✅ Properly tested deltatime_to_utctime function

Turn the ICESat-2 delta_time to utc_time conversion code in our jupyter notebook into a well tested function! The cool bit is that we can pass in either a dask or numpy backed xarray.DataArray, and get the equivalent output, with dimensions and coordinates preserved! Gotta love [NEP18](https://numpy.org/neps/nep-0018-array-function-protocol.html). Added a chunks statement to test_catalog.yaml, and ensure the file is cached in a relative path. Had to make sure the atl11_dataset is closed after each test or subsequent tests will see a numpy.array instead of a dask.array, should do proper setup/teardown next time or test everything using dask. Also bumping up cftime from 1.1.1.2 to 1.1.3 to bust the CI cache, just in case.

atl11_play.ipynb

@@ -20,7 +20,6 @@
     "import glob\n",
     "\n",
     "import deepicedrain\n",
-    "import pointCollection.is2_calendar\n",
     "\n",
     "import dask\n",
     "import dask.array\n",
@@ -214,18 +213,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "ICESAT2_EPOCH = np.datetime64(pointCollection.is2_calendar.t_0())\n",
-    "# ICESAT2_EPOCH = np.datetime64(datetime.datetime(2018, 1, 1, 0, 0, 0))"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "utc_time = dask.array.asarray(ICESAT2_EPOCH) + ds.delta_time.data\n",
-    "ds[\"utc_time\"] = xr.DataArray(data=utc_time, coords=ds.delta_time.coords)"
+    "ds[\"utc_time\"] = deepicedrain.deltatime_to_utctime(dataarray=ds.delta_time)"
    ]
   },
   {
@@ -284,7 +272,9 @@
     "        ymin=-699564.516934089,\n",
     "        ymax=-365489.6822096751,\n",
     "    ),\n",
-    "    \"antarctica\": deepicedrain.Region(\"Antarctica\", -2700000, 2800000, -2200000, 2300000),\n",
+    "    \"antarctica\": deepicedrain.Region(\n",
+    "        \"Antarctica\", -2700000, 2800000, -2200000, 2300000\n",
+    "    ),\n",
     "    \"siple_coast\": deepicedrain.Region(\n",
     "        \"Siple Coast\", -1000000, 250000, -1000000, -100000\n",
     "    ),\n",

atl11_play.py

@@ -24,7 +24,6 @@
 import glob
 
 import deepicedrain
-import pointCollection.is2_calendar
 
 import dask
 import dask.array
@@ -118,12 +117,7 @@
 # in the future.
 
 # %%
-ICESAT2_EPOCH = np.datetime64(pointCollection.is2_calendar.t_0())
-# ICESAT2_EPOCH = np.datetime64(datetime.datetime(2018, 1, 1, 0, 0, 0))
-
-# %%
-utc_time = dask.array.asarray(ICESAT2_EPOCH) + ds.delta_time.data
-ds["utc_time"] = xr.DataArray(data=utc_time, coords=ds.delta_time.coords)
+ds["utc_time"] = deepicedrain.deltatime_to_utctime(dataarray=ds.delta_time)
 
 # %% [markdown]
 # ## Mask out low quality height data

deepicedrain/__init__.py

@@ -5,7 +5,7 @@
 
 import deepicedrain
 from deepicedrain.deltamath import calculate_delta
-from deepicedrain.spatiotemporal import Region
+from deepicedrain.spatiotemporal import Region, deltatime_to_utctime
 
 __version__: str = "0.1.0"
 

deepicedrain/spatiotemporal.py

@@ -3,6 +3,7 @@
 Does bounding box region subsets, coordinate/time conversions, and more!
 """
 import dataclasses
+import datetime
 
 import numpy as np
 import xarray as xr
@@ -55,3 +56,19 @@ def subset(
         )
 
         return ds.where(cond=cond, drop=drop)
+
+
+def deltatime_to_utctime(
+    dataarray: xr.DataArray,
+    start_epoch: np.datetime64 = np.datetime64("2018-01-01T00:00:00.000000"),
+) -> xr.DataArray:
+    """
+    Converts GPS time in nanoseconds from an epoch (default is 2018 Jan 1st)
+    to Coordinated Universal Time (UTC).
+
+    Note, does not account for leap seconds! There are none declared since the
+    last one announced on 31/12/2016, so it should be fine for now as of 2020.
+    """
+    utc_time: xr.DataArray = dataarray.__class__(start_epoch) + dataarray
+
+    return utc_time

deepicedrain/tests/test_calculate_delta.py

@@ -25,6 +25,8 @@ def test_calculate_delta_height():
     npt.assert_allclose(actual=delta_height.mean().data, desired=-0.90124122)
     npt.assert_allclose(actual=delta_height.max().data, desired=9.49908442)
 
+    atl11_dataset.close()
+
 
 def test_calculate_delta_time():
     """
@@ -46,3 +48,5 @@ def test_calculate_delta_time():
     npt.assert_equal(
         actual=delta_time.max().data, desired=np.timedelta64(7846787022726588)
     )
+
+    atl11_dataset.close()

deepicedrain/tests/test_catalog.yaml

@@ -2,16 +2,18 @@ metadata:
   version: 1
 sources:
   atl11_test_case:
+    description: 'An example ATL11 hdf5 file for testing various calculations'
     args:
+      chunks:
+        cycle_number: 2
       urlpath: simplecache::https://github.com/suzanne64/ATL11/raw/125ee1a653d78e6b86864b35c9d0fcfd72d64a85/ATL11_test_case/ATL11_078805_0304_02_v002.h5
       xarray_kwargs:
         engine: h5netcdf
         group: /pt2/corrected_h
       storage_options:
         simplecache:
-          cache_storage: tests/test_data
+          cache_storage: '{{ CATALOG_DIR }}/test_data'
           same_names: True
-    description: 'An example ATL11 hdf5 file for testing various calculations'
     driver: intake_xarray.netcdf.NetCDFSource
     metadata:
       coords:

deepicedrain/tests/test_spatiotemporal_conversions.py

@@ -0,0 +1,47 @@
+"""
+Tests various conversions between geospatial and temporal units
+"""
+import datetime
+
+import dask
+import numpy as np
+import numpy.testing as npt
+import pandas as pd
+import xarray as xr
+
+from deepicedrain import catalog, deltatime_to_utctime
+
+
+def test_deltatime_to_utctime():
+    """
+    Test that converting from ICESat-2 delta_time to utc_time works,
+    and that the xarray dimensions are preserved in the process.
+    """
+    atl11_dataset: xr.Dataset = catalog.test_data.atl11_test_case.to_dask()
+
+    utc_time: xr.DataArray = deltatime_to_utctime(dataarray=atl11_dataset.delta_time)
+
+    assert utc_time.shape == (1404, 2)
+    assert utc_time.dims == ("ref_pt", "cycle_number")
+    assert dask.is_dask_collection(utc_time)
+
+    utc_time = utc_time.compute()
+
+    npt.assert_equal(
+        actual=utc_time.data.min(),
+        desired=np.datetime64("2019-05-19T20:53:51.039891534"),
+    )
+    npt.assert_equal(
+        actual=np.datetime64(pd.DataFrame(utc_time.data)[0].mean()),
+        desired=np.datetime64("2019-05-19 20:54:00.925868"),
+    )
+    npt.assert_equal(
+        actual=np.datetime64(pd.DataFrame(utc_time.data)[1].mean()),
+        desired=np.datetime64("2019-08-18 16:33:47.791226"),
+    )
+    npt.assert_equal(
+        actual=utc_time.data.max(),
+        desired=np.datetime64("2019-08-18T16:33:57.834610209"),
+    )
+
+    atl11_dataset.close()