climate-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From good...@apache.org
Subject climate git commit: ntime % nmonth should always be 0 for reshape / rebin
Date Mon, 17 Apr 2017 20:52:05 GMT
Repository: climate
Updated Branches:
  refs/heads/master 0ed13a19c -> 169f182b2


ntime % nmonth should always be 0 for reshape / rebin


Project: http://git-wip-us.apache.org/repos/asf/climate/repo
Commit: http://git-wip-us.apache.org/repos/asf/climate/commit/169f182b
Tree: http://git-wip-us.apache.org/repos/asf/climate/tree/169f182b
Diff: http://git-wip-us.apache.org/repos/asf/climate/diff/169f182b

Branch: refs/heads/master
Commit: 169f182b29191f9190ca4aa2a388005568556f18
Parents: 0ed13a1
Author: Alex Goodman <agoodm@users.noreply.github.com>
Authored: Mon Apr 17 13:36:45 2017 -0700
Committer: Alex Goodman <agoodm@users.noreply.github.com>
Committed: Mon Apr 17 13:36:45 2017 -0700

----------------------------------------------------------------------
 ocw/dataset_processor.py |  9 +++++++--
 ocw/utils.py             | 33 +++++++++++++++++++--------------
 2 files changed, 26 insertions(+), 16 deletions(-)
----------------------------------------------------------------------


http://git-wip-us.apache.org/repos/asf/climate/blob/169f182b/ocw/dataset_processor.py
----------------------------------------------------------------------
diff --git a/ocw/dataset_processor.py b/ocw/dataset_processor.py
index a99cba9..db4a54f 100755
--- a/ocw/dataset_processor.py
+++ b/ocw/dataset_processor.py
@@ -77,7 +77,12 @@ def temporal_subset(target_dataset, month_start, month_end,
     if average_each_year:
         nmonth = len(month_index)
         ntime = new_dataset.times.size
-        nyear = ntime/nmonth
+        nyear = ntime // nmonth
+        if ntime % nmonth != 0:
+            raise ValueError("Number of times in dataset ({}) does not "
+                             "divide evenly into {} year(s)."
+                             .format(ntime, nyear))
+
         averaged_time = []
         ny, nx = target_dataset.values.shape[1:]
         averaged_values = ma.zeros([nyear, ny, nx])
@@ -536,7 +541,7 @@ def temporal_slice(target_dataset, start_time, end_time):
         target_dataset.lats,
         target_dataset.lons,
         new_times,
-        new_values, 
+        new_values,
         variable=target_dataset.variable,
         units=target_dataset.units,
         origin=target_dataset.origin)

http://git-wip-us.apache.org/repos/asf/climate/blob/169f182b/ocw/utils.py
----------------------------------------------------------------------
diff --git a/ocw/utils.py b/ocw/utils.py
index 0d9eb4e..e9d8d75 100755
--- a/ocw/utils.py
+++ b/ocw/utils.py
@@ -272,7 +272,12 @@ def reshape_monthly_to_annually(dataset):
     values = dataset.values[:]
     data_shape = values.shape
     num_total_month = data_shape[0]
-    num_year = num_total_month/12
+    num_year = num_total_month // 12
+    if num_total_month % 12 != 0:
+        raise ValueError("Number of months in dataset ({}) does not "
+                         "divide evenly into {} year(s)."
+                         .format(num_total_month, num_year))
+
     num_month = 12
     year_month_shape = num_year, num_month
     lat_lon_shape = data_shape[1:]
@@ -600,12 +605,12 @@ def propagate_spatial_mask_over_time(data_array, mask):
             new_mask = mask
             for it in np.arange(nt):
                 new_data_array[it,:] = ma.array(data_array[it,:],
-                                            mask=new_mask)            
+                                            mask=new_mask)
         else:
             for it in np.arange(nt):
                 new_mask = data_array[it, :].mask | mask
                 new_data_array[it,:] = ma.array(data_array[it,:],
-                                            mask=new_mask)                
+                                            mask=new_mask)
 
         return new_data_array
 
@@ -628,7 +633,7 @@ def _force_unicode(s, encoding='utf-8'):
 def calculate_temporal_trends(dataset):
     ''' Calculate temporal trends in dataset.values
     :param dataset: The dataset from which time values should be extracted.
-    :type dataset: :class:`dataset.Dataset'     
+    :type dataset: :class:`dataset.Dataset'
 
     :returns: Arrays of the temporal trend and standard error
     :rtype: :class:`numpy.ma.core.MaskedArray`
@@ -640,19 +645,19 @@ def calculate_temporal_trends(dataset):
     trend = np.zeros([ny, nx])-999.
     slope_err = np.zeros([ny, nx])-999.
     for iy in np.arange(ny):
-        for ix in np.arange(nx): 
-            if dataset.values[:,iy,ix].count() == nt: 
+        for ix in np.arange(nx):
+            if dataset.values[:,iy,ix].count() == nt:
                 trend[iy,ix], slope_err[iy,ix] = calculate_temporal_trend_of_time_series(
-                                               x, dataset.values[:,iy,ix]) 
+                                               x, dataset.values[:,iy,ix])
     
-    return ma.masked_equal(trend, -999.), ma.masked_equal(slope_err, -999.)    
+    return ma.masked_equal(trend, -999.), ma.masked_equal(slope_err, -999.)
 
 def calculate_ensemble_temporal_trends(timeseries_array, number_of_samples=1000):
-    ''' Calculate temporal trends in an ensemble of time series  
-    :param timeseries_array: Two dimensional array. 1st index: model, 2nd index: time. 
+    ''' Calculate temporal trends in an ensemble of time series
+    :param timeseries_array: Two dimensional array. 1st index: model, 2nd index: time.
     :type timeseries_array: :class:`numpy.ndarray'
 
-    :param sampling: A list whose elements are one-dimensional numpy arrays 
+    :param sampling: A list whose elements are one-dimensional numpy arrays
     :type timeseries_array: :class:`list'
 
     :returns: temporal trend and estimated error from bootstrapping
@@ -676,13 +681,13 @@ def calculate_ensemble_temporal_trends(timeseries_array, number_of_samples=1000)
 def calculate_temporal_trend_of_time_series(x,y):
     ''' Calculate least-square trends (a) in y = ax+b and a's standard error
     :param x: time series
-    :type x: :class:`numpy.ndarray'     
+    :type x: :class:`numpy.ndarray'
 
     :param x: time series
-    :type x: :class:`numpy.ndarray'     
+    :type x: :class:`numpy.ndarray'
 
     :returns: temporal trend and standard error
     :rtype: :float:`float','float'
     '''
     slope, intercept, r_value, p_value, std_err = stats.linregress(x,y)
-    return slope, std_err                   
+    return slope, std_err


Mime
View raw message