (inFileName.endsWith(".vcf")) {
inFileType = FileType.VCF;
if (outFileName.endsWith(".vcf"))
return inFileType;
if (previousFileType == FileType.VCF || previousFileType == null) {
return FileType.VCF;
}
}
if (inFileName.endsWith(".bcf")) {
inFileType = FileType.BCF;
if (outFileName.endsWith(".bcf"))
return inFileType;
if (previousFileType == FileType.BCF || previousFileType == null) {
return FileType.BCF;
}
}
for (String extension : AbstractFeatureReader.BLOCK_COMPRESSED_EXTENSIONS) {
if (inFileName.endsWith(".vcf" + extension)) {
inFileType = FileType.BLOCK_COMPRESSED_VCF;
if (outFileName.endsWith(".vcf" + extension))
return inFileType;
if (previousFileType == FileType.BLOCK_COMPRESSED_VCF || previousFileType == null) {
return FileType.BLOCK_COMPRESSED_VCF;
}
}
}
if (inFileType == FileType.INVALID)
System.err.println(String.format("File extension for input file %s is not valid for CatVariants", inFile));
else
System.err.println(String.format("File extension mismatch between input %s and output %s", inFile, outFile));
System.err.println(String.format("File extension for input file %s is not valid for CatVariants", inFile));
printUsage();
return FileType.INVALID;
}
@ -241,10 +234,10 @@ public class CatVariants extends CommandLineProgram {
else
priorityQueue = new PriorityQueue<>(10000, positionComparator);
FileType fileType = FileType.INVALID;
FileType fileType = null;
for (File file : variant) {
// if it returns a valid type, it will be the same for all files
fileType = fileExtensionCheck(file, outputFile);
fileType = fileExtensionCheck(file, fileType);
if (fileType == FileType.INVALID)
return 1;

75
public/gatk-tools-public/src/test/java/org/broadinstitute/gatk/tools/CatVariantsIntegrationTest.java
View File

@ -33,6 +33,7 @@ import org.broadinstitute.gatk.engine.GATKVCFUtils;
import org.broadinstitute.gatk.utils.BaseTest;
import org.broadinstitute.gatk.utils.MD5DB;
import org.broadinstitute.gatk.utils.MD5Mismatch;
import org.broadinstitute.gatk.utils.Utils;
import org.broadinstitute.gatk.utils.runtime.ProcessController;
import org.broadinstitute.gatk.utils.runtime.ProcessSettings;
import org.broadinstitute.gatk.utils.runtime.RuntimeUtils;
@ -44,6 +45,10 @@ import org.testng.annotations.Test;
import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.LinkedHashSet;
import java.util.List;
import java.util.Set;
public class CatVariantsIntegrationTest {
private final MD5DB md5db = new MD5DB();
@ -69,13 +74,13 @@ public class CatVariantsIntegrationTest {
}
public final String getCmdLine() {
return String.format("java -cp %s %s -R %s -V %s -V %s -out %s",
return String.format("java -cp \"%s\" %s -R %s -V %s -V %s -out %s",
StringUtils.join(RuntimeUtils.getAbsoluteClassPaths(), File.pathSeparatorChar),
CatVariants.class.getCanonicalName(), BaseTest.b37KGReference, file1, file2, outputFile);
}
public String toString() {
return "CatVariantsTestProvider " + outputFile;
return String.format("CatVariantsTestProvider %s + %s -> %s", file1.getName(), file2.getName(), outputFile.getName());
}
}
@ -87,17 +92,22 @@ public class CatVariantsIntegrationTest {
new CatVariantsTestProvider(CatVariantsVcf1, CatVariantsVcf2, BaseTest.createTempFile("CatVariantsTest", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
new CatVariantsTestProvider(CatVariantsBcf1, CatVariantsBcf2, BaseTest.createTempFile("CatVariantsTest", ".bcf"), "6a57fcbbf3cae490896d13a288670d83");
for (String extension : AbstractFeatureReader.BLOCK_COMPRESSED_EXTENSIONS) {
final File file1 = new File(CatVariantsDir, "CatVariantsTest1.vcf" + extension);
final File file2 = new File(CatVariantsDir, "CatVariantsTest2.vcf" + extension);
final File outputFile = BaseTest.createTempFile("CatVariantsTest", ".vcf" + extension);
new CatVariantsTestProvider(file1, file2, outputFile, "33f728ac5c70ce2994f3619a27f47088");
for (String extension1 : AbstractFeatureReader.BLOCK_COMPRESSED_EXTENSIONS) {
for (String extension2 : AbstractFeatureReader.BLOCK_COMPRESSED_EXTENSIONS) {
final File file1 = new File(CatVariantsDir, "CatVariantsTest1.vcf" + extension1);
final File file2 = new File(CatVariantsDir, "CatVariantsTest2.vcf" + extension2);
new CatVariantsTestProvider(file1, file2, BaseTest.createTempFile("CatVariantsTest.", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
new CatVariantsTestProvider(file1, file2, BaseTest.createTempFile("CatVariantsTest.", ".bcf"), "6a57fcbbf3cae490896d13a288670d83");
new CatVariantsTestProvider(file1, file2, BaseTest.createTempFile("CatVariantsTest.", ".vcf" + extension1), "33f728ac5c70ce2994f3619a27f47088");
}
new CatVariantsTestProvider(CatVariantsVcf1, CatVariantsVcf2, BaseTest.createTempFile("CatVariantsTest.", ".vcf" + extension1), "33f728ac5c70ce2994f3619a27f47088");
new CatVariantsTestProvider(CatVariantsBcf1, CatVariantsBcf2, BaseTest.createTempFile("CatVariantsTest.", ".vcf" + extension1), "f1a55575f59707f80b8c17e2591fbf53");
}
//Test list parsing functionality
new CatVariantsTestProvider(catVariantsTempList1, CatVariantsVcf2, BaseTest.createTempFile("CatVariantsTest", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
new CatVariantsTestProvider(CatVariantsVcf1, catVariantsTempList2, BaseTest.createTempFile("CatVariantsTest", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
new CatVariantsTestProvider(catVariantsTempList1, catVariantsTempList2, BaseTest.createTempFile("CatVariantsTest", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
new CatVariantsTestProvider(catVariantsTempList1, CatVariantsVcf2, BaseTest.createTempFile("CatVariantsTest.", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
new CatVariantsTestProvider(CatVariantsVcf1, catVariantsTempList2, BaseTest.createTempFile("CatVariantsTest.", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
new CatVariantsTestProvider(catVariantsTempList1, catVariantsTempList2, BaseTest.createTempFile("CatVariantsTest.", ".vcf"), "d0d81eb7fd3905256c4ac7c0fc480094");
return CatVariantsTestProvider.getTests(CatVariantsTestProvider.class);
}
@ -106,7 +116,7 @@ public class CatVariantsIntegrationTest {
public void testExtensions(final CatVariantsTestProvider cfg) throws IOException {
ProcessController pc = ProcessController.getThreadLocal();
ProcessSettings ps = new ProcessSettings(cfg.getCmdLine().split("\\s+"));
ProcessSettings ps = new ProcessSettings(Utils.escapeExpressions(cfg.getCmdLine()));
pc.execAndCheck(ps);
MD5DB.MD5Match result = md5db.testFileMD5("testExtensions", "CatVariantsTestProvider", cfg.outputFile, cfg.md5, false);
@ -116,35 +126,46 @@ public class CatVariantsIntegrationTest {
}
}
@Test(expectedExceptions = IOException.class)
public void testMismatchedExtensions1() throws IOException {
@DataProvider(name = "MismatchedExtensionsTest")
public Object[][] makeMismatchedExtensionsTestProvider() {
return new Object[][]{
{".vcf", ".vcf.gz"},
{".vcf.gz", ".vcf"},
{".bcf", ".vcf.gz"},
{".vcf.gz", ".bcf"},
{".vcf", ".bcf"},
{".bcf", ".vcf"}
};
}
String cmdLine = String.format("java -cp %s %s -R %s -V %s -V %s -out %s",
@Test(dataProvider = "MismatchedExtensionsTest", expectedExceptions = IOException.class)
public void testMismatchedExtensions1(final String extension1, final String extension2) throws IOException {
String cmdLine = String.format("java -cp \"%s\" %s -R %s -V %s -V %s -out %s",
StringUtils.join(RuntimeUtils.getAbsoluteClassPaths(), File.pathSeparatorChar),
CatVariants.class.getCanonicalName(),
BaseTest.b37KGReference,
CatVariantsVcf1,
CatVariantsVcf2,
new File(CatVariantsDir, "CatVariantsTest1" + extension1),
new File(CatVariantsDir, "CatVariantsTest2" + extension2),
BaseTest.createTempFile("CatVariantsTest", ".bcf"));
ProcessController pc = ProcessController.getThreadLocal();
ProcessSettings ps = new ProcessSettings(cmdLine.split("\\s+"));
ProcessSettings ps = new ProcessSettings(Utils.escapeExpressions(cmdLine));
pc.execAndCheck(ps);
}
@Test(expectedExceptions = IOException.class)
public void testMismatchedExtensions2() throws IOException {
@Test(dataProvider = "MismatchedExtensionsTest", expectedExceptions = IOException.class)
public void testMismatchedExtensions2(final String extension1, final String extension2) throws IOException {
String cmdLine = String.format("java -cp %s %s -R %s -V %s -V %s -out %s",
String cmdLine = String.format("java -cp \"%s\" %s -R %s -V %s -V %s -out %s",
StringUtils.join(RuntimeUtils.getAbsoluteClassPaths(), File.pathSeparatorChar),
CatVariants.class.getCanonicalName(),
BaseTest.b37KGReference,
CatVariantsVcf1,
CatVariantsBcf2,
new File(CatVariantsDir, "CatVariantsTest1" + extension1),
new File(CatVariantsDir, "CatVariantsTest2" + extension2),
BaseTest.createTempFile("CatVariantsTest", ".vcf"));
ProcessController pc = ProcessController.getThreadLocal();
ProcessSettings ps = new ProcessSettings(cmdLine.split("\\s+"));
ProcessSettings ps = new ProcessSettings(Utils.escapeExpressions(cmdLine));
pc.execAndCheck(ps);
}
@ -195,7 +216,7 @@ public class CatVariantsIntegrationTest {
@Test(dataProvider = "IndexDataProvider")
public void testCatVariantsVCFIndexCreation(VCFIndexCreatorTest testSpec) throws IOException{
String cmdLine = String.format("java -cp %s %s -R %s -V %s -V %s --variant_index_type %s --variant_index_parameter %s -out %s",
String cmdLine = String.format("java -cp \"%s\" %s -R %s -V %s -V %s --variant_index_type %s --variant_index_parameter %s -out %s",
StringUtils.join(RuntimeUtils.getAbsoluteClassPaths(), File.pathSeparatorChar),
CatVariants.class.getCanonicalName(),
BaseTest.b37KGReference,
@ -206,14 +227,14 @@ public class CatVariantsIntegrationTest {
BaseTest.createTempFile("CatVariantsVCFIndexCreationTest", ".vcf"));
ProcessController pc = ProcessController.getThreadLocal();
ProcessSettings ps = new ProcessSettings(cmdLine.split("\\s+"));
ProcessSettings ps = new ProcessSettings(Utils.escapeExpressions(cmdLine));
pc.execAndCheck(ps);
}
@Test()
public void testCatVariantsGVCFIndexCreation() throws IOException{
String cmdLine = String.format("java -cp %s %s -R %s -V %s -V %s -out %s",
String cmdLine = String.format("java -cp \"%s\" %s -R %s -V %s -V %s -out %s",
StringUtils.join(RuntimeUtils.getAbsoluteClassPaths(), File.pathSeparatorChar),
CatVariants.class.getCanonicalName(),
BaseTest.b37KGReference,
@ -222,7 +243,7 @@ public class CatVariantsIntegrationTest {
BaseTest.createTempFile("CatVariantsGVCFIndexCreationTest", "." + GATKVCFUtils.GVCF_EXT));
ProcessController pc = ProcessController.getThreadLocal();
ProcessSettings ps = new ProcessSettings(cmdLine.split("\\s+"));
ProcessSettings ps = new ProcessSettings(Utils.escapeExpressions(cmdLine));
pc.execAndCheck(ps);
}
}