From ca5b2e07ec183f7c1252fd3cfb3df58befa05bc2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=D0=90=D0=BD=D0=B4=D1=80=D0=B5=D0=B9=20=D0=9F=D0=B0=D1=80?=
 =?UTF-8?q?=D0=B0=D0=BC=D0=BE=D0=BD=D0=BE=D0=B2?= <paramon@lims.acdlabs.ru>
Date: Thu, 11 Jan 2018 13:08:37 +0300
Subject: [PATCH 1/3] PowerNorm: do not clip negative values

---
 lib/matplotlib/colors.py | 13 +------------
 1 file changed, 1 insertion(+), 12 deletions(-)

diff --git a/lib/matplotlib/colors.py b/lib/matplotlib/colors.py
index 84d4bbcf1fb0..356e523a91a5 100644
--- a/lib/matplotlib/colors.py
+++ b/lib/matplotlib/colors.py
@@ -53,7 +53,6 @@
 from collections import Sized
 import itertools
 import re
-import warnings
 
 import numpy as np
 import matplotlib.cbook as cbook
@@ -1177,7 +1176,7 @@ def autoscale_None(self, A):
 class PowerNorm(Normalize):
     """
     Normalize a given value to the ``[0, 1]`` interval with a power-law
-    scaling. This will clip any negative data points to 0.
+    scaling.
     """
     def __init__(self, gamma, vmin=None, vmax=None, clip=False):
         Normalize.__init__(self, vmin, vmax, clip)
@@ -1197,7 +1196,6 @@ def __call__(self, value, clip=None):
         elif vmin == vmax:
             result.fill(0)
         else:
-            res_mask = result.data < 0
             if clip:
                 mask = np.ma.getmask(result)
                 result = np.ma.array(np.clip(result.filled(vmax), vmin, vmax),
@@ -1208,7 +1206,6 @@ def __call__(self, value, clip=None):
             resdat /= (vmax - vmin) ** gamma
 
             result = np.ma.array(resdat, mask=result.mask, copy=False)
-            result[res_mask] = 0
         if is_scalar:
             result = result[0]
         return result
@@ -1230,10 +1227,6 @@ def autoscale(self, A):
         Set *vmin*, *vmax* to min, max of *A*.
         """
         self.vmin = np.ma.min(A)
-        if self.vmin < 0:
-            self.vmin = 0
-            warnings.warn("Power-law scaling on negative values is "
-                          "ill-defined, clamping to 0.")
         self.vmax = np.ma.max(A)
 
     def autoscale_None(self, A):
@@ -1241,10 +1234,6 @@ def autoscale_None(self, A):
         A = np.asanyarray(A)
         if self.vmin is None and A.size:
             self.vmin = A.min()
-            if self.vmin < 0:
-                self.vmin = 0
-                warnings.warn("Power-law scaling on negative values is "
-                              "ill-defined, clamping to 0.")
         if self.vmax is None and A.size:
             self.vmax = A.max()
 

From 5a1ed892eadf86d606f8e5a9dbc244c107ac9a3e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=D0=90=D0=BD=D0=B4=D1=80=D0=B5=D0=B9=20=D0=9F=D0=B0=D1=80?=
 =?UTF-8?q?=D0=B0=D0=BC=D0=BE=D0=BD=D0=BE=D0=B2?= <paramon@lims.acdlabs.ru>
Date: Mon, 15 Jan 2018 12:11:29 +0300
Subject: [PATCH 2/3] Fix PowerNorm description

---
 lib/matplotlib/colors.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/matplotlib/colors.py b/lib/matplotlib/colors.py
index 356e523a91a5..c3c779b38266 100644
--- a/lib/matplotlib/colors.py
+++ b/lib/matplotlib/colors.py
@@ -1175,8 +1175,8 @@ def autoscale_None(self, A):
 
 class PowerNorm(Normalize):
     """
-    Normalize a given value to the ``[0, 1]`` interval with a power-law
-    scaling.
+    Linearly map a given value to the 0-1 range and then apply
+    a power-law normalization over that range.
     """
     def __init__(self, gamma, vmin=None, vmax=None, clip=False):
         Normalize.__init__(self, vmin, vmax, clip)

From 0c1b027522e946a2e08ccbee30ed875acc5e8d55 Mon Sep 17 00:00:00 2001
From: Paramonov Andrey <Paramonov Andrey>
Date: Wed, 20 Jun 2018 23:58:51 +0300
Subject: [PATCH 3/3] Fix tests

---
 lib/matplotlib/colors.py            |  1 +
 lib/matplotlib/tests/test_colors.py | 17 ++++++++++-------
 2 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/lib/matplotlib/colors.py b/lib/matplotlib/colors.py
index c3c779b38266..03d3856a0b9b 100644
--- a/lib/matplotlib/colors.py
+++ b/lib/matplotlib/colors.py
@@ -1202,6 +1202,7 @@ def __call__(self, value, clip=None):
                                      mask=mask)
             resdat = result.data
             resdat -= vmin
+            resdat[resdat < 0] = 0
             np.power(resdat, gamma, resdat)
             resdat /= (vmax - vmin) ** gamma
 
diff --git a/lib/matplotlib/tests/test_colors.py b/lib/matplotlib/tests/test_colors.py
index 3b507e8fbd21..90b11866fdc7 100644
--- a/lib/matplotlib/tests/test_colors.py
+++ b/lib/matplotlib/tests/test_colors.py
@@ -192,6 +192,15 @@ def test_PowerNorm():
     assert pnorm(a[-1], clip=True) == expected[-1]
 
 
+def test_PowerNorm_translation_invariance():
+    a = np.array([0, 1/2, 1], dtype=float)
+    expected = [0, 1/8, 1]
+    pnorm = mcolors.PowerNorm(vmin=0, vmax=1, gamma=3)
+    assert_array_almost_equal(pnorm(a), expected)
+    pnorm = mcolors.PowerNorm(vmin=-2, vmax=-1, gamma=3)
+    assert_array_almost_equal(pnorm(a - 2), expected)
+
+
 def test_Normalize():
     norm = mcolors.Normalize()
     vals = np.arange(-10, 10, 1, dtype=float)
@@ -711,13 +720,7 @@ def __add__(self, other):
                  mcolors.SymLogNorm(3, vmax=5, linscale=1),
                  mcolors.PowerNorm(1)]:
         assert_array_equal(norm(data.view(MyArray)), norm(data))
-        if isinstance(norm, mcolors.PowerNorm):
-            assert len(recwarn) == 1
-            warn = recwarn.pop(UserWarning)
-            assert ('Power-law scaling on negative values is ill-defined'
-                    in str(warn.message))
-        else:
-            assert len(recwarn) == 0
+        assert len(recwarn) == 0
         recwarn.clear()