Ticket #405 (closed defect: fixed)

The Bio-Formats Exporter plugin feeds the input dataset's OME-XML directly to the output writer, which is fine in the simple case but can often cause problems. (Upon initial inspection, the bfconvert command line utility does not appear to have these problems, though it would be good to check the code to verify.)

For example, the Exporter is unable to export datasets opened from series beyond the first when the dimensions differ:

run("Bio-Formats Importer", "open=[/Volumes/DATA/leica-lif/andrew/Par6 GFP delaminating cells.lif] view=Hyperstack stack_order=XYCZT series022");
run("Bio-Formats Exporter", "compression=Uncompressed save=/Users/curtis/Desktop/out.ome.tif");

Results in an exception:

java.lang.IllegalArgumentException: ZCT size vs image count mismatch (sizeZ=5, sizeC=1, sizeT=2, total=51)
	at loci.formats.FormatTools.getZCTCoords(FormatTools.java:293)
	at loci.formats.out.OMETiffWriter.close(OMETiffWriter.java:125)
	at loci.formats.out.TiffWriter.saveBytes(TiffWriter.java:170)
	at loci.formats.out.TiffWriter.saveBytes(TiffWriter.java:191)
	at loci.formats.out.OMETiffWriter.saveBytes(OMETiffWriter.java:171)
	at loci.formats.FormatWriter.saveBytes(FormatWriter.java:89)
	at loci.plugins.exporter.Exporter.run(Exporter.java:367)
	at loci.plugins.LociExporter.run(LociExporter.java:73)

Another example is if you have imported only a subset of the data, using split channels, or specify ranges:

run("Bio-Formats Importer", "open=/Volumes/DATA/deltavision/brian/lung-lentiCMtriple-1_R3D.dv autoscale split_channels view=Hyperstack stack_order=XYCZT");
close();
run("Bio-Formats Exporter", "save=/Users/curtis/Desktop/lung_c2.ome.tif compression=Uncompressed");

Here the OME-TIFF file successfully exports, but SizeC is still 2 instead of 1, resulting in missing planes, and a fallback behavior to the vanilla TIFF reader on subsequent import.