Make AffineScalarFunc hashable for Pandas, Pint and Pint-Pandas

appveyor.yml

@@ -47,7 +47,7 @@ install:
   - conda init cmd.exe
   - conda info -a
   # Create a conda virtual environement 
-  - "conda create -n uncty-env numpy nose python=%PYTHON_VERSION%"
+  - "conda create -n uncty-env future numpy nose python=%PYTHON_VERSION%"
   - activate uncty-env
 
 
@@ -56,4 +56,5 @@ test_script:
   - "cd C:\\projects\\uncertainties"
   - activate uncty-env # Activate the virtual environment
   - python setup.py egg_info
-  - python setup.py nosetests -sv
+  # Try to work around nose-1.3.7 not working with modern setuptools (> Python 3.5)
+  - nosetests -sv uncertainties/

uncertainties/core.py

@@ -1849,6 +1849,15 @@ def std_dev(self):
     # Abbreviation (for formulas, etc.):
     s = std_dev
 
+    def __hash__(self):
+        if not self._linear_part.expanded():
+           self._linear_part.expand()
+        combo = tuple(iter(self._linear_part.linear_combo.items()))
+        if len(combo) > 1 or combo[0][1] != 1.0:
+           return hash(combo)
+        # The unique value that comes from a unique variable (which it also hashes to)
+        return id(combo[0][0])
+
     def __repr__(self):
         # Not putting spaces around "+/-" helps with arrays of
         # Variable, as each value with an uncertainty is a
@@ -2821,7 +2830,19 @@ def __hash__(self):
         # variables, so they never compare equal; therefore, their
         # id() are allowed to differ
         # (http://docs.python.org/reference/datamodel.html#object.__hash__):
-        return id(self)
+
+        # Also, since the _linear_part of a variable is based on self, we can use
+        # that as a hash (uniqueness of self), which allows us to also
+        # preserve the invariance that x == y implies hash(x) == hash(y)
+        if hasattr(self, '_linear_part'):
+            if (
+                hasattr(self._linear_part, 'linear_combo')
+                and self in iter(self._linear_part.linear_combo.keys())
+            ):
+                return id(tuple(iter(self._linear_part.linear_combo.keys()))[0])
+            return hash(self._linear_part)
+        else:
+            return id(self)
 
     def __copy__(self):
         """

uncertainties/test_uncertainties.py

@@ -2404,3 +2404,21 @@ def test_correlated_values_correlation_mat():
         assert arrays_close(
             numpy.array(cov_mat),
             numpy.array(uncert_core.covariance_matrix([x2, y2, z2])))
+
+    def test_hash():
+        '''
+        Tests the invariance that if x==y, then hash(x)==hash(y)
+        '''
+
+        x = ufloat(1.23, 2.34)
+        y = ufloat(1.23, 2.34)
+        # nominal values and std_dev terms are equal, but...
+        assert x.n==y.n and x.s==y.s
+        # ...x and y are independent variables, therefore not equal as uncertain numbers
+        assert x != y
+        assert hash(x) != hash(y)
+
+        # the equation (2x+x)/3 is equal to the variable x, so...
+        assert ((2*x+x)/3)==x
+        # ...hash of the equation and the variable should be equal
+        assert hash((2*x+x)/3)==hash(x)