doesn't dump file unless asked, reads good var in assert func and sav… (

#28)

* doesn't dump file unless asked, reads good var in assert func and save with correct name

* no restriction on numpy

.circleci/config.yml

@@ -17,7 +17,7 @@ aliases:
   - &create_conda_env
     name: create_conda_env
     environment:
-       PKGS: "cdms2 cdat_info testsrunner udunits2 'numpy<1.14.6'"
+       PKGS: "cdms2 cdat_info testsrunner udunits2 numpy"
        CHANNELS: "-c cdat/label/nightly -c conda-forge -c cdat"
     command: |
        export PATH=$WORKDIR/miniconda/bin:$PATH
@@ -181,10 +181,10 @@ workflows:
   genutil:
     jobs:
       - macos_genutil_py2
-      - macos_genutil_py3:
-           requires:
-              - macos_genutil_py2
+      - macos_genutil_py3
+          # requires:
+          #    - macos_genutil_py2
       - linux_genutil_py2
-      - linux_genutil_py3:
-           requires:
-              - linux_genutil_py2
+      - linux_genutil_py3
+          #     requires:
+          #    - linux_genutil_py2

tests/test_genutil_statistics.py

@@ -3,46 +3,56 @@
 import unittest
 import cdat_info
 import numpy
+dump = False
 
 class GENUTIL(unittest.TestCase):
-    def assertArraysEqual(self,A,B):
-        self.assertTrue(numpy.ma.allclose(A,B))
+    def assertArraysEqual(self,A,nm,skip=False, dump=dump):
+        B = self.good(nm)
+        print("A-B max difference", numpy.abs(A-B).max(), skip)
+        print("A-B % difference", numpy.abs((A-B)/A).max(), skip)
+        if dump:
+            self.out.write(A, id=nm)
+            self.out.sync()
+
+        if not skip:
+            self.assertTrue(numpy.ma.allclose(A,B))
 
     def testStatisitcs(self):
         f=cdms2.open(os.path.join(cdat_info.get_sampledata_path(),'clt.nc'))
 
         u=f('u',time=slice(0,1),level=slice(0,1),squeeze=1)
         v=f('v',time=slice(0,1),plev1=slice(0,1),squeeze=1)
 
-        f=cdms2.open(os.path.join(cdat_info.get_sampledata_path(),"genutil_statistics.nc"))
-        self.out = cdms2.open("genutil_statistics_new.nc","w")
+        self.good = cdms2.open(os.path.join(cdat_info.get_sampledata_path(),"genutil_statistics.nc"))
+        if dump:
+            self.out = cdms2.open("genutil_statistics_new.nc","w")
         print('Lagged correlation')
         nm = "lagged_corr"
-        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0), f( nm+"_1"))
-        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0,lag=4), f( nm+"_2"))
-        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0,lag=4,noloop=1), f( nm+"_3"))
-        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0,lag=[4,8,10]), f(nm+"_4"))
+        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0),  nm+"_1")
+        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0,lag=4), nm+"_2")
+        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0,lag=4,noloop=1), nm+"_3")
+        self.assertArraysEqual(genutil.statistics.laggedcorrelation(u,v,axis=0,lag=[4,8,10]), nm+"_4")
 
 
         print('Lagged covariance')
         nm = "lagged_cov"
-        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0), f(nm+"_1"))
-        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0,lag=4), f(nm+"_2"))
-        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0,lag=4,noloop=1), f(nm+"_3"))
-        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0,lag=[4,8,10]), f(nm+"_4"))
+        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0), nm+"_1")
+        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0,lag=4), nm+"_2")
+        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0,lag=4,noloop=1), nm+"_3")
+        self.assertArraysEqual(genutil.statistics.laggedcovariance(u,v,axis=0,lag=[4,8,10]), nm+"_4")
 
         print('Auto correlation')
         nm = "auto_corr"
-        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0), f(nm+"_1"))
-        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0,lag=4), f(nm+"_2"))
-        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0,lag=4,noloop=1), f(nm+"_3"))
-        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0,lag=[4,8,10]), f(nm+"_4"))
+        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0), nm+"_1")
+        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0,lag=4), nm+"_2")
+        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0,lag=4,noloop=1), nm+"_3")
+        self.assertArraysEqual(genutil.statistics.autocorrelation(u,axis=0,lag=[4,8,10]), nm+"_4")
 
         print('Auto covariance')
         nm = "auto_cov"
-        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0), f(nm+"_1"))
-        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0,lag=4), f(nm+"_2"))
-        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0,lag=4,noloop=1), f(nm+"_3"))
-        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0,lag=[4,8,10]), f(nm+"_4"))
+        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0), nm+"_1")
+        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0,lag=4), nm+"_2")
+        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0,lag=4,noloop=1), nm+"_3")
+        self.assertArraysEqual(genutil.statistics.autocovariance(u,axis=0,lag=[4,8,10]), nm+"_4")
         f.close()