Related to #7

CHANGELOG_CURRENT

@@ -1,6 +1,7 @@
 New feature: optional short column names in `export...` action to simplify further data analysis using data table processing libraries like Pandas or R/DataFrames. (`-s` / `--no-spaces` in `exportAlignments` and `exportClones`)
 New export fields: `-defaultAnchorPoints` outputs positions of default anchor point in aligned reads or clonal sequence (this column is added to default output format), `-positionOf` outputs position of specified anchor point, `-lengthOf` outputs lengt of specified gene feature
 Added `V5UTRBeginTrimmed` anchor point, `V5UTR` gene feature renamed to `V5UTRGermline`, trimmed `V5UTR` gene feature added
+Added `--filter-out-of-frames` and `--filter-stops` options in `exportClones`
 minor: some column names in output tab-delimited files slightly changed
 minor: NPE in exportAlignmentsPretty fixed
 minor: New anchor poins added to exportAlignmentsPretty output

src/main/java/com/milaboratory/mixcr/basictypes/Clone.java

@@ -58,7 +58,11 @@ public void setParentCloneSet(CloneSet set) {
     public double getFraction() {
         if (parent == null)
             throw new NullPointerException("Parent not set yet.");
-        return 1.0 * count / parent.getTotalCount();
+        return getFraction(parent.getTotalCount());
+    }
+
+    public double getFraction(long totalCount) {
+        return 1.0 * count / totalCount;
     }
 
     public GeneFeature[] getAssemblingFeatures() {

src/main/java/com/milaboratory/mixcr/basictypes/CloneSet.java

@@ -28,6 +28,7 @@
  */
 package com.milaboratory.mixcr.basictypes;
 
+import cc.redberry.primitives.Filter;
 import com.milaboratory.mixcr.reference.Allele;
 import com.milaboratory.mixcr.reference.AlleleId;
 import com.milaboratory.mixcr.reference.GeneFeature;
@@ -120,4 +121,19 @@ public long getTotalCount() {
     public Iterator<Clone> iterator() {
         return clones.iterator();
     }
+
+    /**
+     * WARNING: in will be destroyed
+     */
+    public static CloneSet transform(CloneSet in, Filter<Clone> filter) {
+        List<Clone> newClones = new ArrayList<>(in.size());
+        for (int i = 0; i < in.size(); ++i) {
+            Clone c = in.get(i);
+            if (filter.accept(c)) {
+                c.parent = null;
+                newClones.add(c);
+            }
+        }
+        return new CloneSet(newClones, in.usedAlleles, in.alignedFeatures, in.assemblingFeatures);
+    }
 }

src/main/java/com/milaboratory/mixcr/cli/ActionExport.java

@@ -42,7 +42,7 @@ protected ActionExport(ActionExportParameters parameters) {
 
     @Override
     public void go(ActionHelper helper) throws Exception {
-        if (parameters.fields) {
+        if (parameters.listFields) {
             helper.getDefaultPrintStream().print(parameters.printFieldsHelp());
             return;
         }

src/main/java/com/milaboratory/mixcr/cli/ActionExportClones.java

@@ -28,17 +28,19 @@
  */
 package com.milaboratory.mixcr.cli;
 
+import cc.redberry.primitives.Filter;
+import com.milaboratory.core.sequence.AminoAcidSequence;
+import com.milaboratory.core.sequence.NSequenceWithQuality;
 import com.milaboratory.mixcr.basictypes.Clone;
 import com.milaboratory.mixcr.basictypes.CloneSet;
 import com.milaboratory.mixcr.basictypes.CloneSetIO;
 import com.milaboratory.mixcr.basictypes.IOUtil;
 import com.milaboratory.mixcr.export.InfoWriter;
+import com.milaboratory.mixcr.reference.GeneFeature;
 import com.milaboratory.mixcr.reference.LociLibraryManager;
 import com.milaboratory.util.CanReportProgressAndStage;
 import com.milaboratory.util.SmartProgressReporter;
 
-import java.io.BufferedInputStream;
-import java.io.FileInputStream;
 import java.io.InputStream;
 import java.util.List;
 
@@ -52,6 +54,8 @@ public void go0() throws Exception {
         try (InputStream inputStream = IOUtil.createIS(parameters.inputFile);
              InfoWriter<Clone> writer = new InfoWriter<>(parameters.outputFile)) {
             CloneSet set = CloneSetIO.read(inputStream, LociLibraryManager.getDefault());
+            if (parameters.filterOutOfFrames || parameters.filterStopCodons)
+                set = CloneSet.transform(set, new CFilter(parameters.filterOutOfFrames, parameters.filterStopCodons));
             writer.attachInfoProviders((List) parameters.exporters);
             ExportClones exportClones = new ExportClones(set, writer);
             SmartProgressReporter.startProgressReport(exportClones);
@@ -65,6 +69,29 @@ public String command() {
         return "exportClones";
     }
 
+    private static final class CFilter implements Filter<Clone> {
+        final boolean filterOutOfFrames, filterStopCodons;
+
+        public CFilter(boolean filterOutOfFrames, boolean filterStopCodons) {
+            this.filterOutOfFrames = filterOutOfFrames;
+            this.filterStopCodons = filterStopCodons;
+        }
+
+        @Override
+        public boolean accept(Clone clone) {
+            if (filterOutOfFrames) {
+                NSequenceWithQuality cdr3 = clone.getFeature(GeneFeature.CDR3);
+                if (cdr3 == null || cdr3.size() % 3 != 0)
+                    return false;
+            }
+            if (filterStopCodons)
+                for (int i = 0; i < clone.numberOfTargets(); i++)
+                    if (AminoAcidSequence.translateFromCenter(clone.getTarget(i).getSequence()).containStops())
+                        return false;
+            return true;
+        }
+    }
+
     private static final class ExportClones implements CanReportProgressAndStage {
         final CloneSet clones;
         final InfoWriter<Clone> writer;

src/main/java/com/milaboratory/mixcr/cli/ActionExportParameters.java

@@ -64,7 +64,7 @@ public ActionExportParameters(Class clazz, String defaultPreset) {
         description[0].add("-h, --help");
         description[1].add("print this help message");
 
-        description[0].add(FIELDS_SHORT + ", " + FIELDS_LONG);
+        description[0].add(LIST_FIELDS_SHORT + ", " + LIST_FIELDS_LONG);
         description[1].add("print available fields to export");
 
         description[0].add(PRESET_SHORT + ", " + PRESET_LONG);
@@ -77,20 +77,28 @@ public ActionExportParameters(Class clazz, String defaultPreset) {
         description[1].add("output short versions of column headers which facilitates analysis with Pandas, " +
                 "R/DataFrames or other data tables processing library.");
 
+        if (clazz.equals(Clone.class)) {
+            description[0].add(FILTER_OUT_OF_FRAMES);
+            description[1].add("exclude out of frames (fractions will be recalculated)");
+
+            description[0].add(FILTER_STOP_CODONS);
+            description[1].add("exclude sequences containing stop codons (fractions will be recalculated)");
+        }
+
         this.helpString =
                 "Usage: export(Type) [options] input_file output_file\n" +
                         "Options:\n" +
-                        Util.printTwoColumns(4, description[0], description[1], 20, 50, 5, "\n") + "\n" +
+                        Util.printTwoColumns(4, description[0], description[1], 23, 50, 5, "\n") + "\n" +
                         "Examples:\n" +
                         "    exportClones -p all -nFeature CDR1 input.clns output.txt\n" +
                         "    exportAlignments -pf params.txt -nFeature CDR1 -dAlignments input.clns output.txt\n";
         description = FieldExtractors.getDescription(clazz);
         this.fieldsHelpString = "Available export fields:\n" + Util.printTwoColumns(
-                description[0], description[1], 20, 50, 5, "\n");
-
+                description[0], description[1], 23, 50, 5, "\n");
     }
 
-    public Boolean fields = false;
+    public Boolean listFields = false;
+    public boolean filterOutOfFrames = false, filterStopCodons = false;
     public String inputFile;
     public String outputFile;
     public ArrayList<FieldExtractor> exporters;
@@ -105,16 +113,23 @@ public String printHelp() {
 
     public final void parseParameters(String[] args) throws ParameterException {
         trim(args);
-        for (String arg : args) {
-            if (arg.equals(FIELDS_SHORT) || arg.equals(FIELDS_LONG)) {
-                fields = true;
-                return;
-            }
-            if (arg.equals("-h") || arg.equals("--help")) {
-                help = true;
-                return;
+        for (String arg : args)
+            switch (arg) {
+                case LIST_FIELDS_SHORT:
+                case LIST_FIELDS_LONG:
+                    listFields = true;
+                    return;
+                case "-h":
+                case "--help":
+                    help = true;
+                    return;
+                case FILTER_OUT_OF_FRAMES:
+                    filterOutOfFrames = true;
+                    break;
+                case FILTER_STOP_CODONS:
+                    filterStopCodons = true;
+                    break;
             }
-        }
 
         if (args.length < 2)
             throw new ParameterException("No output file specified.");
@@ -127,7 +142,10 @@ public final void parseParameters(String[] args) throws ParameterException {
                 break;
             }
 
-        if (args.length == 2 || (outputMode == OutputMode.ScriptingFriendly && args.length == 3))
+        int i = ((outputMode == OutputMode.ScriptingFriendly) ? 1 : 0)
+                + (filterOutOfFrames ? 1 : 0)
+                + (filterStopCodons ? 1 : 0);
+        if (args.length - i == 2)
             exporters = getPresetParameters(outputMode, clazz, defaultPreset);
         else
             exporters = parseParametersString(outputMode, clazz, args, 0, args.length - 2);
@@ -143,6 +161,9 @@ public static ArrayList<FieldExtractor> parseParametersString(OutputMode outputM
             String arg = args[i];
             if (isParsingFriendlyFlag(arg))
                 continue;
+            //skip options
+            if (arg.equals(FILTER_OUT_OF_FRAMES) || arg.equals(FILTER_STOP_CODONS))
+                continue;
             if (arg.charAt(0) == '-') {
                 if (!exporter.isEmpty()) {
                     FieldExtractor exp = FieldExtractors.parse(outputMode,
@@ -258,10 +279,12 @@ private static void trim(String[] args) {
             PRESET_LONG = "--preset",
             PRESET_FILE_SHORT = "-pf",
             PRESET_FILE_LONG = "--presetFile",
-            FIELDS_SHORT = "-l",
-            FIELDS_LONG = "--listFields",
+            LIST_FIELDS_SHORT = "-l",
+            LIST_FIELDS_LONG = "--listFields",
             PARSING_LONG = "--no-spaces",
-            PARSING_SHORT = "-s";
+            PARSING_SHORT = "-s",
+            FILTER_OUT_OF_FRAMES = "--filter-out-of-frames",
+            FILTER_STOP_CODONS = "--filter-stops";
 
     public static boolean isPresetParameter(String string) {
         return string.equals(PRESET_SHORT) || string.equals(PRESET_LONG);

src/main/java/com/milaboratory/mixcr/cli/Util.java

@@ -29,6 +29,7 @@
 package com.milaboratory.mixcr.cli;
 
 import com.milaboratory.mixcr.reference.Locus;
+import gnu.trove.map.hash.TIntObjectHashMap;
 
 import java.io.File;
 import java.io.FileOutputStream;
@@ -156,11 +157,17 @@ public static String printTwoColumns(int offset, List<String> left, List<String>
         return sb.toString();
     }
 
-    public static String spacer(int sep) {
-        StringBuilder sb = new StringBuilder(sep);
-        for (int i = 0; i < sep; ++i)
-            sb.append(" ");
-        return sb.toString();
+    private static TIntObjectHashMap<String> spacesCache = new TIntObjectHashMap<>();
+
+    public static synchronized String spacer(int sep) {
+        String s = spacesCache.get(sep);
+        if (s == null) {
+            StringBuilder sb = new StringBuilder(sep);
+            for (int i = 0; i < sep; ++i)
+                sb.append(" ");
+            spacesCache.put(sep, s = sb.toString());
+        }
+        return s;
     }
 
     private static int lineBreakPos(String str, int width) {