From 0a77469f6271b2309dd3f2e2f47386f3852b17c7 Mon Sep 17 00:00:00 2001
From: Thomas A Caswell <tcaswell@gmail.com>
Date: Mon, 18 Mar 2019 15:45:40 -0400
Subject: [PATCH] Backport PR #12419: Add DivergingNorm (again, again, again)

---
 doc/api/colors_api.rst                        |   1 +
 .../colormap_normalizations_diverging.py      |  44 +++++++++
 lib/matplotlib/colorbar.py                    |  31 +++---
 lib/matplotlib/colors.py                      |  70 +++++++++++++
 .../mpl-data/sample_data/topobathy.npz        | Bin 0 -> 45224 bytes
 lib/matplotlib/tests/test_colorbar.py         |  21 +++-
 lib/matplotlib/tests/test_colors.py           |  90 +++++++++++++++++
 lib/matplotlib/tests/test_contour.py          |   5 +-
 tutorials/colors/colormapnorms.py             |  92 +++++++++++-------
 9 files changed, 300 insertions(+), 54 deletions(-)
 create mode 100644 examples/userdemo/colormap_normalizations_diverging.py
 create mode 100644 lib/matplotlib/mpl-data/sample_data/topobathy.npz

diff --git a/doc/api/colors_api.rst b/doc/api/colors_api.rst
index d691b9905c55..ccd1cf18aa07 100644
--- a/doc/api/colors_api.rst
+++ b/doc/api/colors_api.rst
@@ -25,6 +25,7 @@ Classes
 
    BoundaryNorm
    Colormap
+   DivergingNorm
    LightSource
    LinearSegmentedColormap
    ListedColormap
diff --git a/examples/userdemo/colormap_normalizations_diverging.py b/examples/userdemo/colormap_normalizations_diverging.py
new file mode 100644
index 000000000000..7a5a68c29b73
--- /dev/null
+++ b/examples/userdemo/colormap_normalizations_diverging.py
@@ -0,0 +1,44 @@
+"""
+=====================================
+DivergingNorm colormap normalization
+=====================================
+
+Sometimes we want to have a different colormap on either side of a
+conceptual center point, and we want those two colormaps to have
+different linear scales.  An example is a topographic map where the land
+and ocean have a center at zero, but land typically has a greater
+elevation range than the water has depth range, and they are often
+represented by a different colormap.
+"""
+
+import numpy as np
+import matplotlib.pyplot as plt
+import matplotlib.cbook as cbook
+import matplotlib.colors as colors
+
+filename = cbook.get_sample_data('topobathy.npz', asfileobj=False)
+with np.load(filename) as dem:
+    topo = dem['topo']
+    longitude = dem['longitude']
+    latitude = dem['latitude']
+
+fig, ax = plt.subplots(constrained_layout=True)
+# make a colormap that has land and ocean clearly delineated and of the
+# same length (256 + 256)
+colors_undersea = plt.cm.terrain(np.linspace(0, 0.17, 256))
+colors_land = plt.cm.terrain(np.linspace(0.25, 1, 256))
+all_colors = np.vstack((colors_undersea, colors_land))
+terrain_map = colors.LinearSegmentedColormap.from_list('terrain_map',
+    all_colors)
+
+# make the norm:  Note the center is offset so that the land has more
+# dynamic range:
+divnorm = colors.DivergingNorm(vmin=-500, vcenter=0, vmax=4000)
+
+pcm = ax.pcolormesh(longitude, latitude, topo, rasterized=True, norm=divnorm,
+    cmap=terrain_map,)
+ax.set_xlabel('Lon $[^o E]$')
+ax.set_ylabel('Lat $[^o N]$')
+ax.set_aspect(1 / np.cos(np.deg2rad(49)))
+fig.colorbar(pcm, shrink=0.6, extend='both', label='Elevation [m]')
+plt.show()
diff --git a/lib/matplotlib/colorbar.py b/lib/matplotlib/colorbar.py
index 59f6c9e932c3..fba292890b1a 100644
--- a/lib/matplotlib/colorbar.py
+++ b/lib/matplotlib/colorbar.py
@@ -875,8 +875,8 @@ def _process_values(self, b=None):
                                       + self._boundaries[1:])
                 if isinstance(self.norm, colors.NoNorm):
                     self._values = (self._values + 0.00001).astype(np.int16)
-                return
-            self._values = np.array(self.values)
+            else:
+                self._values = np.array(self.values)
             return
         if self.values is not None:
             self._values = np.array(self.values)
@@ -1113,20 +1113,19 @@ def _locate(self, x):
             b = self.norm(self._boundaries, clip=False).filled()
             xn = self.norm(x, clip=False).filled()
 
-        # The rest is linear interpolation with extrapolation at ends.
-        ii = np.searchsorted(b, xn)
-        i0 = ii - 1
-        itop = (ii == len(b))
-        ibot = (ii == 0)
-        i0[itop] -= 1
-        ii[itop] -= 1
-        i0[ibot] += 1
-        ii[ibot] += 1
-
-        y = self._y
-        db = b[ii] - b[i0]
-        dy = y[ii] - y[i0]
-        z = y[i0] + (xn - b[i0]) * dy / db
+        bunique = b
+        yunique = self._y
+        # trim extra b values at beginning and end if they are
+        # not unique.  These are here for extended colorbars, and are not
+        # wanted for the interpolation.
+        if b[0] == b[1]:
+            bunique = bunique[1:]
+            yunique = yunique[1:]
+        if b[-1] == b[-2]:
+            bunique = bunique[:-1]
+            yunique = yunique[:-1]
+
+        z = np.interp(xn, bunique, yunique)
         return z
 
     def set_alpha(self, alpha):
diff --git a/lib/matplotlib/colors.py b/lib/matplotlib/colors.py
index 0ba21c33beba..bfc53ada8654 100644
--- a/lib/matplotlib/colors.py
+++ b/lib/matplotlib/colors.py
@@ -958,6 +958,76 @@ def scaled(self):
         return self.vmin is not None and self.vmax is not None
 
 
+class DivergingNorm(Normalize):
+    def __init__(self, vcenter, vmin=None, vmax=None):
+        """
+        Normalize data with a set center.
+
+        Useful when mapping data with an unequal rates of change around a
+        conceptual center, e.g., data that range from -2 to 4, with 0 as
+        the midpoint.
+
+        Parameters
+        ----------
+        vcenter : float
+            The data value that defines ``0.5`` in the normalization.
+        vmin : float, optional
+            The data value that defines ``0.0`` in the normalization.
+            Defaults to the min value of the dataset.
+        vmax : float, optional
+            The data value that defines ``1.0`` in the normalization.
+            Defaults to the the max value of the dataset.
+
+        Examples
+        --------
+        This maps data value -4000 to 0., 0 to 0.5, and +10000 to 1.0; data
+        between is linearly interpolated::
+
+            >>> import matplotlib.colors as mcolors
+            >>> offset = mcolors.DivergingNorm(vmin=-4000.,
+                                               vcenter=0., vmax=10000)
+            >>> data = [-4000., -2000., 0., 2500., 5000., 7500., 10000.]
+            >>> offset(data)
+            array([0., 0.25, 0.5, 0.625, 0.75, 0.875, 1.0])
+        """
+
+        self.vcenter = vcenter
+        self.vmin = vmin
+        self.vmax = vmax
+        if vcenter is not None and vmax is not None and vcenter >= vmax:
+            raise ValueError('vmin, vcenter, and vmax must be in '
+                             'ascending order')
+        if vcenter is not None and vmin is not None and vcenter <= vmin:
+            raise ValueError('vmin, vcenter, and vmax must be in '
+                             'ascending order')
+
+    def autoscale_None(self, A):
+        """
+        Get vmin and vmax, and then clip at vcenter
+        """
+        super().autoscale_None(A)
+        if self.vmin > self.vcenter:
+            self.vmin = self.vcenter
+        if self.vmax < self.vcenter:
+            self.vmax = self.vcenter
+
+    def __call__(self, value, clip=None):
+        """
+        Map value to the interval [0, 1]. The clip argument is unused.
+        """
+        result, is_scalar = self.process_value(value)
+        self.autoscale_None(result)  # sets self.vmin, self.vmax if None
+
+        if not self.vmin <= self.vcenter <= self.vmax:
+            raise ValueError("vmin, vcenter, vmax must increase monotonically")
+        result = np.ma.masked_array(
+            np.interp(result, [self.vmin, self.vcenter, self.vmax],
+                      [0, 0.5, 1.]), mask=np.ma.getmask(result))
+        if is_scalar:
+            result = np.atleast_1d(result)[0]
+        return result
+
+
 class LogNorm(Normalize):
     """Normalize a given value to the 0-1 range on a log scale."""
 
diff --git a/lib/matplotlib/mpl-data/sample_data/topobathy.npz b/lib/matplotlib/mpl-data/sample_data/topobathy.npz
new file mode 100644
index 0000000000000000000000000000000000000000..9f9b085fa29bf590f1a3083c7c80baf0295528f3
GIT binary patch
literal 45224
zcmb`w3%pfTx%WR2L{v;n%uLONf_Om%#k^rG=GuTJCMKk&=0foX8%4#uVJ)_ZNQg*^
zhPUO`h!6?yL`sXbkcgwEW~OGgA`cRI$kbyV^w|6Tj=3gGw%g8m-~S&!zt1zqnB)5V
z#`BEJoNGJ$xQ?B=N6Eh)(Fxz!{hzr9qDXWhal%F8FFNqT@mK!f>m!dJerkv4JJHp>
zN1t=)*_ZS_GS_>+x&3?h$@M<>qDv-RGU~!JFS=y(Ii?R8)qLqWq+fdesPX3z-{**a
zeRBN{?z?ZF+%>uXw}0D2<)1E(%8x9MiuZ!q%cK1G<x%cB@I&I0mPeISNIP_S6unT3
zqLaXWU|+Bo*d6Quz5|{DkAo)gWzYkB3G4>U*9~+AA6BDiEqEEM0{4R!a0<wSZlJOZ
zO@Y!cqG(2pw!lt>C|ZM@1r8wG5ex?VgY7{9TvUik*MkLMDYzE=8LTNpl_Q#>{C-VQ
z=|ker67SU%RW1Z4k^W8OK;(hoYfVvoCh~OBwj%#+g(&}R;s+4l6C6%>JJR2wo}0j9
z;4Ux+P;cphq%Wn%2~Q=wGjJTDAK~M{e8!Nb^&z}?(5Cq8>ZWj-r<FV{z`SW%kd-Lv
zpGjLuJ*$X62WEp6aa6fuc~mdKXUp=aa^-Tyiu^kAOgJ45`Vo$nN2OP4QE6c<ioOAQ
zf}O!GU|X;qxB@Hyr-S~W5AeCz2{;erHv;ceqv%CY0rS9Aa6UL3bOG{PZ1C&Y*MfKB
zC|VEH&A>twZ3VUiqrjzLOK=c43=9Gnqu-g}W}yBb0zU_@!B+%7ZxJq$?{egIfV}w=
znxg!XV0U=jmORsW&b%tk22X+Y<T1TUxRqyebp|*I9MOP$zCj;1fF6WLAX82z5|)EB
z-kuy${q*yNnK~>Jp_>S9+m;_s-7k{&Wv~KN<EVT~VyDps4PNE<D{$H$^jaR3c7x;h
zYviv*rFqDH;0vG&_@tWf+mUcDp07fl0KN`}19>*D+&T};w>8)XtfuTr@F=(wlz{dt
zzl-R99@u~R?Gs1QQqZA5Jzxo8Ih5z^(KXyf*MN(_am3~G7Vr^vx=&M-&x5Z5>@L4I
z+<h^LS4gk)fv-;}r!Nrj6}Y?y`7pR1Sl5eyxe}cT4h4Gw=iCm+jt#6JuH8%l`kbyH
z<=}C`+Ep4?&S(=i&3T{AW7zSBw01aE7k2}>od@7F^lKgRHTb;|KCfCH<<CcsLSnc1
z!2o_syOOpG{Qka%Uj&DNFM=Lmd(a7N0pz&@atx5y%WF|-GI9jy2Am5%*Upb>U_E7C
z2hRXGo(c5P-9VP#sr1>te+cHp$ynqWdkMbqWWv~yHYmrXDd3ObPe7Z+UW1SRHC%Ur
z&o4AZ`3^uEhX2yffE;4m`7KF%i!gpY@cez`uE4pqHF5(||1k?_i_M_55$=g>0&=k%
zklR=2bC#>qNlW8XGT}5$sg9J^mrc)x^<&2E%a-&@y~wTYrhRqJ(YF$7Ze=nYeg{6U
zAlwWF6aNDED{Z|B?g7VweSzz#HYvYb0l8fRl*V0qwcks?IADMKf^%z8>272*(8uox
zJ`bE9^1J|?59D{|vcv~DhTG}4eIFLnej(8@KBzPq9_}PAxA0N=70|}!fSbS*U=er>
zXy5qoFds$|J`8AwCxC%qXJA?ZICs>Sb|8oHFroqZu`gL&XT$R8JX{ORSI;MTyH`J5
zMlSk;Y<a^Uw3D}Oo-}S>cLy!h(+M~yUIV+3wk^PC6lW}t^6*=_9A0O#4&MW>-y+R5
z_x@TG-3R1KU%N{K@}iI38f*#HRTF=^En)3+Pw>rJRJs9qH5d!NNBmmkATX8i9OP)Q
z4{%LhOdVH%2;`P+8n-%A$M#=;*1HfzhXBXX3+ebyggb4v45kr32?^(6T|5>khy6fD
za1qc~9)R2$Xm106I$8x5g6&C{-&BwF22W}HU&-P0w;Raj=~-p0Ae)wkqeeS(c;4E8
z&n7m~(|m2EHY~qN=YBelb%dR-`s_0cQ4Ve^-vcKskBZaamc3S~33MU+e2p&EqUdaJ
zD98cVQn4M-_TH@~{H{leV@Z=|Id-pgRzn&-0r0$ZFY-=sA=nRmOgVk@`QZ5sw@q>4
z|2{#g`z;H}-eo7`65>erNTqwg+rT~ZT%K(Qp5yr%pnu&1s0ZiCT;QB(tFLTaJIKb<
z@Q0bcO{tQnl3=arhT$cP!)xfBe7gbV^WY^=0_&8S{4H(b*=PC_<Jmpc%cMCz$J8Bk
z1#+yOoWHH$V#a+nT%NZ)%ClDvIDmLh(7EBCSJKDpz*oTmU@y=ebOrLb75K24_~<u~
zy@($LJ|TQ0QValj&;(AYvEG4cgk7sY2ORgE;24nQcXXVbe~v07wu<e!myNCl;{fTU
zAA`HVbb!wecIvntzhi7^$gIAy;ajSS?phi$`^>$;Y_Jx*0*CXF^5WcftiwSL3<LFc
zI6ft2KJly-Oa$f=4+GN%fDxeCaJ)I+*?y$qEWdNfKMc4Qi2j7Tf+jEp_#A&#xe(hu
zZh4eH0@(w(A97D5{yF-j7L{HGvl!1|K<@ScyEY)d8>)#7uSPm%;T~)tI0HBzz6^Q-
zeLVJDx)Dqv+z+fE|LtHr$b)X66G+G8TCyfi{PN&Ja<=^u@@CStOSvu(m-jQk2jqPX
zXs;rzN2#BbFXxncbv```s$dmxuD*=i8o3DhCfJhjj==qf<Mmlb;CP<|_PrID?sK_M
z*HGK)1LSlT=nv#pU+go}M}ur1<tH0|BNLb3lw;R-*ZQ73m)j`~Y5Do!WiS$c-7k9<
z`W)z#a2jo>Me>{90XzOD{e2Lyo(7-W1*z?-qaA?bdZ(J~>(!C^k>f9rUH~J3>$YV~
z`vz!%->JxBz%%4+0gkl~(ANcik&aW2?gmRi2Y8g*M+#ANDmWYL2zCWWfnLC|k<W9|
zMm-ftTUkx!p}LnxdDs~^SLJRe;CPg#?cRWSgy}-AExV3-Hh?Nv2v&o%4tchnDM0=l
zpLx_xE6|pGHV0(6%!WPV8h<fvKackASjA}23|hbpa5`8OhdoI+3+;~0Vz;F?(feX>
z3;jD2yiwz<Ncy9M_XpabzB$#4e7VnZoG*gqU<G&$n5MlB1kQo=#MObEjs_=zkzgTd
zj#Yc=4mR@3*c{*NIGLNL70@5t;)|o9;0Q1ZXv5}N3MK-5R8~*JY6&O3km_IEACQ4O
z_5gCK^qFJT?<n^L{TgA%ZT=3xF)jgeX4o-0kDVh*=gbDujhkPa|AhDga0+NU&TOB9
zjW*Wv+p$siY}x0##Zfd3ECt&al6|ARZcW&^y%o@h?T-ESfxnO7^d0b9`0RiU-(5@e
ze;)EmPy~+Eemi%+1oi~G0qyeRD(e$i)qr|92o%7(qzy(o-w#1v)4*xS<3N@4hk<L6
zcAFyI%M6Rdn(BEr_}@d28%X;AtN@O)6UdI=uzcEwqrhQ6-VV*g(`PB4N_iCK6PB~Q
zVaLA{kY9D+e097_!HYodwk2JjQ{U!%NueE22UEc@U_ACY3poSWU+1R2HXTRmhtqgg
z_95Th;0<tGf&PJ)2_KKV1{+ufrV!o|ct7<wxP1jI1vk|a+jU>-+&vcT3gm2i&<Qw(
zZoqML2Kwih!1G`&aQ#(J+OFX(fo0U;iL`M8xDL3kn@^vmZ{7{aZ&wiZq>LNiQ@XK`
zoOh=e5}UmUc{Iq5*|54+pUx}V34ON@LqTtl@})co?Aw5SaPFC&BIWo9($eyBxD|1E
z)E=F)b3hBQ?-ih5{t#%_`XzO<7kS;E9ZvWPFb&*JT%Pq?={TC=P2n_6S<#j=Yp*sv
zgS;Kl%S{D*1Q-Mc6aPE+G`S|$+NP+o1<+pc!_jl#esBqJud6=fLz~q`cLvT|d2)P?
zQ{IGg`4u3?j_1pysW0cgP#0@Rw=UxsgE8<s328p(#RQ;^-eKI`k#d}l@#I2scX%Rl
zcc9&-&s-n&2JX@K2l9FXI35fIa(y(gUAdAs`BY}Z@@ijF&eQr+>D+k@IF8i~7+=wV
z{ape~dlNhdhJi<c{c?O$!6`uBQ~`Ynn|?a59nWo@;!R^)(g=sOxZ$tOSI#ppc(zp`
zxu@xX90-QO@3mkFcnbUk|GYE)`9EvCW1$aoz?tA+pbicMI{^7q?wFDCn({1v&Sm|S
z^R^dp*WSUvXK$0{9C#NDB0dhF^OF0l>j~@goYVR+Iqm~`f>gJyG4+FmK;M+6Ppl^U
zP}d3h6mmNXoCs3R<V|~(Pj!6&=nZUF7&d)Zkk;K+%3~GCq2sjeIbb1p9Qe#Ua_?Bx
zhun1s+QfE1T}&pfZ+0G^Pq+m*57PErYo~(jINSQ_JpAq!b7usH3kz@snD^m6wuWqN
zU^%vY;PNQn6YPi`UrqZxz)nE_3<qJYl}E8X&?l#y+h=(e>PX)t<Wc|a8r=oRuXFkh
zz!<|lmTgf_=}s`4u<gmOV|C2wc$F&|@3eUHv@&_rb%Amtf%Ecga4N9hsV&N%W7-8c
zKeqz*QJw2=)st=XY{)dMj;on8)8$eQl@Ei(%M$<U+!XUb6_{r&*Z}Nje_*@vJ{MTt
zF|Gp65$C}OU>jXyo&&>ek%!k5YX>FkR;iGz#oNK@o7gMwQ+$6?>E0BT-Y4z9I18}`
zmwrzBM*;PrE`&S^?O-3E&eV}Ss~=&%wFjR&KOKYP68dTP+wT*WbM<Mu^T)ZNo|!k{
zuHypYGl;9V*}$>g9e;X^&X<>Yz6bI|a0ZZ9!zTfK#t}e!-3#;rj>Yz!Uvt{2v#lP}
z_|^sL0BPLwW-E2b&k|rio@>ZC$lHNZe_H|SNqsdDRyWS^M}c)L0*f2r^O0W!X*=zW
z+h_R<zY}Bh3Fd;?U|+c10bY4;oxcKk8FF+}6kUdX9|Aw7eR)c`-IcIwZraBzr}88}
zax3It|F(hj6y{OS)+y&=O#}E(?>ItRaNu2ZxU-}$VGg9@Y%5=*y!|>BoCf6f1W?Gp
z{nWm|ap+fGq>Rw#yJj`&^!-EB8=M2KNl<D*?g-L6*ApHHKG;+)?3qdf$s^FY&?aYr
z7BB-egQ;LX=vYg5bZ*J1_Phc(FJ^){;27{czQ8?HT4!6C&v08N+!Rv>n9#sjeDPs?
zn=lD(FNNC?fOn$7X65HwKp!cm&N=N?X<RP#&vL2Gl-jO+7v|{=%x_#yj|PVW)Af(5
ztBJIp^T4HWIvIH}_#WXOCNer1DSz6%eqG%e*Dg;5K9^f@d;>m{*CT*!+duW!3iN$#
zbd~=czH4P6yo)HYMwQlrkHDM)Yc%q9VEVma5@|ian)p-ow5>P9ZF4#}1#|_6(U&KY
z`tl0)@4sw+^ve<WEN=0`o7rtSpS+iHFV;O~Jp<D?W88-Qej9EtXo~XRMe;pL=`ZxV
zAMI;<y?}mGJ5@j0-<C+x0ce-+0)48_%(E-dpYH)IuWp6)^#*$ZxmC*jI?5e|JP4SM
z{gp~!2ABz^k~Xf9b{+Bt;&31QuJhmma2^;1wB6&uP#~|?xgBVsjmN=P31{?J`nV94
z{)!HEWBj`^cGs#7=mDRW2Y$m7_AB?2?`<IGi%Gj2bOb#>roIqvjz68I{kDRoU=8RW
zC;Q^@@X_|3!un<|OSpB9k=B27rmb$K<15?`m`7QM<Gzis^{)ZruwT9>%14aJHQ4VK
z7Eq5o$dlae0JJ|jv~TNyoakFzYxHy4h2a&za&jm31)fvn+PHknVaop@gi{V|$GU9G
z@I|$#^nGwWkY~9S@Ey)jmy<_s<#;U6hED>=fun)*ZWb5~QXMs9et*W?oAG`bbYsjL
z&_id^x`G2q8-(12^iK-G2EzH`LBhWVuO<jSd+P#oCr;{@SN)<=Kbf|<u$tKF{m45(
z9UjtWpC#L#eayEa{&YEQwI`5Q>+RGCFDAY%eOpR=4wwx(5Z{5mKY{d)u?zUPnw(u<
zN4n0be>tpxRbY7|tb88mb9}xdW#v;|Q!cY|XVSL?hXUt-d^%sO(>X5mVcP8^AjdZ(
z;c$0$S@JCKdm7=9-~{jq<vJjpYp!>$Ee#!Nr+L1`sPqKP<@`aQAK<&Hz-tH6o$EUT
z?RR_l{e2;Pj}x7WoC*F2m<y%{{S}ahfQ2!02}}dpPRd)CYGM!1Ag6*fJ(cMh$NAvi
zXtVUnrnu>Hx!H5l<o5Okw{`e6tUu76KLl$V=t{mmU;tPRw4ryZ38(5kwN3Rde^nqC
zi^1cKxc0Idm?yPobuE|MgPp+EV5^4IZ<;R5x1>>aK2kWBMv;C2I2)V?<n?lJCOD@d
z_dy;2oG<qKVala?Xh=B=-&Gxq90qqMBF{sf)d1g0=Jx{ZJ;FDW`yuh!emhr@=lkLG
z0T9j>$QKHs4Znpf-9UInoY+b;GHtuKEV)<iN1tBG^hsT3xk~eYO19xQ``o&X=iyXn
zzpnu6b*?zqP6rdfY~cKx0!9FzYquW&d6hqXj8J#-pfr6hFkQ~y0q+BCT`8|}sMJ2S
z>6Bmj6mn~tb7&Tr4dyhYT<r^nf<fR|Fc|2A)3tR0c_%h78ky07=lT3LaHgH&k1I`}
zKp1GJehZsF3h8tC){b{V%IzWG^ThW=g77|>b*uE>z&+NYLeka&$iu*#II#<L7y1A<
zUFc5*=)<NXP2YjMQ^=z&W%?Sz+FUkmGI@3)|Kc|3=JB60pHE<Ya@h=4gEzoZa3XzW
zzYxCBeU7^9t95RuCOX#^<ubLYlw;-HAmzg6>w%of)#?U}?*-NoU(>+5jc3LU+m>@h
zekOn^z<RCo6X4u1&o0Ohkm@X*2U&e&!}=)4x&aI=VE>HqHXt|bp~ClbE0BKz?_s0*
zWY**Sp#WYhI|6vG>_MDwoc(S&-1ALp)U_0uwx?c41O2Q1(DWi<_dheMiBC!MZb!LR
zo;CTLeE$mV@!DQ~=fxE2vtFhBm<4(<zn!D*o$*x_=6Klqe?%YT=tZy;$mQGyT9GMc
z5uO2*OTf!uO+&6mz5-SN`L#aNg!5ok1FlQ9^EgQP)gG-=j;&v5J#PcsHZ3#G($jE+
zpUkF(Z>;trZGD`~YdH{)1JmXF0x*uU^8S6m+FI$49`<ZXzL(YB@#p!i(aSSD$3KUC
z^QDCUop7cPA?$og)6#X-cz)TY^tt5enaP*t|6j|>rn=HH*4Kglv?Aw&Rp3Q%9OGfW
z)~{s@UttWcF%Ku}P2k|oYO;U2A9*KG&tf*10`!gAr*^g#atok8{)D(Tr|)$yy@9xO
ztpgu`^<Xu4t${Vjxj;^vK?|^cc{FU=c;MN$F{bdI>l*s|26I$jyDj~HmN;C=S^i6m
zNjt>f`W;rdufR7s#(ZC%X{*QcyaO0a*nO9KYJ79PBVqSn?4$G8aY;LW8$3ciM}eye
z{|<R9NOi275^o+iZB*^kY1uOWH{md68vM@0X8WR(wESiCWgYGJqi@W;((A-~!KLfZ
z7{>54Y1b3yjPE`p=wD8r2dzL}Zvj&qQ2*MQ{`6tcfpN&2TvdTQ3b}nZBjwmU+VetS
z-R8d?r0Zyr@abR}7y+6<Iu2y%75d3qS*p?JzY~VrJbR{!=htbZy#-!_N9-@}d7d+F
zbRKE%0Lsc){u@ABWgb*U6L#(8w<@l=?z2n(3Fd<7l>Hun(*V~`^cMX7{CHEpQoo$(
zXXx{pwPb8<)7!>(qRzHySs##nwg`Vv+2j*kYoe3EzMuzv=1k|l!nvrPDn}D${1yC=
zW6ez^uKu;JDwqqFf*D{MNO{n%gmx$Y@+I%)Zw1Ri1z68|V4dP+@FFnJ3a|i72j>I%
zbY30<`hz|o9owPweI=MpU)dXl{jGBne)U78Kk)22#^!HHIs^H_R{efGthub2rRTv@
zz%`fiZQ$DdwESYz>LA~N{4eo*i=?b~9pRhkUC>`gus%i?Ann(`l4sYFyGQr7o8@a8
z|E#sRVXx_a**XqKC-*0M@w=qxXt061s(DC!RM>00irkiQq7V1_K`;2S@-+PQ4U9+I
zeFn4w`I-&nQU2t&8Khj_Nm%~Z0J&QW9tU!0o`=D6jksyzQ6Q)KgA&kY<ySt1&xXaD
zbmG1Q-TFI{aL4c~`izZ+y9V}1+F>4_8u-71IKMdwcNh5D;9LKTy#Gdd{k3;iza@`%
z8=QeEL%_b2=WOJhD6JzOKHoVIY~1<LLi>fJ57ExZ=Q4fUjJGyCIj;7e_s^8Sds8^{
zjWuLb8TdUU8>jwLHr45RD!=*}eQ$Ic@g0dT19uT#Pu=v}aJZ8@i7}2PpB%`U+_VBY
zm;mJ3bLKi^XJiksf$%`&R$xBiUP$jp%y&2OX~1&UXT4Q06C4A2g5kh9xfrYkZS`?P
zA=#hKVvgQSKes{Of1uCodBWY3ym`jq9zFbqkF!hg+1e!cQI)SDbCkh;!yWTSlr0n1
zR(tTA{c*VG7*5!GO4ee1d04{-kk*0tB>FHCERFF~U=GOk^MAY7jqt1f%a+ZCN6|($
zUaD*!_x-S3q<Qek;kR;}eZo86-I0Hy@8}}jL%xB|<`ZY0hxvOBHas1E<wCAn!5pCd
zDn9@_*OIe}Jj&}5(mnx}*^;=t?gCa5UkBt@o4*B=!R<g>9R{@1o^hhHw9m@*q-_98
z!BK@|@7t9yb5p+>ej}!Bx)vM%HL@RE$m7$EF`i5OOvZHrY3w(`y80H9I}3lqQu+tu
zc71Z6%o?SS_q-hLXwCBhSV-FkfhujcfC&wZK-$metDAJavHt@%riI^noPaJCk?$4Y
zJl>d(^8U-l|EF*_%FoQyF_1pq4JOmaUf@OeT*J6}!lCQQ0f7Cc>yq=deqB>kyp?Au
zw|7&Q+$y!@$AKK~0p$5;us@LJLEs3`8_4ZzK(60KIzIWG3|z;DgL$CMcuK!4B>S1?
zV(O(|Q;4r3d<gP%<Q~aftgr_R=ayykTYfyha*rHr4x0`9Z;_;Xmt5(=v&I?M_uS#P
z4ch3jJo9d&56>BY>0R3N{_zglbI<b{ZO#D`Kr?V3^{KTKp4yC~-MD&jUN_dxkjFh#
zb{{p~b(u8bzqa*8luy(9(4X1BwM%<=8E)B|gflPt3H-We`H1J&knUPE3Jl?y{<IY+
zg|_P&x*A0Evm5v#I10$G_9{98ZTu7P26!2~2$ll*9Rajo*I3th`#+XGJOzFR9xWvM
z1^0v4YFJ~r6ZHN!*ii@g<Qs$vKGfg*m3~7W_d<IUX5Z-gr43iuD}{Zu@jc;A|2+iA
zv1>5=xwj1KEB5VPDzvc`?Q9R$#mO2w4Jo%N+Unc*UgX~_y|Ve<q46yIuDGM)s3p3n
z;A<=7&GNgk9?}EPZRSp$GLJU&X5Inx?Ok}ifqLcg7mRBV<8dvvt>cI@-=iye_GiGH
zuHd`FUf6ji7c;<YFb(Ku<?1D%eTvn<b$ENw0cfwL$*Xf>1+a|$S_ImzRmajUzS8f$
z^6Z66PZEBQzV!y4N0@t|Z~7qTYD*k^cIj5qwfBqZGke_NcYjWJFynLYv^V^D2F`;X
zpy7XAk4iIWi@lWZ1oO<3=m_euJ^jRa;4q+mUZkxV;IsHO&sIQPJ13O>+sIQJel|<5
zgshu=6A<jS!F$U6XVqt)CW2M)%Xs86=+JY3dw$mD=nbCzH}ZnUcrW7_>sYYeyQ`ac
zokF+;OlV+MCNB4Dz%$@=PzSGo<v?zq2TL0GB671aGXB;$>D#sRkNFqQo(~}J0k;6c
z(eX&1-A7oz1Rw4R!`ia~`53&q4*v(^fZzOSj6<L6Sr>Z``^hDgWv=M=!#D6JA_s$`
zDgS-&7w|Oj4E7-DBazOjoj}U3vTgtO;`!2S`oK@;X<Np45gya@h_J|I(AIvt4NCI6
zK70u5%Um!8ECJ3%^`f3UEA5f=OWSrIk-wO=_a^jmA^gf&)^0}=PmywL+&z-<7H|uY
zQ+Zwm<~8s>axQohe1SI8F)6(}y*p0UIQ3niFIBi=Ee-ot^joTfpYiM>B>p(uEpAO7
z?_%K8_kCgQ<$M(GD+<Jq0s3U^cPO$49Lt$@DW?Yk+dGvyZ=|jv<fFdujUjg=Va>P`
zow$d4gm|h;Y$e-&%WlpsW6F-NZTNX~a7Zojb@vnQ3+l`8YoH8Dpa}9n`%TMtrCoI;
z$NE1x9zohHpnc2+bAY;>23o*n^i4cR_-e+8PU@w`{%U45;aHyKPcBbyNW&vC7(loQ
zD7Di?U^<u#^y%KSPeeX!xSjFi|HAj!W9dVLKBuDd-vai0!A930*&_tMy$E>|eP#@v
z`CMbe`pSC%@8YyU&&ZVn0qc(E9@o~u?KzYiNtqU~0Ia7hdA+9$?}gTqW?QULwjDTj
zP1nE4>p<kY$k%}5*onBhYi~^Lr7fs#`X+48TH@!_a~^a9^<|ql&Jj-el|wNFxc^Ea
z$300K4Vu9O&;%Bu8|>FP@4Ba-F8vqr5uA-?fSK5@yi5Ttpc%-mQl8{iE|tT8+=~97
z3g!dT7lXFrT-Z*UbMr;;IQ_bg{#;3a=F%T{kt^T%R5}yJKD}=aYvl#V9l#38RLJW&
z$Gm@|EPLwG%U}-Aj{-fwp~SJraPPtWR)za^?NuM{``vKQ<$2h>+kQyqdhj2dJ;Hkk
z^{ejMj^}?9-=~_~-|tdm9Rf#!FM&5nzaN|r%oBkMycRRuTK|j~+W=F*N;n<ffcAVl
z=tTN}26`fycis_%?@+ksasTJOHN4;EZlic3X)~$`<TJ}}mALD&e9EyH5AFodfjU?T
zMuP?Mr}eR%^v|+m$0E;bfa{6#P8)uf{%W7>`TaXa`W(~bBD~Kr-`n61q_MBi_QT!3
zygD}Q-E!fZ9KR903b;OF<KAWHg9F#_9R4QKImm0kc<Q>1I@Is!Am#FuOn6u(oIcM=
zd6UC8!FILed*Lr5JA;=<Q*Y;k!-4$jr?uU*j_z^Nb`w%=h4w4QO8GVI!wm4#;T+Go
zG3ZOb!&)5fu(+>rPZDfbUR!{2I4FW!z*1nprZ+Gb`6S4G8#*~o>^tpeR`y|R?b)$;
zI?|8(;R0QIXAx~p-2F6Xpz!_M9|=DN{sQ1E*!zy;yNEDrVfdY>d!h@--+?+_XKoB*
zOxpK(!1Hb%><zG2-zNlHj*)}#T~oM!FA*kRz888piMrM^7Im18q1Xs_B|S}RPj*8m
z>Pc=t0o&A)-vLBOxt#|l19j^>@NQ}*h(KE3V%m~JVc32co(fWav(J1!44=ll^t=%E
z(c{pII&r^oDPiV&XjlI{y3u|fnFAw087u&czzSgh-v?=*l(zYNym?yoc{kd~#?!F-
zgLlDc^sf{A+(Cc5qw-!!JJW`LPa6KlcMR^IgAI>@Pi>q%V%Rgw=c&MZX(e+xcLQPe
zT9t9gW^e#F0}%JUN(H|i;;!9$Qa5W@SZjCt7vo6#ot3v#6Ta?8s+*2Ad?=7t<BNfL
z#)D&kdhG(*^UZzHqqZua>Qw$xyh2!{apgSFDNfD^-$PIAC)~5DJHJ5*`%v~I!G24$
zc?z%(UjzfdFks*8V+GjftlehEY1sUS$L(|ZS$=vK`m`%xpBwh|oMp8o_pzamhY<dE
z!mGjU<b&JN4;X`76?pb4<M5ply!y?4{%q{>`@~sOi_?hfo4*5&2PY7Zk>5me4i0*F
zlloZGOS7oEDbx32Cfs(MX<VHQttEQ-n6N&%OC$ag@iIvDslC=4kXvmx57N4n<$UsO
z)!IP`mp$Vs?5UJ}GU!BDuAKu@!EVI2rT>S4eHqKo0s8aoo4Xl)OVwn4$xjoI+rA(V
zeD*lF6YNIcGyMsB8~y2n)SF3fH_U#j-So8F#nf>*b<G5<x9*kmzW?|JaBqyC4ex00
zAw5SNj>5N*+&_BO^Zjpyyx|?U{`NTVZNT~8aOD!Db-Aa)Ke%@b-zxo(c8>z1GyThF
z!fF4qQeC_c&aNfjhvkra0N=A{yK<{t<UqDu;B#5RZ4nuP>~q6&<8V(A&OB2HyB`<^
z2DCwc;+=r}>bn<%?Fz|zkrNo(wde_-RsI3|o?lIDPk!@Y9}t0V4OwZYAJg#rk5SLl
z)OkFTyNOTr^#Ib*<EPTwjwy}1_F<ou4sfsohyix0zs(;FCqJRSX9`idg*5rX&xY^k
zPC>$5aWYa~w<9m-9N+up<^St|y8N!rxPBsh<9suHW4=e~D(%DnB-PL7YxufaR4O4)
z2M2?LKzFc~=eL6#&_=CKpR4o?B*#K-jdM!e*mkEl**7ml&IWVAEHD8~0q8u;L;E`d
zGy`qj`n!R8KKwQ<>`P`aMtqWIrEovorMhWtIGk`-a17WD=v%Vm{*bzW>$u<ehj0ED
zA+vc5A5GdJne=N2kIcl$<Ndzh)ud&^9Txiq@5l4*Z6nYZz)|oQ?3eX9+G*Z<N$fhT
zrC(vL-bNZchi@WXgO<_1<=`LG_f^t&CC#(ct@OdZvA+vCYTNJ6ifgyp>ckp#ASZ#7
zz#y;>&~~3Bk8_~iI^Z&gF5BbQ{QB8GKpy2-pZ73mZOEm_xnMMq^A^w@XuEZ|ErUE5
zhV8kgaW@;@U-zjdYi?hp;b(yU$NsvW^aQp)F;429*Qigw;Qpu|b!iJbgLcP7S<>~r
z6>KWa6W+fbobXioX`}4L)a6?H9QZZ79Kk)zwQ$3JDx8P@lRCad_*~N1Gln}#>I}Af
zA?f{q?{N=DGH=4)DPRrpT|;<ZRHmJiz+PZJeeVkFb1DsgT;15mw6EbkoqD*gh95!B
z24jIfx(Rp=dYf|EZkKk-mnu=PRc+RAR!`cZoXM3O!%yJO{=EiPXQVz(?wtdcv)}8$
z2cR!}<nsyMfa8zR;X8$7t?ge;?i;TlpFTa?ej2vEK0traA8CK;)OR)Yyh%Mn0eL;g
zhIj0LCH;dW&2;(kJx=%zVSn=IS99=v=%zfr^9p}s5Psb==l7(HcM|)7nE1BHlL70h
z_ZQ)v5@XPRhxdZ!<sKuf&*$;{@6?B14}Z7gG$eD;cU9qia|``<Jo-5Gl#a8^n9{Iv
zcr}^pm(;imt3}ZR;1+Nt_%=8ad>%YczVktO*{1Ss>k9rj!>`ZfMV)Im`bPUA&r^W*
zy9&tj8lXLC$8&)Cx6A}2>sojxCO3KN6wHJ0`?6v+;r2`PF*~ko_!Y`EQO`=`czp9i
z)b}#+Zp7g=za_fVAKpQDAo?AQp1p%=C10bSJbDlJ5vysJeBn0&>;uA`V79G6q|Jq2
zx#Hffat&a;^*%LlcQw+r^lX4#hk5!5Ww7_~zTjTu0`ODXJ{zzuguli71d?^h^K;->
zy{Jpark?DB;cQ<FyH`5A#=afgRg0oKkT)PFfDzz8umxB|zI{Npd>%cQ8_1^1qj7W@
z_*L466rJKk-`c76S_SG`TU`QfCvTm4#ahCbF}8;jgm}L3Yyf#rsWKPnpZ+D=hitf+
zeAd$+iB7^E<xhpg9++3*4wN%Tc%Si0xbf~}?@hiqeB-BHS@*)<aK)#3$6|VTH`|#w
z_Zp%7-t==lc$?>*bKy3eXGaodovpKvDssN9OhitlopFTwBYOh;L**>;OhHDp54ZVF
zg!cvdX#0E^aP4(0>UA;5jx`(ZhE6^N=hir%)ws_?%JcULW6#krpx;(^uFqXSw!Gnd
zgWqyPW}l<W;1BJG{W%?&CeJIu2*NYKY|skiwi$HeS)Tet8EE79r||x4QDIZM`36)!
z-KT6H40{)$JczRLd=5O}Z}T0%7U=O4{!X}e20hSY9w<4JhIerA7WQ)Z-mss=m#TB0
zg*{Ri@?!7qwNl;q4!CkI`3lJ$Q}|Ajy_5Uk$^_CTv7fq=IBRj`3go58F9G>=@3lAe
z-A}vBwXo-+p77g*f6xa0A&f_zI;K4OX**uy!_mv@V0bP3rZH+o-e1FZz%+0nh=J##
zm&i96)R(nW*Fe%hzQM(?IPtsMrd(>P+T$xg?pncaK#uz$2LSVZNL(I<BMr;D^C55l
zslSYr|JQ(F$7i|r#`YO$u4lpj?}Glli}22Xy9C!z@Bi|bq2JMfd!g|Cak{UWg09~}
z*BwcJkTCu;+#7vJn&(^X`2p%q+r64R><^re;X9l&sPnso_XpT(aXN9=fT6(smNv}(
zIK1oMjk?~a{0rbO)UhMat%v^k?l<^m^_RBwS@Owh@V)+?F`82gf8U@qm$3X!0cU~!
zpfgxR-eW*|CdxOsEH|X_6wPtMgSOcMW&v$gJ{JS~CbYe&Kv}1}Qs}4kLE8iO5(9_}
z^X4<{^d&CepEbUA()~SYSo5w&ci3ZC57D3BDEfU(Weaq?H~PO8`89OPy?w=fl(v36
z`PR`N?$Yu_;_QF(2a|RmWwBT5+t?TS?tN9BJ#kqdd^y0!`u{`-caRrT?*M?$&|d9_
zd%3WGb6)I4{B`Q6f>!}+lXrLF%%Seuj_r1I(Kc<1TJX!^?*!ahWBw6u1vi7Mz}JE2
zC2f2%$ettf4Lz5UDZ<?!b|y!o<78cRuQUOqHf&yb84mJ<<=r(G{}|?ib6_H2`Mnri
z0EoNS@V6I%b2;q;@|3qt>ux)?YCCC-Jlcl8Aqd~hOhu=zk=!fzt~H!Xu<Nk?4Tp=X
zdFDQ8JD$&@57Q{SCvi9rdnEj+-#UhS<uj4CaWc}okw?huJm5acy_&zh$jw9!0@@<}
zJlOaF#3SO)5&Us@Z@3ouWAHdwOnoDPYm{?educn~G=66lebkcQsn4i!ekME*DYwXt
zwmX{eSAe#AKV`as;<C-_mB%_d)~@75U#Z@mUvfAMIR*3t&aXUqMZF=%ljbvgWw@K_
zf$RsE58-bjw=%Xl#N8*~hOQ1m2l_?VMRn@k%>c@$W6Mh0ZNw|Am1#JY-rf8A9pBS=
zJ}sAs<2!xd<2POP7IY5Jo_E7<D6pse$;9!+;axiCpYX0|8S-WD7=0NEY*P--CXI7V
zr6c8NKYt}YVIk6W7QXYB0Q|G>#=_j4OMdyB1@JM>?GSz*`2*tGuC}s%)40;HY?M>0
ziNC$Bmh6$`c4dt@0v;i(?ZSC99t;7VH(sI4DIlx6G~7a4?u(uSYk+I#9ALV93VD~u
ze1lWx#PCM?N~At~HK70D8wb~ygTMwjokSWo6TU?ljIK^X7cU^OnH-!{Vszb$=V^aF
zLux15oPYm5%^UVod!yeisOL1KduQ&3D_g?tEaK?jJ#TIb&#`0eD}Ol9NB@F)^cnqu
z^J_OCr)m3B2yaQ>vGw%tBh+uCuA3--Zc<0s3;&Mio_E+chHrn5Lmt}z<%3?<P}gXX
z+KBo~+ezgW)nt!6q83H>);ME<6@;HgK9L~U?k%KocNP5d-pB|nqU_-yw`}uzn`pBc
z$d6%dwG()S_PZk|0R5<(uO&^6^`&)qmS58>^9I<V5YD4vPx=mFZFXnmOUO}R4*=)k
z%#Sbgycyn;o=3ROvvi#QLb?YHzX^aN_szja7t!_6K;FJh8n#^7g6GVm%GZH=TlZO0
zXhXlE&O3s1y!e&ivrccww0*~pZwdcD1?zO*G>$f?H{6@p=PQ9W*#m3^JooE=#cN<1
zNb74)>Qm=|PPK%un~)FInEyb2E8t1eehTJ-+dvC&4(tQuwnF*oAP2I(S)0^0<;<}2
zS2#}jJgSh~oyom(;6w7rdtXq7U&}sB`mTj2d}qtN4Zl->gYceZHRXMFEAj%cg1qF(
z4?<7wcSey9U+Ou-zmpVvNIPqi`>0e$VJ|~mpKLkzqTvpy8<Mrrcl6;K<(0H|C~e~>
z!~39xw99;OuNC%U`fGfM>qOWK$U{%i2{`^L7{qhd%<#RH{`W_OYhXTTq1@m`*$>FO
zCIfwdb*1&S<u$J1FMymL489HIbzzNrFz_sR2K<V&B}sZzMryZm`%SPnNYA>ZDt;A2
zAj@ysu6CGj*rS}u<5VE0_EF#M{Cpg_EpjQc6L4NIN5dV%&kM=@&(TQsi}Dxl;#uFq
zy$ACpeBUhxL(sAN&=Y{?jBhsa|0mnTVX(2ZZw=|3@EwAC*l^}-p^oo>y(y1vdcPFz
zbzMU%;2HRI?)dCneD!tYy9!Xoy9~>Ob=!G>j|pRHVN7$tI{JDeb^IAD13yf{VK4O_
zdB*~M0QrKQ>fh9nb7VA7f2mH}j3dm`(`!-mU2q-XyrXRfK7Wl|1s)~sVXz4N7~BnR
z0+va+tyhyf$kH<IBS9T30>&#9Bz*2eU)pP5f2GZK2OGe7j1S$Wdkuec9?rD#i%!CC
zY_L_&8NSnW?7nyLe#bR0%ya7WuFrb`?*hU(gS~6;)fXU{4{2Y#`|)h5J*H`u5#;GY
zeLcXw;2h%LLtX=V1LjIN|KK~qyF7Rg?{MFu9>Tt>3wM{=v%Khc4k5fdaNnWdklQ2Z
zBYP>&I@)URzsDgz;n_eW`{`h-^BegsCnkbEU?n<Hr+M_-o^C&1OZG<dkw1gWe*@0}
zd~x(562BZh0`3KJugy*aR{(vocDuHk{B7S7JPxBxb`{CHK4b-cyYRd(c!Otc`>+}w
zrWcam<8<XYYo_0$_%6i#L-?NTEXL11ChSR=r(xgt0eLtR`rE4T+p^i@#ilDA$TyVw
z_G*;Pk;nTz?;b{>^RJWt?>xii)4sUAhVNk3XVU(S`X^Eb?!$SunX<go@$Wpyb@<MS
zeR24G&}7QFx03^Xj`LMcSxdv3(}DJ|OWzlVeVz6vx7*QfKYR(c?i|t&=Og5!<Z}#F
z@@Th8>(*BFch2)yKwr>SpT>*iQx2&`(G9i4MzvM_@GpTJ%PsN9wYC*Z2Q6SC7z2iZ
zo<N^G6KJpPwbfo4bJ)F*-|)Nd^IeGN4&PaZ-x?r2PX>Nh5k3PwUPyc_&wJAMwQ&3z
z>DW?v63<xMJTruQKGwjX#}Ak{A5!-l)UBP%i+e2U3-94(k}uOY_ey@-;@`P5UE2@u
zbMU9(u3#PY45Xcl$P-hKb+-cT6FUoMlpiC>7v4=>fGqQD9MUo9&#x!G6?mC3V4t3!
ziZ>@?3-9IOI?N%@W&4s(JDmplfF_XX?`Gkgy00d8g<WgO-}rG1+Uo;gIcf0d9iwNS
zNPn!&@+`U&Oa<qIuY+U2;oys4Ep?6u+4~FYeWH-`rJfJcbDlQjUehywxWjedu{UFW
zvk>NOIJa<q^bS(K^nvaZJxAux2UnBVc|907ZuI26dboq#g6GEq_DsI-40~tS-}lLb
z&kf&!;%|MQuD!d5_DthDw^UEzo$Hr)M*fh-Tncwk`1|m!;V9$`)_@zxa}ec!LEkv%
z+Q&RHzc<j&x~HQ4aKC|n3GX(Zq0idxsbCNo$(XCWub)mn{E5Fgao?@a$}<+f9|%5q
z0ORb|j*UjOgp0F~Uj|<V=K^f7^dMLOenK2OjaC3`HhPG(Dah{w@&##p0?-eC1$-Vj
zKR$%h1z;*@Z%%Dr_*8oSjAWnUocBB2^7V|R7l23i7UBO_%9`eP%kClN%Wp}-?=ntf
z{P@c7Ek66#;PdMF<nOUxK^>jJ5c0VH!G~6I<YA5Tj@~_$Yh^e;x~5{s;lBOP)Xnp3
z|I2fDb_pC#MMjj{9w}c(1LltR8M&6Gus72u_|7l<4*o)(<G({+X*+y3VOxg~W*rFo
zCB_}@z4QzJV0`4S-$5ScK+wU4B%j~PhCPULIIG)B(EB%m-yaM>ei!@zV1MBr;-|=Y
zK>p<YS^(F9<L60#g0Nf;14A;9*Y2P*c%Hha18p|j_P+|d4)I*S6W&kn!I<?|=Mp~<
zj3Lf?C_g#Yp`e3Xd43+@HHF~ADi@&x>^FQrN;urhIIrDT#iSoh+J}_E&T|XMyCb~e
z-?W49{n^E&6#<--ryzHyOn0PqqrEr!Z#mxqmEo}nyju+4%aF%)*mrZ~R??Vn{&qar
z-Cl(COODU-1;9Sedh;FV*JJdf1{Q*+=p*yS@8WWIk%uw*n;?CMcN&3L_aGAhYf#cJ
z_Av5tAs7w54Q>SYfJL?B9^x^=^4cGyTr7s8NjzT#o+bTepnZC0TnEmRZp60%w-c7z
zGDyqInSI6I=}$wtebcTpdD5_Zlwygz@`Jw)@4Y6FrjP0ewqUHBC*;okm3xlxJ5Kak
zzn<rpgC6A3PV|fJH&$VzgLpoPH0(v5%QsT)E$Vo-5cVjZ8+|wDw`u-np-#QQKQAE9
zKhgJP=z0ou52en-$UB($56Syu<o78D>NioQiL~2DD<ggW7-g7q;oKv~aGRTqJcPW?
zC&ugkJM=|vAEGVRV9V60!?~Ov30!M0qOa~f+w;ZE=>L3hE*Jx_%ix3OCo<e4wWIH#
zcWkrtG-aF%JwciEvdCH;!O?IK?)}+^wJGlz-Mehpp*$Qp#<qDj#`Oo3_a44Ho3WEW
z{0)$flw+;S;S=54I=8ippyNKIT}qmJqb&*Jcgy!6&!P<T!?iB>X!fCDpLGIh_}y@a
z(H(gtupMm&yO2LQ3+tP<w-dn62YcF$yhj20=e!B;Zg`fL3pn@pN_F~}gWF;Y5^l<K
z3Exkc`QqRB3-5Q7+U|*<FJabF*YfatJ$$rhU)Lo6|Jcq^*WGaMc3mO4@7QRs8GUy_
zcTGUQEQW*Q!HGbw&zGZ2{<dN5@oVJMuC>v&&)ddFqOZB|x&wV=>_IR6=_mH_nPu~F
zc&6>v@1~67Mt5a=oqGyxQ=Npr<s;A2ss9qv&IR~#eOY}G`R5a#n520p<vGf;kiOe}
zPm#H!4!t+@9@x9#@EbjRqrN!Ik-5|}1+<V4pX%>11F!l7=2LjzIt2L*>bRDClL`Cm
z9OTu&xkbEq3(ppiPJi_K<#~je2jRP!1E}|p=!w0t-^_$}ynB*oD*a@Rg|k*4B=bAm
zNqP1ek?5#h9m?wnpq;qx9tQRU`+$_6X3Be3YR~@V_%nQ}-;spBd)25X{Jn>R8OJX4
zcM^SMyq+mSzD6GVy|G=>yk|i=*X#F@|1j_^#tPrwDY^bsIs^AfpQEms=;W(}j|KZD
zdd@wG#9p+e@E+zhzM1Grcx%$oec)ajM)&^Rl%l%dg1CB==Q=iw-IUcYb0_S%wjnIe
zrS55zu|Fq)Q^|J|{kt7`FVB$fnaj8G>~O#w(q|V*&;2OLQ(i&%PU@BK;k5T2W1z2L
zUF9qh_72~r?D+uQg>R*1q1$u7!~cTYhE9V%^YGal<Un^&O3o&m={lP~4ZC0R-Cr7S
z$nc(x`5pH4jWqYA!G_es8H|Z>hWkF|?56Sqle?8R^{L-rzk5>V3c~u^Q2_s|?p$A8
zm%}|DHW+w%fOOhY&wd}_cPakOyl_wC8m=!qE9r+kmeo1+=1(Et+J;}YY-jR)lV?lG
zvjQpqUn8CQ<XP5nxGxTMA-z{D&LvIG=&OD|?Dv_M?wOqn`irvl{gix57|$=sdnI}4
zQ@Hm!j=Vp_KkUl$>qvuJ{dhPJau4I3X85*|bKEA~WqhsnnAKrFIPD8|16zQsO@0yn
z+$&}C6D}<y?R4-Y&s}#L&%@sg;QSEo8wM9Pru$t<k-mog&Pe37^nDUz>YU{_)5lDE
z-qUz@m_si){Tu!&?5TTE=N&vdFVTzpUDtHahQ;|n-)k7@zBK$6{b}@eHTm$P`pP0Y
zR)67aihjf1fO_hr*Xgg{s|8#C4tXC1tfe{n?%$)-|2PKsa$&CwFXdUJ!L4PT7uurt
zjec|NdvtwHojKy)(5gR9{^iI;<Yj!$!!S?LU-(VJp`<bQ+-rpL_{4C}aWc={Q)P72
zZn!i290XiP_0M~P&w;KDS%Npu)G2olZVHF{#$(8rv0vXSZuD(m*qk@dX~oHmhxzV%
zt1>zdd&`c*3*ef~^QUbp{r<=IGG%pkA?47q?|l5WD2G178R-Jztb_XE`qKa|>b8ZB
z<um;K+GDAQxsZE=^7vZq)jIXlo=3D*eSaO^-3Nu=MbfTor+colw)PNt9;CijbbSN3
z4KN47_Z3Hw58IY=ZM6OX;a0*wMB=C2-@6a<oqdtM)t{rz#en%5-V^Ojp2gJ19^UT{
zf=?g7v(J&P-Ew9N-z2Pu+s4|J)l;|&Oz-c<V|V+*kNYauVSV(M!8YLF8uzrw%>3}4
zHI0{k%d@*dPw*)9m~SWApU*fCq(7&EFK*(?{eSpd59r@>fWO~y%weC@i@e|Cx!8(v
zICnGcreV*&-ZN?!^6TDNpBwhY-FVg=z*$)T=J6c6)y^vWlYTk%&Zd9psD2~md0xh5
ziml|a4tQ~|W85{iXnDE39?AUiyjh%ytR`{iSnwhEUwwmhdI#hC2g?LsZynAr+V<YE
zjGqY{)!>c!mir0#75S0gRfK%ZCGSDQo>4BZq|f*p|0Z#`quv#%-FkoF9dcGrN7hJ#
zi~GT1@Fw^QY5LgFKpWN%&m`S_S)*)tk1?mPF<yF#JU<8Lf+Bg-Hde>Uw;@f8^EAft
z>ZUS&uNLle&gR)f#>2SPdHrU_G8!FWGhu#f2kHqOyPw!tH*NTvAN_DxyB?z~Hs)M$
zFYY{X4U~8HzhTdOZ^DWDIBiV*cyADVym_7^j4kRr^tJU?@;^;o@SS@cDW7nst+|G4
z=S7~CZJYEWbJTU-{j_tyXU@|cT-gWDKz<uj{1J7~hcMT+qc6|%-tZ>k@UQ*qbHZ8i
zVB~J#V)6@g<(*sjwqiS^+-?P0;ikRw&@FJW9J~#7sqr2WXuA_Y3m5}-0}Z~ydzLqu
zJ2MKK(!A^V378Cyp`DalCGRv!H^I+D`Z#WLZZ_uAuUteP#^JloFxMHYoYvt;e^^9c
zVNU}0*|V{>t-i|-BY(~_=8%3>+fl#TPH_rh*0%C8(y{OGUf}cer59~>CEk(e@MIm@
zuYW7ly=?H+%$NF8fOWh6BG1r$k#@>=QSL(0<|2ns{#x>%3z!#j>7BGvzU%s7d0WBr
z(@Fa&ar*3<5Wb<nPTVKCw+ij7M3Ub%A>3&>9>PWXs7@-3IsDDt4sgJEJ-jc-a$9fg
zgMH_AMKyUxuqE6a2;_DWxDnh|3;$1*aDOOQXTs|<fc;+h8y%#DJIL>lw-4w72GCxX
z-*@46MIrFx-Th|IviS@@PF~h>^_}09ar8otM8@DM#y1r?7GQ51>p33X;DhVvH>}${
zS9irp#LJXtO)$N<9Jve~;`j6ko{7TxJ)H0kJf}=<HS!tqF)!-aqjwmdt>x48wfrn;
z*l_M%!V7tJIL~e+Je}~t)VGknsDHTCHm$>c>u1Z%OZR4#uTq|Ql>0GhNY8fOtvT;J
z>(xom)p*97aQ=jK3x8z)!d;8~E~EdGz>DbGb8eQ~RHqSK%!AwKK}W({ff!r`?x}_U
z|5fP+gk66dWy0QJNnz99bjOeTz9@bEDd}3@AI?5x-1r;kMw-X_Q+;Y0Z!5!DV>Dwr
z23<8XPIysIo~7KAZI=Hu>hDAPL5bdi?V!stJb7=Z{)<n57kCa=<?HFouNVjW&K$NH
z>|_k-Yr&($7m!yjir_%v_atrCe@2*j<b8%V?Yi$-%rz$bJvj6n?u+w0CtrRP`Dv@J
zFSvkw*C3Ce{C%XuecAbES#;`oxl9?$*@q&0>qEjh?+WrUPr^B1Tk_wF9(MrVr=AYx
zwOMmhj<<jt?RIOx-YMKOUjb(ifhUNk`H<n8>-!7g-POi)@0It<<kMHM_HTTC0%67+
zzN?Mea6+Aak6CGE4Cv4Cc*ac2`kPDrTDg^H%>A&gp{rnfgNgrWQ{GVK0{G&58t#11
zrS_XYj`~MX$6ds40gDNvJJ&<sv1xy0+Re>H;?L`^HuQc!atZn2AkQ50UZip*dGMd|
zZr*YU+(~*z`tdU3p2|2E(1&N4Ypl)TjMPFN?B2T+*9PaT@23No&l6`(`Aud8&RqY^
z>)b5Dd*~1T%Dq_l|D#c7p8ke?&y(nN18vR#J;7r5)o(U|;Tf*0RlW%T1K>y>>zZ4^
zXMaMzlw0pYdQ;zP$YGiEW2%YmM{Vvr@Ua`qS`YVJ;dfHbxkjC?jcGXCCp~~}(Z73Z
z_dV&eaKG~%o-wz*w=H6iVNI1YbdZCakQW~jzHK-W{r&-c?Lht>;1u#Ocgib~iz$zu
z>y)YVMz24^f1tm*YaY+*%$MSC$p@!((#y!4eB(Dfi-+Hb$lG^GW8V2&0q-8e-li*K
zgjapE_UtqFeEORrWAkk6y9M{p;l1Kf)FE%o?QnL)u8a8n^3O<P4ivHB;!WhmKPZdT
z6Yes0BmeKv<JEwDYj`K5{nCE;4p^D;HoKbao83RV)-I@pcL*DwYx@&`<<l~$>_%D7
zQ`&Ev{kUt%#_xBWPprS;JLpW?-aBRMbbhJ_&mj6s@0Y^w5U0b}57FCK89QrGZW%hm
zcdB#ky-u3Hp9_18576;L<UN5j>>}JleVcM{QHR?)+-bk{dwJHKJm^2f-$#B3{++xa
zhmJi587{v~_z~pw<nM%TZ>B$Yf|WpDHiEHXyY6r63&{(Q;ho1M>XBRL0o;}ECeE1k
z2jNcZ%jBaE-n;vbE3E1Fk#7d+hf(G%Byg`-C(Scr_%04^L%Y||_FUS&8?=DYKs!$P
zX|A&80(-;j^)=p!fJX>>9!m4H<!vs{z6L6k8xPWa|CciCH*QGAn%3c-x%ex__fs&7
z@f^r_4kE3E@V4kdj*nxk=sEc7rATyDoX)cc3t=5{ucYr&kM3W}_&d)M;T^y-(mo;W
zkDx~S7fA0xeGd}<4fXt;=YK-}8i|d2Pf>rJ`0s(XjepOzBJl-&GZKE|FqHPLrC)uh
z<452Z<e@*F-}D=G=7sHhe;V#Yj_3Ju;@GL@So`UHqTi^8`N!M{^EFSof#f+PsmDFM
z{vdn@JBc=*V!XV=40p^6k@{$TbIQY%Ds~CJ09Wple^g7>T5Z^KZkn$l{o4qo)=#Im
z&G)~FyOtH_GB))EkG|iwocoy^oE1U1e>#(ASCDoS@>0ge9fWsra-aJ%`Rau4BmEzR
z#K+9Vb{V_xhV;3vr{Ua59Yy?LnKIqB8XVRCM4V^ddApY7sH2QsFWO>oREGcDn>^PJ
zEf3fF^`iFtQ&6TH-1+X%_lV&>4sN_V4D(_h`JHc^EzIXRSB`R&4QHt%2{T8$3(4zW
z@WcMbPkwXM?f#&Q?+Wj@dLRdYy+Kc~6?tdTZU^8VM^3fd*`O7qoV8Sw?_cEcT6l9G
zwTSQxa5z}R^E5BAbPH*jc=&F(M<IOQ6~6TzNI3iae+Y-WclB`{W5FNiu1)xLZ>?{t
zzXWb1eGzf|Z*DPp(QjV9cOp%le?%U1=e>KlyJ7DfzH59DU44t^=&ATi@;pZQ`Gi|Z
zTa%P4S15M|b?rm^2RvI(zW2cDq)e_(7@ph*g*z$2?ybW+AbrZ7=y`D=e6y9CN?!bL
z;1XH<H|l(vIDCZPt@NcW`mB!)zZW=`y6&aSo#bPFr{^O1)u(y>4)0mcK^{OI|D>Id
z0Y1t7b@--rHC(!GP6Mu+eZcs}eriRP`zzoXl{===e6WzP+#Uhu@$7t1W?fBrno8X8
z4uwRQ-1UUt<-Cb5JD^|r^^O4l5$<X`(l6H~%Q#=sdNR`UfP0*<jy_2Kcc>GeQ&wM9
za4+M375oCATlX@-R}DmmPl1)-X3}<~{q=EpFY0gPD(FeO4gMM5T7HW#`nH|=8<fG<
zYA603H`~v_Z|;S}uKR8|IH_B`j{w+9?t02ROkU=oc3CICdndo2_8mZR74eD4M|noL
z_#41{a;*v9RP0Ik0>U}Q_-DrW8_MA`Tyt{FLBFZ-n}o0zWKM*$&rwKqxi4T1ajzJD
z^K=F}-45uNwb^0d_RK!{Wq5i6d<mU@s}{a_E6pKn+F{6}K!$_xt#OHcv_8P}0}9Fe
zgi*A|88+OntKTC?-x_!}!GDBre!oO|rY`4lHm+~-T%nyRy#ow)#9wCY^8P#0D@fL(
zatqIfq6_DQwgg{gI4)nu*r(Cn<4J#fC*@sa{b$rsP5kZiTjP)6IjpB}<e5lY(r#T(
z>(<43kXuWc-y$jJy{OP$!}mASh|^xUFZ?Fs@{X+^ar{bsIr%Q54*Dy%^=FaROZ~z3
zTi5O6f$#FYq%ltW8Gdv59C8YJ{2b}>{3`K}sJlD)w*Z|<XD)e$3*Rj7i1hs8erf_(
znc=p7jXhp9IRjl)<F1hSb>L*MGw2Q?a5#A~yo7Jr(>V6w-+Bn&ne0RR-=OU$NIMZ+
zM*83AoAdN1gl{G-Er;|x=lxxp<~J;!4fIp>Cm6dNPek5LKIivb(#zy|gfx6?aT!3*
zav#=b^cD6P8`!t18_tjJLBqL`IZ=nR^2_A?2X$!c*jMoD&+!bqDKiIhe<NK^*CFNR
z71IBU_<s_{&)R+%(^aIeAf5FfwDl{_uJ`gBU*KHwyc6!kZY9rrFqN^K3I>z@0D$YF
z`R7soLV%C)eMtCT`VsoQ3-Lk3-z+451MYRCHmh&`9eNxG_63YJ?5Cz6`+#!eOuB%u
z+_LX2-3DgYn1jeEK%VuXoxw`-oDWKE&a(3KMx6Yn!?|}8ed<sDkAyF5R=?;ygy+`c
zWcqpvIB^q~a_V=Z;kVHZ&U4!B492Xzy#SU`ADjgquoc&2b>*DTy+%6v&Mib@A7%HP
zchW9ncFu=y%^pC~zWe1Oyo7gk&r|Ooz`K-PgOoRTuEV8wg08pqUjf5#74DNakoP|b
zGp{_ac+cd!N9UGjuF`e1a}IIx=a?JT6V7Su72NxW_l*Z4uO`kIi;S`CJf1@s|L$FL
zxSOPp?%lP6V5<iZ#&*NEkar;O0r<%9jSK5x_+61_%*TPY+t6c>>#D(SOLy0}<E$lr
z*ROy)66_5=<hk$RX9Ddv&)WNPys3^sh2-}ww3FX~e%ye}A)h9FF#RCz+9Kb++YI)>
z*s|~P!d-QyPvIQHe!;)d;(k~gaNdXe$0gM99ARyUxu5$x@|Wa6$Lh>?diqGu*x}xn
zITh~4@a;M3kuUFv%gX@u6yY@oZ?5ZMkNYRm-y}|+uqVEQxHd-JIc(oK5`2f{ucPi2
zwDlJ-2dv@j1()T2lII=D{1M3<D%T0$1Ak+Yzo3ujl7_wK?gh+M_wViz%kp{vWoXZD
zc^rdt#yhw0_boUBx_1fd&Sdm^DA*Nr0CJ*l_C7&f-jAG94ZgR8FZPTR-VOPl=oL`G
zPCLcPT~?j2wp#>Q9wt%mAg~J<2_FS?JBs%I8{k{Q@5Zs`@D0O2#&9oVxQ{WxS@?f=
zWa`S+8-9E9ImXzNe2gu854bC1???FusOL$roO<wib?p$n<PYKbR!I2u9?5rt@~fSP
zZ(^}u?-7iv%W%j1?@iG=oQZ~mGWEF!^4@;<mV=XXkaqZQw5k35i0Ar_R>JUFB&~8R
z<Gqr$9s(aThnq=f{%HHb-?N_RbM)PETmCt5+V-AUU*`M>-!KoS4E1}~3%~ihg1V0<
z+?lbs7uQ~Y4Q>bD2Q2`e!|$xOM!pD7^4kj1JBd39?^a8E@l0f@>*3YpF6$U%n$|{!
z?~oWz_zm4P)W4c_PCl4>`s8qr^H=(C7kHC#kL20W^!a?`#o)VOXY#hGH_Y+N8EXr9
z$0XzR9g63`@ILPr^n$L6FYye$l_!zTy6m^e;XTR|JVWp9QT!e)-0_})+?}-F&<|`P
zXP(z-<FCjMk%Y^C0dS}O%jiG&=#SZhvHx1si9HxL?t8!DTAshwsP`h;WIp-rOL%8<
z40U6h??mK(kGyY!-+-S0xU5$w`#S)SVZL`K9ox;_MSXeF&I0frzJ<RTStR~>>VFj-
zVyo^U!Z-NXbNHPJe$l@t9sWNB?wQ31km_{*8ukrN1_y#PoxA>x;cjrO&9ZaYm`1+v
zyVu^d-vmc{6q36k_o**|r_kf6Jll)D$uU00zpbL*%*|j-k0DuW!d(;NPV224%$UZb
zhe<r2Mn3YE7a;FNVh81U0Kes)C+H47!tXCfkbWO&%aQnM`S$<O67KEpr!LBPc5*M`
zdD3@l#nnh`v%ZGsgyq2Zy?!^~9?dh9JonYE;n}lMc>`(hfF<O0?wn4YuhY(j<ZmSn
zT?D&zUUcRx^(N`qvS%^xshm^AS13<kZ700LWX*64j<rI4?j6JX&dZR;@oXE~_ys!D
zUS|N$1Kh`k-+La1Wc>@jSC-p{!9*aB-fif6<681g=;LbgTgvdeS)Qj{N2Is&t-0%K
zIExj~<wELl9XK8=A-p|(xE}lgxF`5AdOemg+)kPAfioGedx@(U!#6kaYn$P_aM$WH
zczy|G%H+i!D`yi{N7~n~kc&w>kMQ*XzPn*F?rV!rlfE-L`FDIXoQ1gn$KH*22c6#<
zOib!5{|4PxDDON$pWXx4|BU=6lDf)Oo;T)z^}6PlUqCX~!r$2JOr6>wb1>}nt|L5)
z=g$y_W9>D^v-(?<!&l^%5~eNQ{|!H`W2f#IjsJQ?BhTtJvA6|x1hylk`JxNYyI{g)
zqt7|;!tq!B;OirgAAV|w=sVHXy+@yO>Dia`J~G#Pz`6Z<_sR7>_o7QCTr%pyGcUSi
z^f{&v8P$C0Iiz2D{;2Wi5Z|ZY!F~Je+b4HT?*IAUHbXj%DLmVCOyQ)R#uVzik13qH
z_n5+e95|+M)nQ|xX-r|tF=Gllo;ap3=hQKU+*xA^4~`jAIONhXh04`q3PUDSb~<Hm
zrR-glolDt;lzoD-70Rxp><g5Am9ncTyN<H&QT8Lsb{b2&V`+CR?T)40v9vpucE{50
zSlS&+yJKm0EbWe^-LbSgmUhR|E@h`v_EyT?McKKOT}as{C|jZIO3J=K*;gsMnzHLC
z`yORKqHL#T+HI!YX4-A0-DcWtrrl=RZKmC3+HI!YX4-A0-DcWtrrl=RrR;Rd-b&fK
zC_9(33n}{qWh<0jN!b@D`zmEuQ+6F?-=pkDl<hQ*cE{1~INBXYyW?ng9PM&E%8#Sn
zakM**cE{1~INBXYyW?ng9PLteI%RLA>|K<dOWB2#eS)$T%C4mB3zU78va2b(j<WAj
z_M>ry&8On;I=nvl?GAj#1G+jDM@{(uXC@xe@BiaO%xzQamXBxm&d29{H6I^7G9Pzr
z&d1NS<l|d@n2%3cl#h4)Wj=oW5Bd1^_w(^d+YXAm?l~x~96l(XaKfOt*SUk@7bXsh
zuld2Cc<=dx;%6#@;_<&56n}pGp!lIqaeVx)alEEq9A9%x9Pe;u9N%|Y93MJ8j+eE@
z@o_(k<KO-|j!#_^$G`tLj?dV!5Wlu>AwD@@h+p_-As#%g5HGp55brmq5dUydA@2Ok
zLVWQb3h^uN72>0}X^P9arZ_sRDL&))rg-t$O>wuYn&NY25}w->cX_HQKIJz}@qK@3
zivRl0rudLA4359kXK?(pcyRoWQwGO}UNkr!cirH4?j3{U*B=@jcU?I+KKhNp@!0nV
z$G2@WBz`J4B>wYPhQwPRKP2Ao>>=@xD~H6RZyFL`^P?g0Ek7R;|M-O=@z38H60iF2
zA@R#Q4vkmuJ2YN9aA^FukwfFZj~yDnJ85XV{<fj<UmqMAzgZg^|NhmX@w0y$8V^6N
z!xr5;ZgNp4d}H^2<{sb^dC*0#)%aIpo^YVMX#7R`s<x%tI<`%vhEUIDZmw<7wi9tv
a(H9=vd|D1aZp+R-?Z|)B8ol*XWB&({QJmWV

literal 0
HcmV?d00001

diff --git a/lib/matplotlib/tests/test_colorbar.py b/lib/matplotlib/tests/test_colorbar.py
index a2f5faa55a48..0244101ebcb0 100644
--- a/lib/matplotlib/tests/test_colorbar.py
+++ b/lib/matplotlib/tests/test_colorbar.py
@@ -4,7 +4,8 @@
 from matplotlib import rc_context
 from matplotlib.testing.decorators import image_comparison
 import matplotlib.pyplot as plt
-from matplotlib.colors import BoundaryNorm, LogNorm, PowerNorm, Normalize
+from matplotlib.colors import (BoundaryNorm, LogNorm, PowerNorm, Normalize,
+                               DivergingNorm)
 from matplotlib.cm import get_cmap
 from matplotlib.colorbar import ColorbarBase, _ColorbarLogLocator
 from matplotlib.ticker import LogLocator, LogFormatter, FixedLocator
@@ -539,3 +540,21 @@ def test_colorbar_inverted_ticks():
     cbar.ax.invert_yaxis()
     np.testing.assert_allclose(ticks, cbar.get_ticks())
     np.testing.assert_allclose(minorticks, cbar.get_ticks(minor=True))
+
+
+def test_extend_colorbar_customnorm():
+    # This was a funny error with DivergingNorm, maybe with other norms,
+    # when extend='both'
+    N = 100
+    X, Y = np.mgrid[-3:3:complex(0, N), -2:2:complex(0, N)]
+    Z1 = np.exp(-X**2 - Y**2)
+    Z2 = np.exp(-(X - 1)**2 - (Y - 1)**2)
+    Z = (Z1 - Z2) * 2
+
+    fig, ax = plt.subplots(2, 1)
+    pcm = ax[0].pcolormesh(X, Y, Z,
+                           norm=DivergingNorm(vcenter=0., vmin=-2, vmax=1),
+                           cmap='RdBu_r')
+    cb = fig.colorbar(pcm, ax=ax[0], extend='both')
+    np.testing.assert_allclose(cb.ax.get_position().extents,
+                               [0.78375, 0.536364, 0.796147, 0.9], rtol=1e-3)
diff --git a/lib/matplotlib/tests/test_colors.py b/lib/matplotlib/tests/test_colors.py
index 107162bb0a22..b08721407219 100644
--- a/lib/matplotlib/tests/test_colors.py
+++ b/lib/matplotlib/tests/test_colors.py
@@ -221,6 +221,96 @@ def test_Normalize():
     assert 0 < norm(1 + 50 * eps) < 1
 
 
+def test_DivergingNorm_autoscale():
+    norm = mcolors.DivergingNorm(vcenter=20)
+    norm.autoscale([10, 20, 30, 40])
+    assert norm.vmin == 10.
+    assert norm.vmax == 40.
+
+
+def test_DivergingNorm_autoscale_None_vmin():
+    norm = mcolors.DivergingNorm(2, vmin=0, vmax=None)
+    norm.autoscale_None([1, 2, 3, 4, 5])
+    assert norm(5) == 1
+    assert norm.vmax == 5
+
+
+def test_DivergingNorm_autoscale_None_vmax():
+    norm = mcolors.DivergingNorm(2, vmin=None, vmax=10)
+    norm.autoscale_None([1, 2, 3, 4, 5])
+    assert norm(1) == 0
+    assert norm.vmin == 1
+
+
+def test_DivergingNorm_scale():
+    norm = mcolors.DivergingNorm(2)
+    assert norm.scaled() is False
+    norm([1, 2, 3, 4])
+    assert norm.scaled() is True
+
+
+def test_DivergingNorm_scaleout_center():
+    # test the vmin never goes above vcenter
+    norm = mcolors.DivergingNorm(vcenter=0)
+    x = norm([1, 2, 3, 5])
+    assert norm.vmin == 0
+    assert norm.vmax == 5
+
+
+def test_DivergingNorm_scaleout_center_max():
+    # test the vmax never goes below vcenter
+    norm = mcolors.DivergingNorm(vcenter=0)
+    x = norm([-1, -2, -3, -5])
+    assert norm.vmax == 0
+    assert norm.vmin == -5
+
+
+def test_DivergingNorm_Even():
+    norm = mcolors.DivergingNorm(vmin=-1, vcenter=0, vmax=4)
+    vals = np.array([-1.0, -0.5, 0.0, 1.0, 2.0, 3.0, 4.0])
+    expected = np.array([0.0, 0.25, 0.5, 0.625, 0.75, 0.875, 1.0])
+    assert_array_equal(norm(vals), expected)
+
+
+def test_DivergingNorm_Odd():
+    norm = mcolors.DivergingNorm(vmin=-2, vcenter=0, vmax=5)
+    vals = np.array([-2.0, -1.0, 0.0, 1.0, 2.0, 3.0, 4.0, 5.0])
+    expected = np.array([0.0, 0.25, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0])
+    assert_array_equal(norm(vals), expected)
+
+
+def test_DivergingNorm_VminEqualsVcenter():
+    with pytest.raises(ValueError):
+        norm = mcolors.DivergingNorm(vmin=-2, vcenter=-2, vmax=2)
+
+
+def test_DivergingNorm_VmaxEqualsVcenter():
+    with pytest.raises(ValueError):
+        norm = mcolors.DivergingNorm(vmin=-2, vcenter=2, vmax=2)
+
+
+def test_DivergingNorm_VminGTVcenter():
+    with pytest.raises(ValueError):
+        norm = mcolors.DivergingNorm(vmin=10, vcenter=0, vmax=20)
+
+
+def test_DivergingNorm_DivergingNorm_VminGTVmax():
+    with pytest.raises(ValueError):
+        norm = mcolors.DivergingNorm(vmin=10, vcenter=0, vmax=5)
+
+
+def test_DivergingNorm_VcenterGTVmax():
+    vals = np.arange(50)
+    with pytest.raises(ValueError):
+        norm = mcolors.DivergingNorm(vmin=10, vcenter=25, vmax=20)
+
+
+def test_DivergingNorm_premature_scaling():
+    norm = mcolors.DivergingNorm(vcenter=2)
+    with pytest.raises(ValueError):
+        norm.inverse(np.array([0.1, 0.5, 0.9]))
+
+
 def test_SymLogNorm():
     """
     Test SymLogNorm behavior
diff --git a/lib/matplotlib/tests/test_contour.py b/lib/matplotlib/tests/test_contour.py
index 509ffb1e56de..c957d2f59485 100644
--- a/lib/matplotlib/tests/test_contour.py
+++ b/lib/matplotlib/tests/test_contour.py
@@ -404,7 +404,10 @@ def test_contourf_log_extension():
 
 
 @image_comparison(baseline_images=['contour_addlines'],
-                  extensions=['png'], remove_text=True, style='mpl20')
+                  extensions=['png'], remove_text=True, style='mpl20',
+                  tol=0.03)
+# tolerance is because image changed minutely when tick finding on
+# colorbars was cleaned up...
 def test_contour_addlines():
     fig, ax = plt.subplots()
     np.random.seed(19680812)
diff --git a/tutorials/colors/colormapnorms.py b/tutorials/colors/colormapnorms.py
index 0c17c2834b96..c5929ad1f025 100644
--- a/tutorials/colors/colormapnorms.py
+++ b/tutorials/colors/colormapnorms.py
@@ -46,6 +46,7 @@
 import numpy as np
 import matplotlib.pyplot as plt
 import matplotlib.colors as colors
+import matplotlib.cbook as cbook
 
 N = 100
 X, Y = np.mgrid[-3:3:complex(0, N), -2:2:complex(0, N)]
@@ -66,7 +67,7 @@
 
 pcm = ax[1].pcolor(X, Y, Z, cmap='PuBu_r')
 fig.colorbar(pcm, ax=ax[1], extend='max')
-fig.show()
+plt.show()
 
 ###############################################################################
 # Symmetric logarithmic
@@ -103,7 +104,7 @@
 
 pcm = ax[1].pcolormesh(X, Y, Z, cmap='RdBu_r', vmin=-np.max(Z))
 fig.colorbar(pcm, ax=ax[1], extend='both')
-fig.show()
+plt.show()
 
 ###############################################################################
 # Power-law
@@ -135,7 +136,7 @@
 
 pcm = ax[1].pcolormesh(X, Y, Z1, cmap='PuBu_r')
 fig.colorbar(pcm, ax=ax[1], extend='max')
-fig.show()
+plt.show()
 
 ###############################################################################
 # Discrete bounds
@@ -184,53 +185,72 @@
 
 pcm = ax[2].pcolormesh(X, Y, Z, cmap='RdBu_r', vmin=-np.max(Z))
 fig.colorbar(pcm, ax=ax[2], extend='both', orientation='vertical')
-fig.show()
+plt.show()
 
 
 ###############################################################################
-# Custom normalization: Two linear ranges
-# ---------------------------------------
+# DivergingNorm: Different mapping on either side of a center
+# -----------------------------------------------------------
 #
-# It is possible to define your own normalization.  In the following
-# example, we modify :func:`colors:SymLogNorm` to use different linear
-# maps for the negative data values and the positive.  (Note that this
-# example is simple, and does not validate inputs or account for complex
-# cases such as masked data)
-#
-# .. note::
-#    This may appear soon as :func:`colors.OffsetNorm`.
-#
-#    As above, non-symmetric mapping of data to color is non-standard
-#    practice for quantitative data, and should only be used advisedly.  A
-#    practical example is having an ocean/land colormap where the land and
-#    ocean data span different ranges.
+# Sometimes we want to have a different colormap on either side of a
+# conceptual center point, and we want those two colormaps to have
+# different linear scales.  An example is a topographic map where the land
+# and ocean have a center at zero, but land typically has a greater
+# elevation range than the water has depth range, and they are often
+# represented by a different colormap.
+
+filename = cbook.get_sample_data('topobathy.npz', asfileobj=False)
+with np.load(filename) as dem:
+    topo = dem['topo']
+    longitude = dem['longitude']
+    latitude = dem['latitude']
+
+fig, ax = plt.subplots()
+# make a colormap that has land and ocean clearly delineated and of the
+# same length (256 + 256)
+colors_undersea = plt.cm.terrain(np.linspace(0, 0.17, 256))
+colors_land = plt.cm.terrain(np.linspace(0.25, 1, 256))
+all_colors = np.vstack((colors_undersea, colors_land))
+terrain_map = colors.LinearSegmentedColormap.from_list('terrain_map',
+    all_colors)
+
+# make the norm:  Note the center is offset so that the land has more
+# dynamic range:
+divnorm = colors.DivergingNorm(vmin=-500., vcenter=0, vmax=4000)
+
+pcm = ax.pcolormesh(longitude, latitude, topo, rasterized=True, norm=divnorm,
+    cmap=terrain_map,)
+# Simple geographic plot, set aspect ratio beecause distance between lines of
+# longitude depends on latitude.
+ax.set_aspect(1 / np.cos(np.deg2rad(49)))
+fig.colorbar(pcm, shrink=0.6)
+plt.show()
 
-N = 100
-X, Y = np.mgrid[-3:3:complex(0, N), -2:2:complex(0, N)]
-Z1 = np.exp(-X**2 - Y**2)
-Z2 = np.exp(-(X - 1)**2 - (Y - 1)**2)
-Z = (Z1 - Z2) * 2
 
+###############################################################################
+# Custom normalization: Manually implement two linear ranges
+# ----------------------------------------------------------
+#
+# The `.DivergingNorm` described above makes a useful example for
+# defining your own norm.
 
 class MidpointNormalize(colors.Normalize):
-    def __init__(self, vmin=None, vmax=None, midpoint=None, clip=False):
-        self.midpoint = midpoint
+    def __init__(self, vmin=None, vmax=None, vcenter=None, clip=False):
+        self.vcenter = vcenter
         colors.Normalize.__init__(self, vmin, vmax, clip)
 
     def __call__(self, value, clip=None):
         # I'm ignoring masked values and all kinds of edge cases to make a
         # simple example...
-        x, y = [self.vmin, self.midpoint, self.vmax], [0, 0.5, 1]
+        x, y = [self.vmin, self.vcenter, self.vmax], [0, 0.5, 1]
         return np.ma.masked_array(np.interp(value, x, y))
 
 
-fig, ax = plt.subplots(2, 1)
-
-pcm = ax[0].pcolormesh(X, Y, Z,
-                       norm=MidpointNormalize(midpoint=0.),
-                       cmap='RdBu_r')
-fig.colorbar(pcm, ax=ax[0], extend='both')
+fig, ax = plt.subplots()
+midnorm = MidpointNormalize(vmin=-500., vcenter=0, vmax=4000)
 
-pcm = ax[1].pcolormesh(X, Y, Z, cmap='RdBu_r', vmin=-np.max(Z))
-fig.colorbar(pcm, ax=ax[1], extend='both')
-fig.show()
+pcm = ax.pcolormesh(longitude, latitude, topo, rasterized=True, norm=midnorm,
+    cmap=terrain_map)
+ax.set_aspect(1 / np.cos(np.deg2rad(49)))
+fig.colorbar(pcm, shrink=0.6, extend='both')
+plt.show()