doc: improve documentation for frac_masks

doc/hiopy.rst

@@ -79,11 +79,12 @@ All of the above steps can be done using the `hiopy-configure` tool.
 Assumptions
 ~~~~~~~~~~~
 
-hiopy assumes that the variables it should write have the time axis as
-the first dimension and the horizontal axis as its last
-dimension. The variables are enabled to be written using the boolean value of attribute hiopy::enable.
+Variables that are precessed by hiopy need to have set the attribute `hiopy::enable`.
 
-hiopy specific parameters are added in the attributes and are usually prefixed by `hiopy::`:
+hiopy examinates the dimensions of the variable. If the first dimension has a coordinate, and the `axis` attribute of the coordinate is `T`, hiopy assumes that this variable is time dependend with the respective time coordinates. If such a coordiante is not found, hiopy considers the field as "one_shot" field which is handled once in the very beginning of the run.
+Furthermore hiopy assumes that the last dimension is the spatial dimension. It examinates the attribute `grid_mapping` to determine the corresponding grid.
+
+Further hiopy specific parameters can be added in the attributes and are prefixed with `hiopy::`:
 
 .. csv-table::
    :header: "attribute", "entity", "description"
@@ -93,7 +94,7 @@ hiopy specific parameters are added in the attributes and are usually prefixed b
    "hiopy::src_name", "Array", "yac field name of the source field. Default is the name of the array itself."
    "hiopy::time_method", "Array", "Time interpolation method (point, mean). Default: point"
    "hiopy::nnn", "Array", "Number of nearest neighbors to use. Default is 1."
-   "hiopy::frac_mask", "Array", "Array name of the fractional mask that is used when forwarding the data. (see `yac_cput_frac`)"
+   "hiopy::frac_mask", "Array", "Array name of the fractional mask that is used when forwarding the data. (see `yac_cput_frac`). Could also be a field that is processed by hiopy itself in the same run."
    "hiopy::enable", "Boolean", "Enable or disable the variable to be handled by the worker"
 
 `hiopy.configure` can also be used as a python
@@ -105,16 +106,9 @@ from level 6 to 0 and one 3d variable "ta":
    import zarr
    import hiopy.configure as hc
 
-   dates = {"startdate": np.datetime64("2020-01-01"),
-         "enddate": np.datetime64("2020-01-10")}
-
    z = zarr.open("dataset.zarr")
-   hc.create_hierachy(z,
-                      order=6,
-                      startdate=np.datetime64("2020-01-01"),
-                      enddate=np.datetime64("2020-01-10"),
-                      dt=3*60*60)  # 3 hourly time axis
-
+   hc.add_healpix_hierarchy(z, order=6)
+   hc.add_time(z, "2020-01-01", "2020-01-10", dt=3*60*60)  # 3 hourly time axis
    hc.add_height(z, "level", 72)
    hc.add_variable(z,
                    name="ta",
@@ -124,6 +118,8 @@ from level 6 to 0 and one 3d variable "ta":
                    zaxis="level",
                    chunk_shape=(8, 16, 2**12))  # 8M (uncompressed)
 
+   zarr.consolidate_metadata(z.store)  # optional
+
 Setup
 -----