SSUSI bug

CHANGELOG.md

@@ -9,10 +9,11 @@ This project adheres to [Semantic Versioning](https://semver.org/).
   * MAVEN SEP
   * MAVEN in situ key parameters
   * REACH Dosimeter
-  * DMSP SSUSI SDR-disk data
+  * DMSP SSUSI SDR-disk and SDR2-disk data
 * New Features
   * Allow files to be unzipped after download
-  * Added custom `concat_data` method to TIMED-GUVI data
+  * Added custom `concat_data` method to JHUAPL methods, for TIMED-GUVI and
+    DMSP-SSUSI data
   * Added cleaning to TIMED-GUVI SDR imaging data
 * Bug Fixes
   * Fix general clean routine to skip transformation matrices

pysatNASA/instruments/dmsp_ssusi.py

@@ -84,6 +84,9 @@
 
 _test_dates = {inst_id: {tag: dt.datetime(2015, 1, 1) for tag in tags.keys()}
                for inst_id in inst_ids.keys()}
+# TODO(#218, #222): Remove when compliant with multi-day load tests
+_new_tests = {inst_id: {tag: False for tag in tags.keys()}
+              for inst_id in inst_ids.keys()}
 # TODO(pysat#1196): Un-comment when pysat bug is fixed and released
 # _clean_warn = {inst_id: {tag: mm_nasa.clean_warnings
 #                          for tag in inst_ids[inst_id]
@@ -93,12 +96,8 @@
 # ----------------------------------------------------------------------------
 # Instrument methods
 
-
 # Use standard init routine
 init = functools.partial(mm_nasa.init, module=mm_dmsp, name=name)
-# TODO(#218, #222): Remove when compliant with multi-day load tests
-_new_tests = {inst_id: {tag: False for tag in tags.keys()}
-              for inst_id in inst_ids.keys()}
 
 
 def clean(self):
@@ -122,6 +121,38 @@ def clean(self):
     return
 
 
+def concat_data(self, new_data, combine_times=False, **kwargs):
+    """Concatonate data to self.data for DMSP SSUSI data.
+
+    Parameters
+    ----------
+    new_data : xarray.Dataset or list of such objects
+        New data objects to be concatonated
+    combine_times : bool
+        For SDR data, optionally combine the different datetime coordinates
+        into a single time coordinate (default=False)
+    **kwargs : dict
+        Optional keyword arguments passed to xr.concat
+
+    Note
+    ----
+    For xarray, `dim=Instrument.index.name` is passed along to xarray.concat
+    except if the user includes a value for dim as a keyword argument.
+
+    """
+    # Establish the time dimensions by data type
+    time_dims = [self.index.name]
+
+    if self.tag in ['sdr-disk', 'sdr2-dist']:
+        time_dims.append('time_auroral')
+
+    # Concatonate using the appropriate method for the number of time
+    # dimensions
+    jhuapl.concat_data(self, time_dims, new_data, combine_times=combine_times,
+                       **kwargs)
+    return
+
+
 # ----------------------------------------------------------------------------
 # Instrument functions
 #

pysatNASA/instruments/methods/jhuapl.py

@@ -38,19 +38,27 @@ def build_dtimes(data, var, epoch=None, epoch_var='time'):
     skey = 'TIME{:s}'.format(var)
 
     if epoch is None:
-        hours = [int(np.floor(sec / 3600.0)) for sec in data[skey].values]
+        hours = np.array([int(np.floor(sec / 3600.0))
+                          for sec in data[skey].values])
         mins = [int(np.floor((sec - hours[i] * 3600) / 60.0))
                 for i, sec in enumerate(data[skey].values)]
         secs = [int(np.floor((sec - hours[i] * 3600 - mins[i] * 60)))
                 for i, sec in enumerate(data[skey].values)]
         microsecs = [int(np.floor((sec - hours[i] * 3600 - mins[i] * 60
                                    - secs[i]) * 1.0e6))
                      for i, sec in enumerate(data[skey].values)]
+        days = np.array([int(dval) for dval in data[dkey].values])
+
+        # Ensure hours are within a realistic range. Datetime can handle day of
+        # roll-over for non-leap years.
+        days[hours >= 24] += 1
+        hours[hours >= 24] -= 24
+
         dtimes = [
             dt.datetime.strptime(
                 "{:4d}-{:03d}-{:02d}-{:02d}-{:02d}-{:06d}".format(
-                    int(data[ykey].values[i]), int(data[dkey].values[i]),
-                    hours[i], mins[i], secs[i], microsec), '%Y-%j-%H-%M-%S-%f')
+                    int(data[ykey].values[i]), days[i], hours[i], mins[i],
+                    secs[i], microsec), '%Y-%j-%H-%M-%S-%f')
             for i, microsec in enumerate(microsecs)]
     else:
         dtimes = [
@@ -444,3 +452,76 @@ def clean_by_dqi(inst):
                     inst.data[dat_var].values[dqi_bad] = np.nan
                     inst.meta[dat_var] = {inst.meta.labels.fill_val: np.nan}
     return
+
+
+def concat_data(inst, time_dims, new_data, combine_times=False, **kwargs):
+    """Concatonate data to inst.data for JHU APL SDR data.
+
+    Parameters
+    ----------
+    inst : pysat.Instrument
+        Object containing a JHU APL Instrument with data
+    time_dims : list
+        List of the time dimensions
+    new_data : xarray.Dataset or list of such objects
+        New data objects to be concatonated
+    combine_times : bool
+        For SDR data, optionally combine the different datetime coordinates
+        into a single time coordinate (default=False)
+    **kwargs : dict
+        Optional keyword arguments passed to xr.concat
+
+    Note
+    ----
+    For xarray, `dim=Instrument.index.name` is passed along to xarray.concat
+    except if the user includes a value for dim as a keyword argument.
+
+    """
+    # Concatonate using the appropriate method for the number of time
+    # dimensions
+    if len(time_dims) == 1:
+        # There is only one time dimensions, but other dimensions may
+        # need to be adjusted
+        new_data = pysat.utils.coords.expand_xarray_dims(
+            new_data, inst.meta, exclude_dims=time_dims)
+
+        # Combine the data
+        inst.data = xr.combine_by_coords(new_data, **kwargs)
+    else:
+        inners = None
+        for ndata in new_data:
+            # Separate into inner datasets
+            inner_keys = {dim: [key for key in ndata.keys()
+                                if dim in ndata[key].dims] for dim in time_dims}
+            inner_dat = {dim: ndata.get(inner_keys[dim]) for dim in time_dims}
+
+            # Add 'single_var's into 'time' dataset to keep track
+            sv_keys = [val.name for val in ndata.values()
+                       if 'single_var' in val.dims]
+            singlevar_set = ndata.get(sv_keys)
+            inner_dat[inst.index.name] = xr.merge([inner_dat[inst.index.name],
+                                                   singlevar_set])
+
+            # Concatenate along desired dimension with previous data
+            if inners is None:
+                # No previous data, assign the data separated by dimension
+                inners = dict(inner_dat)
+            else:
+                # Concatenate with existing data
+                inners = {dim: xr.concat([inners[dim], inner_dat[dim]],
+                                         dim=dim) for dim in time_dims}
+
+        # Combine all time dimensions
+        if inners is not None:
+            if combine_times:
+                data_list = pysat.utils.coords.expand_xarray_dims(
+                    [inners[dim] if dim == inst.index.name else
+                     inners[dim].rename_dims({dim: inst.index.name})
+                     for dim in time_dims if len(inners[dim].dims) > 0],
+                    inst.meta, dims_equal=False)
+            else:
+                data_list = [inners[dim] for dim in time_dims]
+
+            # Combine all the data, indexing along time
+            inst.data = xr.merge(data_list)
+    return

pysatNASA/instruments/timed_guvi.py

@@ -60,7 +60,6 @@
 
 import datetime as dt
 import functools
-import xarray as xr
 
 import pysat
 from pysat.instruments.methods import general as mm_gen
@@ -159,51 +158,8 @@ def concat_data(self, new_data, combine_times=False, **kwargs):
 
     # Concatonate using the appropriate method for the number of time
     # dimensions
-    if len(time_dims) == 1:
-        # There is only one time dimensions, but other dimensions may
-        # need to be adjusted
-        new_data = pysat.utils.coords.expand_xarray_dims(
-            new_data, self.meta, exclude_dims=time_dims)
-
-        # Combine the data
-        self.data = xr.combine_by_coords(new_data, **kwargs)
-    else:
-        inners = None
-        for ndata in new_data:
-            # Separate into inner datasets
-            inner_keys = {dim: [key for key in ndata.keys()
-                                if dim in ndata[key].dims] for dim in time_dims}
-            inner_dat = {dim: ndata.get(inner_keys[dim]) for dim in time_dims}
-
-            # Add 'single_var's into 'time' dataset to keep track
-            sv_keys = [val.name for val in ndata.values()
-                       if 'single_var' in val.dims]
-            singlevar_set = ndata.get(sv_keys)
-            inner_dat[self.index.name] = xr.merge([inner_dat[self.index.name],
-                                                   singlevar_set])
-
-            # Concatenate along desired dimension with previous data
-            if inners is None:
-                # No previous data, assign the data separated by dimension
-                inners = dict(inner_dat)
-            else:
-                # Concatenate with existing data
-                inners = {dim: xr.concat([inners[dim], inner_dat[dim]],
-                                         dim=dim) for dim in time_dims}
-
-        # Combine all time dimensions
-        if inners is not None:
-            if combine_times:
-                data_list = pysat.utils.coords.expand_xarray_dims(
-                    [inners[dim] if dim == self.index.name else
-                     inners[dim].rename_dims({dim: self.index.name})
-                     for dim in time_dims if len(inners[dim].dims) > 0],
-                    self.meta, dims_equal=False)
-            else:
-                data_list = [inners[dim] for dim in time_dims]
-
-            # Combine all the data, indexing along time
-            self.data = xr.merge(data_list)
+    jhuapl.concat_data(self, time_dims, new_data, combine_times=combine_times,
+                       **kwargs)
     return