matplotlib · tacaswell · Nov 24, 2016 · Nov 21, 2016 · Nov 21, 2016 · Nov 21, 2016
diff --git a/lib/matplotlib/category.py b/lib/matplotlib/category.py
@@ -14,25 +14,6 @@
 import matplotlib.ticker as ticker
 
 
-#  pure hack for numpy 1.6 support
-from distutils.version import LooseVersion
-
-NP_NEW = (LooseVersion(np.version.version) >= LooseVersion('1.7'))
-
-
-def to_array(data, maxlen=100):
-    if NP_NEW:
-        return np.array(data, dtype=np.unicode)
-    if cbook.is_scalar_or_string(data):
-        data = [data]
-    try:
-        vals = np.array(data, dtype=('|S', maxlen))
-    except UnicodeEncodeError:
-        # this yields gibberish
-        vals = np.array([convert_to_string(d) for d in data])
-    return vals
-
-
 class StrCategoryConverter(units.ConversionInterface):
     @staticmethod
     def convert(value, unit, axis):
@@ -44,7 +25,7 @@ def convert(value, unit, axis):
         if isinstance(value, six.string_types):
             return vmap[value]
 
-        vals = to_array(value)
+        vals = np.array(value, dtype=np.unicode)
         for lab, loc in vmap.items():
             vals[vals == lab] = loc
 
@@ -79,25 +60,6 @@ def __init__(self, seq):
         self.offset_string = ''
 
 
-def convert_to_string(value):
-    """Helper function for numpy 1.6, can be replaced with
-    np.array(...,dtype=unicode) for all later versions of numpy"""
-
-    if isinstance(value, six.string_types):
-        pass
-    elif np.isfinite(value):
-        value = np.asarray(value, dtype=str)[np.newaxis][0]
-    elif np.isnan(value):
-        value = 'nan'
-    elif np.isposinf(value):
-        value = 'inf'
-    elif np.isneginf(value):
-        value = '-inf'
-    else:
-        raise ValueError("Unconvertable {}".format(value))
-    return value
-
-
 class UnitData(object):
     # debatable makes sense to special code missing values
     spdict = {'nan': -1.0, 'inf': -2.0, '-inf': -3.0}
@@ -119,12 +81,11 @@ def update(self, new_data):
         self._set_seq_locs(new_data, value)
 
     def _set_seq_locs(self, data, value):
-        # magic to make it work under np1.6
-        strdata = to_array(data)
+        strdata = np.array(data, dtype=np.unicode)
         # np.unique makes dateframes work
         new_s = [d for d in np.unique(strdata) if d not in self.seq]
         for ns in new_s:
-            self.seq.append(convert_to_string(ns))
+            self.seq.append(ns)
             if ns in UnitData.spdict.keys():
                 self.locs.append(UnitData.spdict[ns])
             else:

diff --git a/lib/matplotlib/cbook.py b/lib/matplotlib/cbook.py
@@ -2582,23 +2582,6 @@ def get_label(y, default_name):
     except AttributeError:
         return default_name
 
-# Numpy > 1.6.x deprecates putmask in favor of the new copyto.
-# So long as we support versions 1.6.x and less, we need the
-# following local version of putmask.  We choose to make a
-# local version of putmask rather than of copyto because the
-# latter includes more functionality than the former. Therefore
-# it is easy to make a local version that gives full putmask
-# behavior, but duplicating the full copyto behavior would be
-# more difficult.
-
-try:
-    np.copyto
-except AttributeError:
-    _putmask = np.putmask
-else:
-    def _putmask(a, mask, values):
-        return np.copyto(a, values, where=mask)
-
 _lockstr = """\
 LOCKERROR: matplotlib is trying to acquire the lock
     {!r}

diff --git a/lib/matplotlib/colors.py b/lib/matplotlib/colors.py
@@ -473,7 +473,7 @@ def __call__(self, X, alpha=None, bytes=False):
             # Treat 1.0 as slightly less than 1.
             vals = np.array([1, 0], dtype=xa.dtype)
             almost_one = np.nextafter(*vals)
-            cbook._putmask(xa, xa == 1.0, almost_one)
+            np.copyto(xa, almost_one, where=xa == 1.0)
             # The following clip is fast, and prevents possible
             # conversion of large positive values to negative integers.
 
@@ -482,15 +482,15 @@ def __call__(self, X, alpha=None, bytes=False):
 
             # ensure that all 'under' values will still have negative
             # value after casting to int
-            cbook._putmask(xa, xa < 0.0, -1)
+            np.copyto(xa, -1, where=xa < 0.0)
             xa = xa.astype(int)
         # Set the over-range indices before the under-range;
         # otherwise the under-range values get converted to over-range.
-        cbook._putmask(xa, xa > self.N - 1, self._i_over)
-        cbook._putmask(xa, xa < 0, self._i_under)
+        np.copyto(xa, self._i_over, where=xa > self.N - 1)
+        np.copyto(xa, self._i_under, where=xa < 0)
         if mask_bad is not None:
             if mask_bad.shape == xa.shape:
-                cbook._putmask(xa, mask_bad, self._i_bad)
+                np.copyto(xa, self._i_bad, where=mask_bad)
             elif mask_bad:
                 xa.fill(self._i_bad)
         if bytes:
@@ -990,7 +990,7 @@ def __call__(self, value, clip=None):
                 mask = (resdat <= 0)
             else:
                 mask |= resdat <= 0
-            cbook._putmask(resdat, mask, 1)
+            np.copyto(resdat, 1, where=mask)
             np.log(resdat, resdat)
             resdat -= np.log(vmin)
             resdat /= (np.log(vmax) - np.log(vmin))

diff --git a/lib/matplotlib/quiver.py b/lib/matplotlib/quiver.py
@@ -707,8 +707,8 @@ def _h_arrows(self, length):
         Y0 = shrink * Y0[np.newaxis, :]
         short = np.repeat(length < minsh, 8, axis=1)
         # Now select X0, Y0 if short, otherwise X, Y
-        cbook._putmask(X, short, X0)
-        cbook._putmask(Y, short, Y0)
+        np.copyto(X, X0, where=short)
+        np.copyto(Y, Y0, where=short)
         if self.pivot == 'middle':
             X -= 0.5 * X[:, 3, np.newaxis]
         elif self.pivot == 'tip':
@@ -728,8 +728,8 @@ def _h_arrows(self, length):
             X1 = np.repeat(x1[np.newaxis, :], N, axis=0)
             Y1 = np.repeat(y1[np.newaxis, :], N, axis=0)
             tooshort = np.repeat(tooshort, 8, 1)
-            cbook._putmask(X, tooshort, X1)
-            cbook._putmask(Y, tooshort, Y1)
+            np.copyto(X, X1, where=tooshort)
+            np.copyto(Y, Y1, where=tooshort)
         # Mask handling is deferred to the caller, _make_verts.
         return X, Y
 

diff --git a/lib/matplotlib/tests/test_category.py b/lib/matplotlib/tests/test_category.py
@@ -13,17 +13,6 @@
 import unittest
 
 
-class TestConvertToString(object):
-    testdata = [("abc", "abc"), ("Здравствуйте мир", "Здравствуйте мир"),
-                ("3.14", 3.14), ("nan", np.nan),
-                ("inf", np.inf), ("-inf", -np.inf)]
-    ids = ["string", "unicode", "decimal", "nan", "posinf", "neginf", ]
-
-    @pytest.mark.parametrize("expected, test", testdata, ids=ids)
-    def test_convert_to_string(self, expected, test):
-        assert expected == cat.convert_to_string(test)
-
-
 class TestUnitData(object):
     testdata = [("hello world", ["hello world"], [0]),
                 ("Здравствуйте мир", ["Здравствуйте мир"], [0]),
@@ -157,7 +146,6 @@ def axis_test(self, axis, ticks, labels, unit_data):
 
     @cleanup
     def test_plot_unicode(self):
-        # Image test would fail on numpy 1.6
         words = ['Здравствуйте', 'привет']
         locs = [0.0, 1.0]
         unit_data = MockUnitData(zip(words, locs))

diff --git a/lib/matplotlib/tri/triinterpolate.py b/lib/matplotlib/tri/triinterpolate.py
@@ -1265,12 +1265,9 @@ def dot(self, V):
         *V* dense vector of shape (self.m,)
         """
         assert V.shape == (self.m,)
-        # For a more generic implementation we could use below kw argument
-        # minlength=self.m of bincount ; however:
-        # - it is new in numpy 1.6
-        # - it is unecessary when each row have at least 1 entry in global
-        #   matrix, which is the case here.
-        return np.bincount(self.rows, weights=self.vals*V[self.cols])
+        return np.bincount(self.rows,
+                           weights=self.vals*V[self.cols],
+                           minlength=self.m)
 
     def compress_csc(self):
         """