Add a start_date attribute to NetCDFWriter

Prior to this version, users had to go to the simution to find the start date.
It is now saved as an attribute, making it easily accessible.

NEWS.md

@@ -3,6 +3,14 @@
 v0.2.9
 -------
 
+## Features
+
+### Add a `start_date` attribute to NetCDFWriter PR [#94](https://github.com/CliMA/ClimaDiagnostics.jl/pull/94).
+
+Prior to this version, users had to go to the simution to find the start date.
+It can now be saved as an attribute, making it easily accessible.
+To do so, users need to pass the kwarg `start_date` when calling `NetCDFWriter`.
+
 ## Bug fixes
 
 ### Acquiring ownership with `compute!` PR [#88](https://github.com/CliMA/ClimaDiagnostics.jl/pull/88).

docs/src/writers.md

@@ -18,7 +18,7 @@ interpolation is needed.
 ## `NetCDFWriter`
 
 The `NetCDFWriter` resamples the input `Field` to a rectangular grid and saves
-the output to a NetCDF file. 
+the output to a NetCDF file.
 
 The `NetCDFWriter` relies on the `Remappers` module in `ClimaCore` to
 interpolate onto the rectangular grid. Horizontally, this interpolation is a
@@ -32,6 +32,9 @@ and the output directory where the files should be saved. By default, the
 exist. The `NetCDFWriter` does not overwrite existing data and will error out if
 existing data is inconsistent with the new one.
 
+Optionally (recommended), you can pass an optional argument `start_date`, which
+will be saved as an attribute of your NetCDF file, easily accessible.
+
 `NetCDFWriter`s take as one of the inputs the desired number of points along
 each of the dimensions. For the horizontal dimensions, points are sampled
 linearly. For the vertical dimension, the behavior can be customized by passing

src/netcdf_writer.jl

@@ -16,7 +16,7 @@ include("netcdf_writer_coordinates.jl")
 A struct to remap `ClimaCore` `Fields` to rectangular grids and save the output to NetCDF
 files.
 """
-struct NetCDFWriter{T, TS, DI, SYNC, ZSM <: AbstractZSamplingMethod} <:
+struct NetCDFWriter{T, TS, DI, SYNC, ZSM <: AbstractZSamplingMethod, DATE} <:
        AbstractWriter
     """The base folder where to save the files."""
     output_dir::String
@@ -64,6 +64,9 @@ struct NetCDFWriter{T, TS, DI, SYNC, ZSM <: AbstractZSamplingMethod} <:
 
     """Set of datasets that need to be synced. Useful when `sync_schedule` is not `nothing`."""
     unsynced_datasets::Set{NCDatasets.NCDataset}
+
+    """Date of the beginning of the simulation (it is used to convert seconds to dates)."""
+    start_date::DATE
 end
 
 """
@@ -103,6 +106,7 @@ Keyword arguments
                    a boolean callable that takes as a single argument the `integrator`.
                    `sync_schedule` can also be set as `nothing`, in which case we let
                    handling buffered writes to disk.
+- `start_date`: Date of the beginning of the simulation.
 """
 function NetCDFWriter(
     space,
@@ -112,6 +116,7 @@ function NetCDFWriter(
     sync_schedule = ClimaComms.device(space) isa ClimaComms.CUDADevice ?
                     EveryStepSchedule() : nothing,
     z_sampling_method = LevelsMethod(),
+    start_date = nothing,
 )
     horizontal_space = Spaces.horizontal_space(space)
     is_horizontal_space = horizontal_space == space
@@ -173,6 +178,7 @@ function NetCDFWriter(
         typeof(preallocated_arrays),
         typeof(sync_schedule),
         typeof(z_sampling_method),
+        typeof(start_date),
     }(
         output_dir,
         Dict{String, Remapper}(),
@@ -184,6 +190,7 @@ function NetCDFWriter(
         preallocated_arrays,
         sync_schedule,
         unsynced_datasets,
+        start_date,
     )
 end
 
@@ -349,6 +356,13 @@ function write_field!(writer::NetCDFWriter, field, diagnostic, u, p, t)
         writer.interpolated_physical_z,
     )
 
+    start_date = nothing
+    if isnothing(writer.start_date)
+        start_date = get(p, :start_date, nothing)
+    else
+        start_date = writer.start_date
+    end
+
     if haskey(nc, "$(var.short_name)")
         # We already have something in the file
         v = nc["$(var.short_name)"]
@@ -367,9 +381,8 @@ function write_field!(writer::NetCDFWriter, field, diagnostic, u, p, t)
         v.attrib["long_name"] = output_long_name(diagnostic)::String
         v.attrib["units"] = var.units::String
         v.attrib["comments"] = var.comments::String
-        if hasproperty(p, :start_date)
-            # FIXME: We are hardcoding p.start_date !
-            v.attrib["start_date"] = string(p.start_date)::String
+        if !isnothing(start_date) && !haskey(v.attrib, "start_date")
+            v.attrib["start_date"] = string(start_date)::String
         end
         temporal_size = 0
     end
@@ -382,8 +395,8 @@ function write_field!(writer::NetCDFWriter, field, diagnostic, u, p, t)
 
     # FIXME: We are hardcoding p.start_date !
     # FIXME: We are rounding t
-    if hasproperty(p, :start_date)
-        nc["date"][time_index] = string(p.start_date + Dates.Second(round(t)))
+    if !isnothing(start_date)
+        nc["date"][time_index] = string(start_date + Dates.Millisecond(1000t))
     end
 
     # TODO: It would be nice to find a cleaner way to do this

test/integration_test.jl

@@ -7,6 +7,8 @@ import NCDatasets
 
 import ClimaDiagnostics
 
+import Dates
+
 import ClimaComms
 @static if pkgversion(ClimaComms) >= v"0.6"
     ClimaComms.@import_required_backends
@@ -38,6 +40,7 @@ function setup_integrator(output_dir; context, more_compute_diagnostics = 0)
         space,
         output_dir;
         num_points = (10, 5, 3),
+        start_date = Dates.DateTime(2015, 2, 2),
     )
 
     function compute_my_var!(out, u, p, t)
@@ -126,6 +129,8 @@ end
                 @test nc["YO"].attrib["short_name"] == "YO"
                 @test nc["YO"].attrib["long_name"] == "YO YO, Instantaneous"
                 @test size(nc["YO"]) == (11, 10, 5, 10)
+                @test nc["YO"].attrib["start_date"] ==
+                string(Dates.DateTime(2015, 2, 2))
             end
 
             NCDatasets.NCDataset(