package org.autoplot.cdf; import gov.nasa.gsfc.spdf.cdfj.CDFDataType; import gov.nasa.gsfc.spdf.cdfj.CDFException; import gov.nasa.gsfc.spdf.cdfj.CDFReader; import gov.nasa.gsfc.spdf.cdfj.CDFWriter; import gov.nasa.gsfc.spdf.cdfj.ReaderFactory; import gov.nasa.gsfc.spdf.cdfj.SupportedTypes; import java.lang.reflect.Array; import org.autoplot.datasource.DataSourceUtil; import org.das2.datum.Units; import org.das2.datum.UnitsConverter; import org.das2.datum.UnitsUtil; import java.io.File; import java.io.FileNotFoundException; import java.io.IOException; import java.io.OutputStream; import java.nio.ByteBuffer; import java.nio.ByteOrder; import java.util.HashMap; import java.util.Iterator; import java.util.Map; import org.das2.util.monitor.ProgressMonitor; import org.das2.qds.QDataSet; import org.das2.qds.QubeDataSetIterator; import org.autoplot.datasource.URISplit; import org.autoplot.datasource.DataSourceFormat; import java.util.logging.Level; import java.util.logging.Logger; import org.das2.util.LoggerManager; import org.das2.qds.DataSetOps; import org.das2.qds.SemanticOps; import org.das2.qds.examples.Schemes; import org.das2.qds.ops.Ops; import org.das2.util.monitor.NullProgressMonitor; /** * Format the QDataSet into CDF tables, using Nand Lal's library. * Datasets will be assigned names if they don't have a NAME property. * If the append=T parameter is set, then variables should have names. * if the bundle=T parameter is set, then bundles should be unbundled into separate variables. * * @author jbf */ public class CdfDataSourceFormat implements DataSourceFormat { Map names; Map namesRev; private static final Logger logger= LoggerManager.getLogger("apdss.cdf"); public CdfDataSourceFormat() { names= new HashMap<>(); namesRev= new HashMap<>(); } //@Override public boolean streamData(Map params, Iterator data, OutputStream out) throws Exception { return false; } private synchronized String nameFor(QDataSet dep0) { String name= names.get(dep0); if ( name!=null ) { return name; } else { logger.log(Level.FINE, "new variable: {0}", dep0); } name = (String) dep0.property(QDataSet.NAME); if ( namesRev.containsKey(name) ) { int i= 1; while ( namesRev.containsKey(name+"_"+i) ) { i=i+1; } name= name + "_"+ i; } Units units = (Units) dep0.property(QDataSet.UNITS); if (name == null) { if ( units!=null && UnitsUtil.isTimeLocation(units)) { name = "Epoch"; } else { name = "Variable_" + namesRev.size(); } } names.put(dep0, name); namesRev.put(name, dep0); return name; } @Override public void formatData( String uri, QDataSet data, ProgressMonitor mon) throws Exception { mon.started(); CDFWriter cdf; try { URISplit split= URISplit.parse( uri ); java.util.Map params= URISplit.parseParams( split.params ); File ffile= new File( split.resourceUri.getPath() ); boolean append= "T".equals( params.get("append") ); boolean insert= "T".equals( params.get("insert") ); boolean useRowMajority= "row".equals(params.get("majority") ); // note default is column, which might have been a poor choice. if ( !append && !insert ) { logger.log(Level.FINE, "create CDF file {0}", ffile); logger.log(Level.FINE, "call cdf= new CDFWriter( false )"); cdf = new CDFWriter( useRowMajority ); } else if ( insert ) { //CDFReader read= ReaderFactory.getReader( ffile.toString() ); String name= (String)data.property(QDataSet.NAME); if ( name==null ) name= params.get("arg_0"); if ( name==null ) { throw new IllegalArgumentException("dataset must have name found within CDF file."); } int itype= 21; String type= CdfUtil.getStringDataType(itype); switch (type) { case "CDF_CHAR": type="string"; break; case "CDF_REAL8": case "CDF_DOUBLE": type="double"; break; case "CDF_FLOAT": case "CDF_REAL4": type="float"; break; case "CDF_UINT1": type="short"; break; default: throw new IllegalArgumentException("Unsupported type: "+type+", must be CHAR, REAL8, REAL4, DOUBLE, FLOAT, or UINT1."); } params.put("type",type); logger.log(Level.FINE, "call cdf= new CDFWriter( {0}, false )", ffile.toString() ); cdf = new CDFWriter( ffile.toString(), useRowMajority ); // read in the old file first } else { CDFReader read= ReaderFactory.getReader( ffile.toString() ); for ( String n : read.getVariableNames() ) { namesRev.put( n,null ); //TODO: there's a problem here, where it puts in a set of timetags for each variable. https://sourceforge.net/p/autoplot/bugs/2514/ } logger.log(Level.FINE, "call cdf= new CDFWriter( {0}, false )", ffile.toString() ); cdf = new CDFWriter( ffile.toString(), useRowMajority ); // read in the old file first } String name1= params.get( "arg_0" ); if ( name1!=null ) { names.put(data,name1); namesRev.put(name1,data); } logger.log( Level.FINE, "handling {0}", nameFor(data)); // allocate a good name QDataSet dep0 = (QDataSet) data.property(QDataSet.DEPEND_0); String dep0name=null; if ( insert ) dep0=null; if ( dep0 != null ) { if ( !append ) { String name= nameFor(dep0); dep0name= name; Map params1= new HashMap<>(); params1.put( "timeType",params.get("timeType") ); addVariableRankN( cdf, dep0, name, true, params1, mon.getSubtaskMonitor("dep0") ); } else { String name = (String) dep0.property(QDataSet.NAME); if ( !namesRev.containsKey(name) ) { name= nameFor(dep0); dep0name= name; Map params1= new HashMap<>(); params1.put( "timeType",params.get("timeType") ); try { addVariableRankN( cdf, dep0, name, true, params1, mon.getSubtaskMonitor("dep0") ); } catch ( Exception e ) { logger.fine("CDF Exception, presumably because the variable already exists."); } } else { dep0name= name; // https://sourceforge.net/p/autoplot/bugs/2514/ use the variable if it has the right name. } } } QDataSet dep1 = (QDataSet) data.property(QDataSet.DEPEND_1); if ( insert ) dep1=null; if (dep1 != null) { if ( !append ) { String name= nameFor(dep1); if ( dep1.rank()==1 ) { addVariableRank1NoVary( cdf, dep1, name, true, new HashMap(), mon.getSubtaskMonitor("dep1") ); } else { addVariableRankN( cdf, dep1, name, true, new HashMap(), mon.getSubtaskMonitor("dep1") ); } } else { String name = (String) dep1.property(QDataSet.NAME); if ( !namesRev.containsKey(name) ) { name= nameFor(dep1); Map params1= new HashMap<>(); try { if ( dep1.rank()==1 ) { addVariableRank1NoVary( cdf, dep1, name, true, params1, mon.getSubtaskMonitor("dep1") ); } else { addVariableRankN( cdf, dep1, name, true, params1, mon.getSubtaskMonitor("dep1") ); } } catch ( Exception e ) { logger.fine("CDF Exception, presumably because the variable already exists."); } } else { names.put(dep1,name); } } } QDataSet dep2 = (QDataSet) data.property(QDataSet.DEPEND_2); if ( insert ) dep2=null; if (dep2 != null) { if ( !append ) { String name= nameFor(dep2); if ( dep2.rank()==1 ) { addVariableRank1NoVary( cdf, dep2, name, true, new HashMap(), mon.getSubtaskMonitor("dep2") ); } else { addVariableRankN( cdf, dep2, name, true, new HashMap(), mon.getSubtaskMonitor("dep2") ); } } else { String name= (String) dep2.property(QDataSet.NAME); if ( !namesRev.containsKey(name) ) { name= nameFor(dep2); Map params1= new HashMap<>(); try { if ( dep2.rank()==1 ) { addVariableRank1NoVary( cdf, dep2, name, true, params1, mon.getSubtaskMonitor("dep2") ); } else { addVariableRankN( cdf, dep2, name, true, params1, mon.getSubtaskMonitor("dep2") ); } } catch ( Exception e ) { logger.fine("CDF Exception, presumably because the variable already exists."); } } else { names.put(dep2,name); } } } QDataSet dep3 = (QDataSet) data.property(QDataSet.DEPEND_3); if ( insert ) dep3=null; if (dep3 != null) { if ( !append ) { String name= nameFor(dep3); if ( dep3.rank()==1 ) { addVariableRank1NoVary( cdf, dep3, name, true, new HashMap(), mon.getSubtaskMonitor("dep3") ); } else { addVariableRankN( cdf, dep3, name, true, new HashMap(), mon.getSubtaskMonitor("dep3") ); } } else { String name= (String) dep2.property(QDataSet.NAME); if ( !namesRev.containsKey(name) ) { name= nameFor(dep3); Map params1= new HashMap<>(); try { if ( dep3.rank()==1 ) { addVariableRank1NoVary( cdf, dep3, name, true, params1, mon.getSubtaskMonitor("dep3") ); } else { addVariableRankN( cdf, dep3, name, true, params1, mon.getSubtaskMonitor("dep3") ); } } catch ( Exception e ) { logger.fine("CDF Exception, presumably because the variable already exists."); } } else { names.put(dep3,name); } } } QDataSet bds= (QDataSet) data.property(QDataSet.BUNDLE_1); if ( bds != null && insert==false ) { if ( !append && data.rank()==2 ) { if ( dep1==null ) { logger.fine("writing bundled datasets to CDF separately."); } else { String name= nameFor(bds); if ( bds.rank()==1 || ( bds.rank()==2 && Schemes.isBundleDescriptor(bds) ) ) { addVariableRank1NoVary( cdf, bds, name, true, new HashMap(), mon.getSubtaskMonitor("bundle1") ); } } } else { String name= nameFor(bds); Map params1= new HashMap<>(); try { addVariableRank1NoVary( cdf, bds, name, true, params1, mon.getSubtaskMonitor("bundle1") ); } catch ( Exception e ) { logger.fine("CDF Exception, presumably because the variable already exists."); } } } if ( bds!=null && dep1==null && "T".equals(params.get("bundle")) ) { for ( int i=0; i params, org.das2.util.monitor.ProgressMonitor mon ) throws Exception { Units units = (Units) ds.property(QDataSet.UNITS); CDFDataType type = CDFDataType.DOUBLE; UnitsConverter uc = UnitsConverter.IDENTITY; if (units != null && UnitsUtil.isTimeLocation(units)) { type = CDFDataType.EPOCH; uc = units.getConverter(Units.cdfEpoch); } if ( ds.rank()==1 ) { //cdf.defineNRVVariable( name, type, new int[0], 0 ); //cdf.createVariable( name, type, new int[0] ); Object array= CdfDataSourceFormat.datasetToArray( ds, uc, type, mon ); logger.log(Level.FINE, "call cdf.addNRVVariable( {0},{1},{2})", new Object[]{name, logName(type), logName( new int[] { ds.length() } ), logName(array) }); cdf.addNRVVariable( name, type, new int[] { ds.length() }, array ); } else if ( Schemes.isBundleDescriptor(ds) ) { String[] array= new String[ ds.length() ]; String[] ss= DataSetOps.bundleNames(ds); int dim=0; // max number of characters for ( int i=0; i 2147483648. ? d - 4294967296. : d ); } private static short encodeUINT2( double d ) { return (short)( d > 32768 ? d - 65536 : d ); } private static byte encodeUINT1( double d ) { return (byte)( d > 128 ? d - 256 : d ); } /** * convert the rank 1 dataset to a buffer. * @param ds rank 1 dataset * @param uc units converter to convert the type. * @param type type code, such as CDF_DOUBLE indicating how the data should be converted. * @return buffer of this type. */ private static ByteBuffer doIt1Nio( QDataSet ds, UnitsConverter uc, CDFDataType type ) { ByteBuffer export; QubeDataSetIterator iter = new QubeDataSetIterator(ds); if ( type==CDFDataType.DOUBLE || type==CDFDataType.EPOCH ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*8 ); buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.putDouble( uc.convert(iter.getValue(ds) ) ); } export= buf; } else if ( type==CDFDataType.TT2000 ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*8 ); buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.putLong( (long)uc.convert(iter.getValue(ds) ) ); } export= buf; } else if ( type==CDFDataType.FLOAT ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*4 ); buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.putFloat( (float)uc.convert(iter.getValue(ds) ) ); } export= buf; } else if ( type==CDFDataType.INT4 ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*4 ); buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.putInt( (int)uc.convert(iter.getValue(ds) ) ); } export= buf; } else if ( type==CDFDataType.INT2 ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*2 ); buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.putShort( (short)uc.convert(iter.getValue(ds) ) ); } export= buf; } else if ( type==CDFDataType.INT1 ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*1 ); //buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.put( (byte)uc.convert(iter.getValue(ds) ) ); } export= buf; } else if ( type==CDFDataType.UINT4 ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*4 ); buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.putInt( encodeUINT4( uc.convert(iter.getValue(ds) ) ) ); } export= buf; } else if ( type==CDFDataType.UINT2 ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*2 ); buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.putShort(encodeUINT2( uc.convert(iter.getValue(ds) ) ) ); } export= buf; } else if ( type==CDFDataType.UINT1 ) { ByteBuffer buf= ByteBuffer.allocate( ds.length()*1 ); //buf.order( ByteOrder.LITTLE_ENDIAN ); while (iter.hasNext()) { iter.next(); buf.put( encodeUINT1( uc.convert(iter.getValue(ds) ) ) ); } export= buf; } else if ( type==CDFDataType.CHAR ) { int maxLenth= 1; ByteBuffer buf= ByteBuffer.allocate( ds.length()*maxLenth ); for ( int i=0; i params, org.das2.util.monitor.ProgressMonitor mon) throws Exception { Units units = (Units) ds.property(QDataSet.UNITS); CDFDataType type = CDFDataType.DOUBLE; String sinsert=params.get("insert"); boolean insert= sinsert!=null && sinsert.startsWith("T"); if ( insert ) { type = cdf.getVariableType(name); } else { String t= params.get("type"); if ( t!=null ) { switch (t) { case "float": type= CDFDataType.FLOAT; break; case "byte": type= CDFDataType.INT1; break; case "int1": type= CDFDataType.INT1; break; case "int2": type= CDFDataType.INT2; break; case "int4": type= CDFDataType.INT4; break; case "uint1": type= CDFDataType.UINT1; break; case "uint2": type= CDFDataType.UINT2; break; case "uint4": type= CDFDataType.UINT4; break; case "double": type= CDFDataType.DOUBLE; break; case "string": type= CDFDataType.CHAR; break; default: logger.log(Level.WARNING, "unsupported type, using double: {0}", t); break; } } else { if ( ds.rank()<3 ) { type= CDFDataType.DOUBLE; } else { type= CDFDataType.FLOAT; } } } boolean compressed= "T".equals( params.get("compressed") ); UnitsConverter uc = UnitsConverter.IDENTITY; if (units != null && UnitsUtil.isTimeLocation(units)) { boolean tt2000= !( "epoch".equals( params.get("timeType") ) ); if ( tt2000 ) { type = CDFDataType.TT2000; uc = units.getConverter(Units.cdfTT2000); units= Units.cdfTT2000; } else { type = CDFDataType.EPOCH; uc = units.getConverter(Units.cdfEpoch); units= Units.cdfEpoch; } } if ( ds.rank()==0 ) { throw new IllegalArgumentException("rank 0 data not supported"); } if ( ds.rank()>4 ) { throw new IllegalArgumentException("high rank data not supported"); } if ( compressed ) { if ( ds.rank()==1 ) { logger.log(Level.FINE, "call cdf.defineCompressedVariable( {0}, {1}, {2} )", new Object[] { name, logName(type), logName(new int[0]) } ); cdf.defineCompressedVariable( name, type, new int[0] ); addData( cdf, name, dataSetToNioArray( ds, uc, type, mon ) ); //TODO: I think I need to compress the channel. } else { switch (ds.rank()) { case 2: defineCompressedVariable( cdf, name, type, new int[] { ds.length(0) } ); break; case 3: defineCompressedVariable( cdf, name, type, new int[] { ds.length(0),ds.length(0,0) } ); break; case 4: defineCompressedVariable( cdf, name, type, new int[] { ds.length(0),ds.length(0,0),ds.length(0,0,0) } ); break; default: break; } Object o= CdfDataSourceFormat.datasetToArray( ds, uc, type, mon ); addData( cdf, name, o ); } } else { if ( ds.rank()==1 ) { if ( !insert ) { defineVariable( cdf, name, type, new int[0] ); addData( cdf, name, dataSetToNioArray( ds, uc, type, mon ) ); } else { addData( cdf, name, doIt1( ds, uc, type ) ); } } else { // this branch doesn't use dataSetToNioArray if ( !insert ) { switch (ds.rank()) { case 2: defineVariable( cdf, name, type, new int[] { ds.length(0) } ); break; case 3: defineVariable( cdf, name, type, new int[] { ds.length(0),ds.length(0,0) } ); break; case 4: defineVariable( cdf, name, type, new int[] { ds.length(0),ds.length(0,0),ds.length(0,0,0) } ); break; default: break; } } addData(cdf, name, CdfDataSourceFormat.datasetToArray( ds, uc, type, mon ) ); } } if ( !insert ) { copyMetadata( cdf, units, name, type, isSupport, ds ); } } /** * return expressions so example testing codes can be written * @param o * @return */ private static String logName( Object o ) { if ( o.getClass().isArray() ) { StringBuilder s= new StringBuilder(o.getClass().getComponentType().toString()+"["); s.append( Array.getLength(o)); if ( Array.getLength(o)>0 ) { o= Array.get(o,0); while ( o.getClass().isArray() ) { s.append(",").append(Array.getLength(o)); o= Array.get(o,0); } } s.append("]"); return s.toString(); } else if ( o instanceof String ) { return "\"" + o + "\""; } else if ( o instanceof CDFDataType ) { return "CDFDataType=" + ((CDFDataType)o).getValue(); } else { return o.toString(); } } private static void write( CDFWriter cdf, String name ) throws IOException { logger.log(Level.FINE, "call cdf.write({0})", new Object[] { logName(name) } ); try { CdfDataSource.cdfCacheReset(); cdf.write( name ); } catch ( FileNotFoundException ex ){ logger.log(Level.WARNING, "first attempt to write \"{0}\" fails, try again for good measure", name); CdfDataSource.cdfCacheReset(); System.gc(); try { Thread.sleep(1000); System.gc(); Thread.sleep(1000); System.gc(); } catch (InterruptedException ex1) { logger.log(Level.SEVERE, null, ex1); } cdf.write( name ); } } private static void defineCompressedVariable( CDFWriter cdf, String name, CDFDataType type, int[] dims ) throws Exception { logger.log(Level.FINE, "call cdf.defineCompressedVariable({0},{1},{2})", new Object[] { logName(name), logName(type), logName(dims) } ); cdf.defineCompressedVariable( name, type, dims ); } private static void defineVariable( CDFWriter cdf, String name, CDFDataType type, int[] dims ) throws Exception { logger.log(Level.FINE, "call cdf.defineVariable({0},{1},{2})", new Object[] { logName(name), logName(type), logName(dims) } ); cdf.defineVariable( name, type, dims ); } private static void addData( CDFWriter cdf, String name, Object d ) throws Exception { logger.log(Level.FINE, "call cdf.addData({0},{1})", new Object[] { logName(name), logName(d) } ); cdf.addData( name, d ); } private static void addVariableAttributeEntry( CDFWriter cdf, String varName, String attrName, CDFDataType type, Object o ) throws CDFException.WriterError { logger.log( Level.FINE, "call cdf.addVariableAttributeEntry( {0}, {1}, {2}, {3} )", new Object[] { logName(varName), logName(attrName), logName(type), logName( o ) } ); if ( type==CDFDataType.CHAR && o.toString().length()==0 ) { o= " "; } cdf.addVariableAttributeEntry( varName, attrName, type, o ); } /** * copy metadata for the variable. * @param units Units object to identify time types. * @param name the variable name * @param ds the dataset containing metadata. * @throws Exception */ private static void copyMetadata( CDFWriter cdf, Units units, String name, CDFDataType type, boolean isSupport, QDataSet ds ) throws Exception { if ( units!=null ) { if (units == Units.cdfEpoch) { addVariableAttributeEntry( cdf, name, "UNITS", CDFDataType.CHAR, "ms" ); } else if ( units==Units.cdfTT2000 ) { addVariableAttributeEntry( cdf, name, "UNITS", CDFDataType.CHAR, "ns" ); } else { addVariableAttributeEntry( cdf, name, "UNITS", CDFDataType.CHAR, units.toString() ); } } else { addVariableAttributeEntry( cdf, name, "UNITS", CDFDataType.CHAR, " " ); } String label = (String) ds.property(QDataSet.LABEL); if (label != null && label.length()>0 ) { if ( units!=null && label.endsWith("("+units+")") ) { label= label.substring(0,label.length()-units.toString().length()-2); } addVariableAttributeEntry( cdf, name,"LABLAXIS", CDFDataType.CHAR, label); } String title = (String) ds.property(QDataSet.TITLE); if (title != null && title.length()>0 ) { addVariableAttributeEntry( cdf, name,"CATDESC", CDFDataType.CHAR, title); } String description = (String) ds.property(QDataSet.DESCRIPTION); if (description != null && description.length()>0 ) { addVariableAttributeEntry( cdf, name,"VAR_NOTES", CDFDataType.CHAR, description); } Number vmax= (Number) ds.property( QDataSet.VALID_MAX ); Number vmin= (Number) ds.property( QDataSet.VALID_MIN ); if ( vmax!=null || vmin !=null ) { if ( units==Units.cdfEpoch ) { //UnitsConverter uc= ((Units)ds.property(QDataSet.UNITS)).getConverter(units); //if ( vmax==null ) vmax= 1e38; else vmax= uc.convert(vmax); //if ( vmin==null ) vmin= -1e38; else vmin= uc.convert(vmin); //cdf.addVariableAttributeEntry( name, "VALIDMIN", CDFDataType.DOUBLE, vmin.doubleValue() ); //cdf.addVariableAttributeEntry( name, "VALIDMAX", CDFDataType.DOUBLE, vmax.doubleValue() ); } else if ( units==Units.cdfTT2000 ) { if ( vmax!=null && vmin !=null ) { cdf.addVariableAttributeEntry( name, "VALIDMIN", CDFDataType.TT2000, new long[] { vmin.longValue() } ); cdf.addVariableAttributeEntry( name, "VALIDMAX", CDFDataType.TT2000, new long[] { vmax.longValue() } ); } } else { if ( vmax==null ) vmax= 1e38; if ( vmin==null ) vmin= -1e38; cdf.addVariableAttributeEntry( name, "VALIDMIN", type, new double[] { vmin.doubleValue() } ); cdf.addVariableAttributeEntry( name, "VALIDMAX", type, new double[] { vmax.doubleValue() } ); } } Number fillval= (Number) ds.property( QDataSet.FILL_VALUE ); if ( fillval!=null ) { if ( units==Units.cdfEpoch ) { } else if ( units==Units.cdfTT2000 ) { //TODO: use long access, if available. cdf.addVariableAttributeEntry( name, "FILLVAL", CDFDataType.TT2000, new long[] { fillval.longValue() } ); } else { cdf.addVariableAttributeEntry( name,"FILLVAL", type, new double[] { fillval.doubleValue() }); } } else { //cdf.addVariableAttributeEntry( name,"FILLVAL",CDFDataType.DOUBLE,-1e31); } Number smax= (Number) ds.property( QDataSet.TYPICAL_MAX ); Number smin= (Number) ds.property( QDataSet.TYPICAL_MIN ); if ( smax!=null || smin !=null ) { if ( units==Units.cdfEpoch ) { //UnitsConverter uc= ((Units)ds.property(QDataSet.UNITS)).getConverter(units); //if ( smax==null ) smax= 1e38; else smax= uc.convert(smax); //if ( smin==null ) smin= -1e38; else smin= uc.convert(smin); //cdf.addVariableAttributeEntry( name,"SCALEMIN", CDFDataType.DOUBLE, smin.doubleValue() ); //cdf.addVariableAttributeEntry( name,"SCALEMAX", CDFDataType.DOUBLE, smax.doubleValue() ); } else if ( units==Units.cdfTT2000 ) { if ( smax==null ) smax= Units.cdfTT2000.parse("1958-01-01T00:00").doubleValue( Units.cdfTT2000); if ( smin==null ) smin= Units.cdfTT2000.parse("2058-01-01T00:00").doubleValue( Units.cdfTT2000); cdf.addVariableAttributeEntry( name, "SCALEMIN", CDFDataType.TT2000, new long[] { smin.longValue() } ); cdf.addVariableAttributeEntry( name, "SCALEMAX", CDFDataType.TT2000, new long[] { smax.longValue() } ); } else { if ( smax==null ) smax= 1e38; if ( smin==null ) smin= -1e38; cdf.addVariableAttributeEntry( name,"SCALEMIN", type, new double[] { smin.doubleValue() } ); cdf.addVariableAttributeEntry( name,"SCALEMAX", type, new double[] { smax.doubleValue() } ); } } String scaleTyp= (String) ds.property(QDataSet.SCALE_TYPE); if ( scaleTyp!=null ) { addVariableAttributeEntry( cdf, name,"SCALETYP",CDFDataType.CHAR,scaleTyp); } String format= (String) ds.property( QDataSet.FORMAT ); if ( format!=null && format.trim().length()>0 ) { addVariableAttributeEntry( cdf, name,"FORMAT",CDFDataType.CHAR,format); } String displayType= (String)ds.property( QDataSet.RENDER_TYPE ); if ( displayType==null || displayType.length()==0 ) { displayType= DataSourceUtil.guessRenderType(ds); } switch (displayType) { case "nnSpectrogram": case "spectrogram": displayType= "spectrogram"; break; case "image": displayType= "image"; break; case "series": case "scatter": case "hugeScatter": displayType= "time_series"; break; default: break; } addVariableAttributeEntry( cdf, name,"DISPLAY_TYPE", CDFDataType.CHAR, displayType ); addVariableAttributeEntry( cdf, name,"VAR_TYPE", CDFDataType.CHAR, isSupport ? "support_data" : "data" ); } @Override public boolean canFormat(QDataSet ds) { return ! ( ds.rank()==0 || SemanticOps.isJoin(ds) ); } @Override public String getDescription() { return "NASA Common Data Format"; } }