Refactor metadata module to be more pythonic

The axis submodule has also been streamlined to drop
the CalibratedAxis dict and instead uses a list of Strings.
This avoids Java import errors if a user imports the axis submodule
before initializing ImageJ.

src/imagej/convert.py

@@ -233,7 +233,7 @@ def java_to_xarray(ij: "jc.ImageJ", jobj) -> xr.DataArray:
     xr_dims = list(permuted_rai.dims)
     xr_attrs = sj.to_python(permuted_rai.getProperties())
     xr_attrs = {sj.to_python(k): sj.to_python(v) for k, v in xr_attrs.items()}
-    xr_attrs["imagej"] = metadata.ImageMetadata.create_imagej_metadata(xr_axes, xr_dims)
+    xr_attrs["imagej"] = metadata.create_imagej_metadata(xr_axes, xr_dims)
     # reverse axes and dims to match narr
     xr_axes.reverse()
     xr_dims.reverse()

src/imagej/dims.py

@@ -216,12 +216,12 @@ def _assign_axes(
                 if cal_axis_type == "DefaultLinearAxis":
                     origin = xarr.attrs["imagej"][ij_dim + "_origin"]
                     scale = xarr.attrs["imagej"][ij_dim + "_scale"]
-                    jaxis = metadata.Axis._str_to_cal_axis(cal_axis_type)(
+                    jaxis = metadata.axis.str_to_calibrated_axis(cal_axis_type)(
                         ax_type, scale, origin
                     )
                 else:
                     try:
-                        jaxis = metadata.Axis._str_to_cal_axis(cal_axis_type)(
+                        jaxis = metadata.axis.str_to_calibrated_axis(cal_axis_type)(
                             ax_type, doub_coords
                         )
                     except (JException, TypeError):

src/imagej/metadata/__init__.py

@@ -0,0 +1,37 @@
+from typing import Sequence
+
+import imagej.dims as dims
+import imagej.metadata.axis as axis
+from imagej._java import jc
+
+
+def create_imagej_metadata(
+    axes: Sequence["jc.CalibratedAxis"], dim_seq: Sequence[str]
+) -> dict:
+    """
+    Create the ImageJ metadata attribute dictionary for xarray's global attributes.
+    :param axes: A list or tuple of ImageJ2 axis objects
+        (e.g. net.imagej.axis.DefaultLinearAxis).
+    :param dim_seq: A list or tuple of the dimension order (e.g. ['X', 'Y', 'C']).
+    :return: Dict of image metadata.
+    """
+    ij_metadata = {}
+    if len(axes) != len(dim_seq):
+        raise ValueError(
+            f"Axes length ({len(axes)}) does not match \
+                dimension length ({len(dim_seq)})."
+        )
+
+    for i in range(len(axes)):
+        # get CalibratedAxis type as string (e.g. "EnumeratedAxis")
+        ij_metadata[
+            dims._to_ijdim(dim_seq[i]) + "_cal_axis_type"
+        ] = axis.calibrated_axis_to_str(axes[i])
+        # get scale and origin for DefaultLinearAxis
+        if isinstance(axes[i], jc.DefaultLinearAxis):
+            ij_metadata[dims._to_ijdim(dim_seq[i]) + "_scale"] = float(axes[i].scale())
+            ij_metadata[dims._to_ijdim(dim_seq[i]) + "_origin"] = float(
+                axes[i].origin()
+            )
+
+    return ij_metadata

src/imagej/metadata/axis.py

@@ -0,0 +1,47 @@
+from _jpype import JClass
+
+from imagej._java import jc
+
+_calibrated_axes = [
+    "net.imagej.axis.ChapmanRichardsAxis",
+    "net.imagej.axis.DefaultLinearAxis",
+    "net.imagej.axis.EnumeratedAxis",
+    "net.imagej.axis.ExponentialAxis",
+    "net.imagej.axis.ExponentialRecoveryAxis",
+    "net.imagej.axis.GammaVariateAxis",
+    "net.imagej.axis.GaussianAxis",
+    "net.imagej.axis.IdentityAxis",
+    "net.imagej.axis.InverseRodbardAxis",
+    "net.imagej.axis.LogLinearAxis",
+    "net.imagej.axis.PolynomialAxis",
+    "net.imagej.axis.PowerAxis",
+    "net.imagej.axis.RodbardAxis",
+]
+
+
+def calibrated_axis_to_str(axis: "jc.CalibratedAxis") -> str:
+    """
+    Convert a CalibratedAxis class to a String.
+    :param axis: CalibratedAxis type (e.g. net.imagej.axis.DefaultLinearAxis).
+    :return: String of CalibratedAxis typeb(e.g. "DefaultLinearAxis").
+    """
+    if not isinstance(axis, JClass):
+        axis = axis.__class__
+
+    return str(axis).split("'")[1]
+
+
+def str_to_calibrated_axis(axis: str) -> "jc.CalibratedAxis":
+    """
+    Convert a String to CalibratedAxis class.
+    :param axis: String of calibratedAxis type (e.g. "DefaultLinearAxis").
+    :return: Java class of CalibratedAxis type
+        (e.g. net.imagej.axis.DefaultLinearAxis).
+    """
+    if not isinstance(axis, str):
+        raise TypeError(f"Axis {type(axis)} is not a String.")
+
+    if axis in _calibrated_axes:
+        return getattr(jc, axis.split(".")[3])
+    else:
+        return None