Restrict units to what is defined in NGFF schema

src/main/java/com/glencoesoftware/bioformats2raw/Converter.java

@@ -127,6 +127,24 @@ public class Converter implements Callable<Integer> {
   /** NGFF specification version.*/
   public static final String NGFF_VERSION = "0.4";
 
+  /** Units allowed by NGFF specification for length/space axes. */
+  private static final List<String> LENGTH_UNITS = Arrays.asList(
+    "angstrom", "attometer", "centimeter", "decimeter", "exameter",
+    "femtometer", "foot", "gigameter", "hectometer", "inch", "kilometer",
+    "megameter", "meter", "micrometer", "mile", "millimeter", "nanometer",
+    "parsec", "petameter", "picometer", "terameter", "yard", "yoctometer",
+    "yottameter", "zeptometer", "zettameter"
+  );
+
+  /** Units allowed by NGFF specification for time axes. */
+  private static final List<String> TIME_UNITS = Arrays.asList(
+    "attosecond", "centisecond", "day", "decisecond", "exasecond",
+    "femtosecond", "gigasecond", "hectosecond", "hour", "kilosecond",
+    "megasecond", "microsecond", "millisecond", "minute", "nanosecond",
+    "petasecond", "picosecond", "second", "terasecond", "yoctosecond",
+    "yottasecond", "zeptosecond", "zettasecond"
+  );
+
   private volatile Path inputPath;
   private volatile String outputLocation;
 
@@ -2767,16 +2785,20 @@ else if (axis.equals("c")) {
         String unitName = null;
         try {
           if (scale instanceof Length) {
-            unitName = UnitsLength.fromString(symbol).name().toLowerCase();
+            unitName = getLengthUnit(
+              UnitsLength.fromString(symbol).name().toLowerCase());
           }
           else if (scale instanceof Time) {
-            unitName = UnitsTime.fromString(symbol).name().toLowerCase();
+            unitName = getTimeUnit(
+              UnitsTime.fromString(symbol).name().toLowerCase());
           }
         }
         catch (EnumerationException e) {
           LOGGER.warn("Could not identify unit '{}'", symbol);
         }
-        thisAxis.put("unit", unitName);
+        if (unitName != null) {
+          thisAxis.put("unit", unitName);
+        }
       }
       axes.add(thisAxis);
     }
@@ -3210,6 +3232,44 @@ private String getNameFromWavelength(Length v) {
     return v.toString();
   }
 
+  /**
+   * Get a valid unit name for a length/space axis according to NGFF spec.
+   * OME-XML schema allows for some units that are not valid according to
+   * NGFF, e.g. 'pixel'.
+   *
+   * @param unitName length unit from OME schema
+   * @return valid NGFF unit or null
+   */
+  private String getLengthUnit(String unitName) {
+    if (unitName == null) {
+      return null;
+    }
+    if (LENGTH_UNITS.contains(unitName)) {
+      return unitName;
+    }
+    LOGGER.debug("Omitting unit '{}'", unitName);
+    return null;
+  }
+
+  /**
+   * Get a valid unit name for a time axis according to NGFF spec.
+   * OME-XML schema allows for some units that are not valid according to
+   * NGFF, e.g. 'decasecond'.
+   *
+   * @param unitName length unit from OME schema
+   * @return valid NGFF unit or null
+   */
+  private String getTimeUnit(String unitName) {
+    if (unitName == null) {
+      return null;
+    }
+    if (TIME_UNITS.contains(unitName)) {
+      return unitName;
+    }
+    LOGGER.debug("Omitting unit '{}'", unitName);
+    return null;
+  }
+
   private int calculateResolutions(int width, int height) {
     int resolutions = 1;
     while (width > minSize || height > minSize) {

src/test/java/com/glencoesoftware/bioformats2raw/test/ZarrTest.java

@@ -416,6 +416,59 @@ public void testPhysicalSizes() throws Exception {
     }
   }
 
+  /**
+   * Test that physical sizes with invalid units are not saved in
+   * axes metadata.
+   */
+  @Test
+  public void testInvalidUnitPhysicalSizes() throws Exception {
+    input = fake("physicalSizeX", "1pixel",
+      "physicalSizeY", "0.5thou",
+      "physicalSizeZ", "2dam");
+    assertTool();
+
+    List<Map<String, Object>> multiscales = getMultiscales("0");
+    assertEquals(1, multiscales.size());
+    Map<String, Object> multiscale = multiscales.get(0);
+    checkMultiscale(multiscale, "image");
+    List<Map<String, Object>> axes =
+      (List<Map<String, Object>>) multiscale.get("axes");
+    checkAxes(axes, "TCZYX",
+      new String[] {null, null, null, null, null});
+
+    List<Map<String, Object>> datasets =
+      (List<Map<String, Object>>) multiscale.get("datasets");
+    assertEquals(2, datasets.size());
+
+    for (int r=0; r<datasets.size(); r++) {
+      Map<String, Object> dataset = datasets.get(r);
+      List<Map<String, Object>> transforms =
+        (List<Map<String, Object>>) dataset.get("coordinateTransformations");
+      assertEquals(1, transforms.size());
+      Map<String, Object> scale = transforms.get(0);
+      assertEquals("scale", scale.get("type"));
+      List<Double> axisValues = (List<Double>) scale.get("scale");
+
+      assertEquals(5, axisValues.size());
+      double factor = Math.pow(2, r);
+      // X and Y are the only dimensions that are downsampled,
+      // so the TCZ physical scales remain the same across all resolutions
+      assertEquals(axisValues, Arrays.asList(new Double[] {
+        1.0, 1.0, 2.0, 0.5 * factor, factor}));
+    }
+
+    // Check dimensions and special pixels
+    ZarrGroup z = ZarrGroup.open(output.toString());
+    ZarrArray series0 = z.openArray("0/0");
+    assertArrayEquals(new int[] {1, 1, 1, 512, 512}, series0.getShape());
+    assertArrayEquals(new int[] {1, 1, 1, 512, 512}, series0.getChunks());
+    int[] shape = new int[] {1, 1, 1, 512, 512};
+    byte[] tile = new byte[512 * 512];
+    series0.read(tile, shape);
+    int[] seriesPlaneNumberZCT = FakeReader.readSpecialPixels(tile);
+    assertArrayEquals(new int[] {0, 0, 0, 0, 0}, seriesPlaneNumberZCT);
+  }
+
   /**
    * Test using a different tile size from the default (1024).
    */