From b2a85a76627a8eb33a5367a63f113030d9da0e4a Mon Sep 17 00:00:00 2001
From: mhjensen <morten.hjorth-jensen@fys.uio.no>
Date: Thu, 26 Sep 2019 06:39:20 +0200
Subject: [PATCH] typos and added bootstrap

---
 .../2019/Project1/html/._Project1-bs000.html  |   8 ++++----
 .../2019/Project1/html/Project1-bs.html       |   8 ++++----
 doc/Projects/2019/Project1/html/Project1.html |   8 ++++----
 .../Project1/ipynb/ipynb-Project1-src.tar.gz  | Bin 212 -> 211 bytes
 doc/Projects/2019/Project1/pdf/Project1.p.tex |   8 ++++----
 doc/Projects/2019/Project1/pdf/Project1.pdf   | Bin 276517 -> 276807 bytes
 doc/Projects/2019/Project1/pdf/Project1.tex   |   8 ++++----
 doc/Projects/2019/Project1/pdf/Project1.tex~  |   8 ++++----
 .../Projects/2019/Project1/Project1.do.txt    |   8 ++++----
 9 files changed, 28 insertions(+), 28 deletions(-)

diff --git a/doc/Projects/2019/Project1/html/._Project1-bs000.html b/doc/Projects/2019/Project1/html/._Project1-bs000.html
index 114f1b35f..0e1c861c4 100644
--- a/doc/Projects/2019/Project1/html/._Project1-bs000.html
+++ b/doc/Projects/2019/Project1/html/._Project1-bs000.html
@@ -179,7 +179,7 @@ function</a>.  This
 is a function which has been widely used when testing various
 interpolation and fitting algorithms. Furthermore, after having
 established the model and the method, we will employ resamling
-techniques such as cross-validation in order to perform a
+techniques such as cross-validation and/or bootstrap in order to perform a
 proper assessment of our models. We will also study in detail the
 so-called Bias-Variance trade off.
 
@@ -196,7 +196,7 @@ $$
 The function will be defined for \( x,y\in [0,1] \).  Our first step will
 be to perform an OLS regression analysis of this function, trying out
 a polynomial fit with an \( x \) and \( y \) dependence of the form \( [x, y,
-x^2, y^2, xy, \dots] \). We will also include cross-validation as
+x^2, y^2, xy, \dots] \). We will also include cross-validation (or bootstrap) as
 resampling technique.  As in homeworks 1 and 2, we can use a uniform
 distribution to set up the arrays of values for \( x \) and \( y \), or as in
 the example below just a set of fixed 
@@ -316,7 +316,7 @@ approximately \( 2/3 \) to \( 4/5 \) of the data as training data.
 Implement the \( k \)-fold cross-validation algorithm (write your own
 code) and evaluate again the MSE function resulting
 from the test data. You can compare your own code with that from
-<b>Scikit-Learn</b> if needed.
+<b>Scikit-Learn</b> if needed. You can alternatively write your own bootstrap code.
 
 <h3 id="___sec3" class="anchor">Part c): Bias-variance tradeoff </h3>
 
@@ -374,7 +374,7 @@ of data points, and possibly also your training and test data.
 
 <p>
 Try to make a figure similar to Fig. 2.11 of Hastie, Tibshirani, and
-Friedman, see the references below. You will most likely not get an
+Friedman, see the references below. You should include an analysis of the bias and variance for the test results. Figure 2.11 displays only the test and training MSEs while indicating regions of low/high bias and variance. You will most likely not get an
 equally smooth curve!
 
 <h3 id="___sec4" class="anchor">Part d): Ridge Regression on the Franke function  with resampling </h3>
diff --git a/doc/Projects/2019/Project1/html/Project1-bs.html b/doc/Projects/2019/Project1/html/Project1-bs.html
index 114f1b35f..0e1c861c4 100644
--- a/doc/Projects/2019/Project1/html/Project1-bs.html
+++ b/doc/Projects/2019/Project1/html/Project1-bs.html
@@ -179,7 +179,7 @@ function</a>.  This
 is a function which has been widely used when testing various
 interpolation and fitting algorithms. Furthermore, after having
 established the model and the method, we will employ resamling
-techniques such as cross-validation in order to perform a
+techniques such as cross-validation and/or bootstrap in order to perform a
 proper assessment of our models. We will also study in detail the
 so-called Bias-Variance trade off.
 
@@ -196,7 +196,7 @@ $$
 The function will be defined for \( x,y\in [0,1] \).  Our first step will
 be to perform an OLS regression analysis of this function, trying out
 a polynomial fit with an \( x \) and \( y \) dependence of the form \( [x, y,
-x^2, y^2, xy, \dots] \). We will also include cross-validation as
+x^2, y^2, xy, \dots] \). We will also include cross-validation (or bootstrap) as
 resampling technique.  As in homeworks 1 and 2, we can use a uniform
 distribution to set up the arrays of values for \( x \) and \( y \), or as in
 the example below just a set of fixed 
@@ -316,7 +316,7 @@ approximately \( 2/3 \) to \( 4/5 \) of the data as training data.
 Implement the \( k \)-fold cross-validation algorithm (write your own
 code) and evaluate again the MSE function resulting
 from the test data. You can compare your own code with that from
-<b>Scikit-Learn</b> if needed.
+<b>Scikit-Learn</b> if needed. You can alternatively write your own bootstrap code.
 
 <h3 id="___sec3" class="anchor">Part c): Bias-variance tradeoff </h3>
 
@@ -374,7 +374,7 @@ of data points, and possibly also your training and test data.
 
 <p>
 Try to make a figure similar to Fig. 2.11 of Hastie, Tibshirani, and
-Friedman, see the references below. You will most likely not get an
+Friedman, see the references below. You should include an analysis of the bias and variance for the test results. Figure 2.11 displays only the test and training MSEs while indicating regions of low/high bias and variance. You will most likely not get an
 equally smooth curve!
 
 <h3 id="___sec4" class="anchor">Part d): Ridge Regression on the Franke function  with resampling </h3>
diff --git a/doc/Projects/2019/Project1/html/Project1.html b/doc/Projects/2019/Project1/html/Project1.html
index f350ddddb..f84e992cd 100644
--- a/doc/Projects/2019/Project1/html/Project1.html
+++ b/doc/Projects/2019/Project1/html/Project1.html
@@ -134,7 +134,7 @@ function</a>.  This
 is a function which has been widely used when testing various
 interpolation and fitting algorithms. Furthermore, after having
 established the model and the method, we will employ resamling
-techniques such as cross-validation in order to perform a
+techniques such as cross-validation and/or bootstrap in order to perform a
 proper assessment of our models. We will also study in detail the
 so-called Bias-Variance trade off.
 
@@ -151,7 +151,7 @@ $$
 The function will be defined for \( x,y\in [0,1] \).  Our first step will
 be to perform an OLS regression analysis of this function, trying out
 a polynomial fit with an \( x \) and \( y \) dependence of the form \( [x, y,
-x^2, y^2, xy, \dots] \). We will also include cross-validation as
+x^2, y^2, xy, \dots] \). We will also include cross-validation (or bootstrap) as
 resampling technique.  As in homeworks 1 and 2, we can use a uniform
 distribution to set up the arrays of values for \( x \) and \( y \), or as in
 the example below just a set of fixed 
@@ -271,7 +271,7 @@ approximately \( 2/3 \) to \( 4/5 \) of the data as training data.
 Implement the \( k \)-fold cross-validation algorithm (write your own
 code) and evaluate again the MSE function resulting
 from the test data. You can compare your own code with that from
-<b>Scikit-Learn</b> if needed.
+<b>Scikit-Learn</b> if needed. You can alternatively write your own bootstrap code.
 
 <h3 id="___sec3">Part c): Bias-variance tradeoff </h3>
 
@@ -329,7 +329,7 @@ of data points, and possibly also your training and test data.
 
 <p>
 Try to make a figure similar to Fig. 2.11 of Hastie, Tibshirani, and
-Friedman, see the references below. You will most likely not get an
+Friedman, see the references below. You should include an analysis of the bias and variance for the test results. Figure 2.11 displays only the test and training MSEs while indicating regions of low/high bias and variance. You will most likely not get an
 equally smooth curve!
 
 <h3 id="___sec4">Part d): Ridge Regression on the Franke function  with resampling </h3>
diff --git a/doc/Projects/2019/Project1/ipynb/ipynb-Project1-src.tar.gz b/doc/Projects/2019/Project1/ipynb/ipynb-Project1-src.tar.gz
index 4cae777dd814318424f0bcdda00a756e0083a765..44c25d08c828fb46602b6d1996937ad95c0513cb 100644
GIT binary patch
literal 211
zcmb2|=3toY&=bqR{Pz6$EG9#N<Avw#j@lfYl(T3a_ty)~j$ZK!yEaN^3(t0Wv9I<o
z^S6+ErKR`JKRs@6*x<8KeMQYY^RKJ>^4v5{w}$mu&i3<Ld8@c~(N~`>Gb0wS$%<C?
zl1_f>apkSwuZ{XwpB~oyd+hk5$LovRH-7zfYTjq<qLcQIw(`5q6LWqWJY~kJ_+?-8
zy62eHg_a7JImdWSudy^!{e8anL|lx_k<!=|imuyb-xpP#ez(}<arJ&?24wL6nE3=L
Ko(2XD1_l7|PhvU%

literal 212
zcmb2|=3sbf(-X_U{Pz6$EG9#N<Avw#j@lfYG~@2qm<!JQteUlrH+t{>P~Yg;^6vI^
z`wR2Fs4e~RH>Qrqn8$ei=jAs~@6n1+Eiw06xozFkDVMidE}s?i%6RSaMZUq(sZp0R
z0&mSY*5h*b{gU|hm>jElm)6Ig(2JRC-M2Q({98=i#QA^J?^VrzT4^}Nj*WBUj?^CM
z$R#_<-n)j^*|#h$x%f0Vc+zj1A5V_7Ip$pd`6zkA?T>ax`%}tyO8x(8&43I(80=#_
MZD{kGL4$z-0QA{sS^xk5

diff --git a/doc/Projects/2019/Project1/pdf/Project1.p.tex b/doc/Projects/2019/Project1/pdf/Project1.p.tex
index ee1530565..6c4d050b5 100644
--- a/doc/Projects/2019/Project1/pdf/Project1.p.tex
+++ b/doc/Projects/2019/Project1/pdf/Project1.p.tex
@@ -175,7 +175,7 @@ function}.  This
 is a function which has been widely used when testing various
 interpolation and fitting algorithms. Furthermore, after having
 established the model and the method, we will employ resamling
-techniques such as cross-validation in order to perform a
+techniques such as cross-validation and/or bootstrap in order to perform a
 proper assessment of our models. We will also study in detail the
 so-called Bias-Variance trade off.
 
@@ -189,7 +189,7 @@ f(x,y) &= \frac{3}{4}\exp{\left(-\frac{(9x-2)^2}{4} - \frac{(9y-2)^2}{4}\right)}
 The function will be defined for $x,y\in [0,1]$.  Our first step will
 be to perform an OLS regression analysis of this function, trying out
 a polynomial fit with an $x$ and $y$ dependence of the form $[x, y,
-x^2, y^2, xy, \dots]$. We will also include cross-validation as
+x^2, y^2, xy, \dots]$. We will also include cross-validation (or bootstrap) as
 resampling technique.  As in homeworks 1 and 2, we can use a uniform
 distribution to set up the arrays of values for $x$ and $y$, or as in
 the example below just a set of fixed 
@@ -300,7 +300,7 @@ approximately $2/3$ to $4/5$ of the data as training data.
 Implement the $k$-fold cross-validation algorithm (write your own
 code) and evaluate again the MSE function resulting
 from the test data. You can compare your own code with that from
-\textbf{Scikit-Learn} if needed.
+\textbf{Scikit-Learn} if needed. You can alternatively write your own bootstrap code.
 
 
 
@@ -351,7 +351,7 @@ of your model complexity (the degree of the polynomial) and the number
 of data points, and possibly also your training and test data.
 
 Try to make a figure similar to Fig.~2.11 of Hastie, Tibshirani, and
-Friedman, see the references below. You will most likely not get an
+Friedman, see the references below. You should include an analysis of the bias and variance for the test results. Figure 2.11 displays only the test and training MSEs while indicating regions of low/high bias and variance. You will most likely not get an
 equally smooth curve!
 
 \paragraph{Part d): Ridge Regression on the Franke function  with resampling.}
diff --git a/doc/Projects/2019/Project1/pdf/Project1.pdf b/doc/Projects/2019/Project1/pdf/Project1.pdf
index 963505aa08a0b6bed125df5ad8fa3bf6f30bf5e0..0380e19af76b2943c7bd35872c66fdadb451610a 100644
GIT binary patch
delta 15049
zcmV;)IyS|n@DRuH5RfDSIXIKRJt%+W8q0Fq$nmbPK=~4|RvgV6fVXn+Zq{3sIO~n2
zB$aY}fCxy$E(yp0Xo>mybWaZ;00JrcaY?0E#d%LpKc}Y~Jh%;b@bejeru^aR?BccL
zfrywCMh90nP?bz1fi;F{83$Kc@IL&tt`<ur!r$^$3zf)(f{Lo7@~`P?0pow`qRgq|
zO`g_eQQn>}K3x4dkP(Ze43<!;#P6BqX}0O-zg@M}wS$q@)DxXkDY-~+lNlV~$8=C*
zldA82Ms!4wu*6D>*enf)x(b#$VM-VR{in1|Ve?ppugY|@Z;A$6I!w!qx*a;QLkBuK
zbg1?EUzf|pQiPXRe}O&C1+;%Pl%rF^Ib;w6Z0o3M6q8W|6c8*Ckv~R$m)7kzFBdWi
zk*->W)eV)uuCWu2wrb9)=kI0l3H$Q8DOxu;Uh%fsRBoyaf2r#Gg^a_Lrp*&?Twc5u
zc0jhsGA>Q9REousqT^oXcd)%up|l7`>7F$tLV%eu(h*%~%}fkbw-$e7$us<2ep}~F
zQ*<(uKGNQ`GOECwbbGfUVnchIw`-I>8D<Ru8=Vx`8$14sK)`9W2(*fsQUG)dK(E%o
zL22S}n-(S3qA)GC)P<D8vRyAki08n@9f~^8jzk@Mx{<b`^-a6W_I}83xvgpsCd=Ej
z*iik`0-@DKwQFcvJ)nPpI}@P>(6WXUWr9s)<3LNxK<EgaqFilunWK%2IGh~8Mk0J$
zLnmO7qFB~Y*;7BrZPU`+%a6Mhpi}?9x#4f$TtXpeba0e&EKK;W$Zm5Q9Gx55j#spe
z(!j;d{@~$-CM$x6f0H(ivm<7(BU>!BHpC|!8PbG$(%KIpwXuI}ceSsrsx6B6HRx$Z
zJ@*B0(}Nhu5w6NtICotZA9s1f+=&H9d3RzRvq;E5bIEih>BRqB#4^nBgaUTnY^a7<
z*GN{ac_$6n59gLqcu)N(E#2gns&`1Vu-fd)YFnh6W;`*aqM=#&s=$>i>TFS)?{W8P
znH69$kaFqf2@Zdr>MG&Ppp`)!Zg%CWb*$4o4p+b}pcR>fIr|yBck^B&=5<>BjzIH2
zG_<<kYiLpc1~Ada9GF=6iBu*Iso^@;x?dM7+;&aPb?O&g6Xtn0t;ljOS#vl>F*DZl
zY1cUZgJ|Bbak;W!Q1hnsMuej(>RuJGqW~Q#@eplZcYJ?ITW_0qXN@?q?{S}Wb6eFQ
zVp|ZPrLl(Sw5vf8YcvAq-EqZK6BL+2%RNuho7P=((gOB)icbUq+HmUhdQ&uOhY@r*
z9mmSns|B1q-(Xkdk2#v;)hBX#teiX|QE8N9*nt7naN@eBw$m1;EC)1M8O^Nm^me)g
zoI{i2)Cqqf+OUD<jz+KpRJWmWRaZ^31SVLZAg;+gt#K71A;OEQrslQ7l5nlsrmfRE
z+Qb=K^f~G*_j8>|hORp|JFjo5dg~$kfx5ymxHT7O7Z2g4C9_G#o2CKF^mr@FD50AI
z%bf!p6E705_8Sxg4VNOA3KQluN@T_bpyd_S&`y7Jb~*95B#+kx)ZI0d``OjfYpQGh
zP^8WB9|!?tI4xIhqLZ{KCW^BN?m*&Fon|>**%ipKVjccTGBJ8eKpES?m5J9o$hPt}
zAP#+Br^(&}AXRq!#-_W>m7~5JbqkBzHI}UtXy|QrTdFwE1!mtkvSLsE;B~P?^1~W3
zlQ@5=26wTcS!pwmJGU%W9`~DSgYpf$@J@poIt_xw3HV}$#oRk5H-6U8cQoM800cLV
z7v{`H=pN~A&cpX```U;+uE}1)FlF^%H$kHEM?wWW7Vy3;@Qx-ZXJLQ3I{SFmqZuVk
z1~OJ)+I6tnp1uFTgA9g{rRY`PyNTOCgMfeIqD^pl_BZnI$48-!BFzL)!GY866yT(*
zC_p0mQy@DWDq0TL!;BbZhjATs_(<uIPKI72_+c8o!Qo8KXUGGNvL51(<W6Sj9kbZj
zp5C({6COX3c@~(fb#tWS1x>;eBoVrf9VEqU_h^t$St2C|_tu`o6wUz75w1uTu?T;9
zp7LdaAxLckwE>$Bpa~ASF!ARSbw5i~ZOjQoRo{fDK!6y=PavvFm~po3AW;)|%St^<
z)R9a}csaxXlRMGwWLnMQX>>f6>F-E~r{tc|VWq5lg?<Aq?zj3nJ(j2Hu{v3g&50ss
zrg|(6>G5}9E2Rv@0jUD(8H?l@u~L5#U?Jo;(NzUvr9hC!v`^-gHHH`I$;z61F6H*K
zl+%yeJf+;2CR)Lp9c|7Ds%Dr_M@XXa#V;nBCiuLnMbL4Asx1{4Okvyz{4=HP6KxyP
zOU&JvG2@8>i)k>r^>ufq-%1<SVJH88FU{y{?7K)92O6HjmFF)_cvPW%AFzMrfA!8(
zUv+1KQUME3py>&BCO%)IIhsDbGVLKfP9|4obdp3rA-|ZrLjlLraDDy`^;->4HD7Rr
zf^B$;*N_5W8?6nNv-8HbJJQMU$#L@C`6cSyi=I9zdSK9n8i?Km8fYL!dIz&ku^?Zi
zSN3y|J7%lQ400m|YX%al;2nSX$Tu-_>L>ht`AnqRzq_RbUsn!8)k>&Av<zE5W&<^f
zjQB}e?t-rr_k=Z7J@?M->_cps&|(a4`dl4S8C)u%MDNbe(0G6izy}$a?EZeSG$B87
zCMEH5*W*>+KLZI2d1NMZ6CaKf8+=#iJ&YE)u(J-|y2u02uaD_sqiKI#s^c*6+lopb
zTcWtrfk34cl5!|6xa8uC0Qb!$R@5l8jswG_h|ldnZ@Z=*9o}Gw?J)!s+pw+oK0FEY
zcdg%sib=$oCi#doj=O02p>uPfp@{o)ALb-Vvy3Jy9%h&_gkr$f89jxP9Rm}gVN1qK
z_RoL;!B5l#OqTl?dzpXb<;o=>r6;xYkyNO3Je^)qkDJTHf+u6fo4-G%h6%WAW=g1D
z4VTJ>u3>w#J~UV$`}jtkdIC>Ok+uoA&X2PuYChaDVK(~W4If4)eo37#yVg<4sR-<m
zv|~biIL=B8_$PY8evHBcuev-9inmu;=%wH*)+lY3_wnCp=um&evAZ(do_(-7f*txu
zwNi<@w5pow*G~0EVQPx%;b7WOIm$t)<S}kueB5CS7p;&{;VVDA@Nm}EHg~bps{Y*%
zbsLk0OoP}4p1-IboO`;+bDs`#8Dprs)K5v3UE_PuO7H~iNr7g?#m=#v6@U&#>92Pl
zcTU+)9BlH|?{a_V2iDzgbxjH9hGsgA;$0ea<FW+UK7Lhc#hKxDjLySx3$MYcm3%^i
z<HkgT{Yi8@Ub24%j8$w(Xiw*=ei~cvWXl?O#dP-Hhp>s$+7nM1m!Eh@x0t@YgX+e2
z`;^$G%PiuW-*)bvQo0MX6HEGlx^@=+fx_)3^jQ}oKSqD>kCI<m42sL~sSl99M^$0S
zPEHmbr@u-tPN2t2_Rqk<+$>)IqeVXB`$&#k{7wR`yQdCb<_GRLVH7;;ed4Dn{^Jx9
zXa@E~P|nN@a%SDd3DuOw2A-N5H~gmZdA%IC*>kd6VW&5;HyJQc-5jG<oPZA@$LD?K
zhJ}X^tg(ONHP0voxP8V<$nwk$tKl6uS>cHY{%oC7rgO$*Y4jW5xu<<f@xDtkfd%PE
zjt11H_R^L*T)w~0F=CGsdQ8=={c13|zSe!qvC6Yu?fZ6(PqhwK_)i_M>NBcI|FMEQ
z#4Vab@21VAo`g(5nmzjCuR2`%t)Og{KTft=`5S*hVbV{cMW&XQ;HM4>rtrol13c0q
z^~+H3Je|OObRP@s2d|zX=P*7O4PIf=%&-?sfavG2$!ASg-~%(_!iDH02Udc2yTrHx
zJcZqo742#DXeLvjx=YSvA6gL}_-IQZm`wVohXCzg;&(<!Nu$FT{@X364!+$o-M3p%
z!Eb-J@H-Fsb_+WCZ?}+h%UFOMf<Hr_D-jU<o@@9m7cd*DFJ8WIA7Qoa-OG!M`}=#g
zD=Jo27nsT|^7<m*mhit#+cIr0ZuSkXg*{-kxhQ~j4@(EqkO><>%*$K0UMP+UW)W^T
zFHmQ|QF)hd@{~SifRpb^m+8jpk3V8*!dEbEQCs`FdHG9K{#b3ddA%b49;Rx>Nk}<b
z<;Vl}(bG+_%F8An@${#wv;P7g#Dku*D+~<<0y8(38U!hS8ryE;#`WD_p+_2sCY<|B
zf;=?9Zj&P0q_Kgv$R-6XQ5LZh>5$Uee#y7@oH@g*DA`FD1q$?u<nYYQnRCBPy0~AY
zi$7nb@%i=b)#q;oUwGz?FpJyGf=kJ)aSJ0klWK9hUi_5&a4kiW)rae)!k71{=3iVd
zP3n^G>vd6TX7w}u`7Y0z2z2x7F{^WY=Sfx4*kPN~%iC*jk~%9t;JR^ab1YYfq6K{{
z4qJMw^CsIbuYbP%3n0AYoOx{stA|}t-ZT93`CD#=M_{P{;B>(k+;V26T^R0}ba0$K
zO#J(`3)@PW!bX!%5*>dubyG!<j2_(y0!DH~)QFCBmMLq-=)ApllBgkElEaru1}hCu
zqqJtyPEG^dNaL1fW=o|3R`~Ipr!TLU+}PxK4%fkcRx#t{$$or>=qdBAB|3<Z;tP|)
zVPWSmD=_ptmK;u&T0NdK#nKNaI>sf=4x%7t&M`#D31;|Y^`C#PmpZvbIqp*?v>(s<
z@`ar|P|Vr!9RBlVdWlpp;l^_qN&hXXjIe4dTyyveW0;(5y}698umXf*Jmm#>64o%z
z?PX4#+s<>Q96#Cb1$h$Qg4RzMQ8Y*xO-2PN1d5}3H5yYbQuCh(QXT1Ev#pN1H7$}q
zJ?uc6In0}=Tx)+0wb!7zQ3bMHQ-RLwly;YzC)s)p(+jd-VW1=k_>r&awK-I{U9v)x
zvISFFtmsEs6-^!&tZ4dSiwDG!-Z0|Vu>qqu+8d|W`)o&;PtkUgX&y<oE}BDK+#S!z
zkyaoJs#U4K&al$6)OR|Eaxy0<jybT<5ok_b6fn`X0Q`Rp`$wy=6Rs%7EEfw6=3Sei
zS+9#=!F>v{2y#GX9KL!p?tHQ$KCB}op5)jd`vPfUBnNAi*94r=&^`m&J<(86;%Xm*
zpw$iW2KN}x08YW0_s3lpark(xHOS8KypVydHfhVPd{ylq0-{yZ3YOH#rmpt*rjnf9
zv(JcjV$Oft2xY*4O_nXu=N@5UcsyTro&to&!pfuhfnNcUa9nsEgX{u6&qx)bI1CSo
zJVJY0?eoU~T7PKhXIwEW*90QS-!t$1ltjuhqh-{?79Pojyx0UKj{9WvXbu3D)e&+>
zAxiiWLlk_k^Lq$1f(#IyVS1VEo|~YB@YU*Fw^@I+doHVe0j2;)CCbDFCyZf;Gh^(Z
zUI<;<6H5(F*!3sA<V9rw9Zh+G$=nF0rgM%DLu-T@BGm>XGW@9ObzXx-)u~C|7Ug<k
zjaF5O;Oo3x#Sv5&mmKmM6hI|JF<fUenfZ{_***uj=3+c!ne=vy@B5Ui%g0p4g4;6V
zX?K4ph0pVbQ@~S2PN5=#)g%Fsh;UKUE2(r{WN_H#GaAcb(4d}EFi8_gbl&B`Ex-`P
z$LlzwuB*5p+M;(iZ+fdZCPHeM1FV)ZWzK7|j_<(trarj=j=1E_jKLKUH5xgz+Yr3_
z<d>NZBy+ZN0T9(oGi$<m%A{66U+-dy@JoLeq4#6I&1|qtAwC-HaJsf;LWu#g9L<PJ
zz9KfEUx0v=*|<lb($$;Wt6#6WAUl=7YPC?{vLytKtNqnaKc|Z|{6w1~f%}gEbPs!l
zAY-?<x%wOBLSt&QD>=hJ!f3}<&mjb2-bxp}6-<OESwQSJgi`4MN-^L2*Y|1XJ=K2z
z1}wI)opX$fh#>)i_SWgeQULCr;Qld#5<tPcrQl5P2v*GOCz#{D-S4Y(j@O)V5H)yM
ztp_4BFFbn6T00_u1Kxv(D6hsS`H=u<KwtngAk{<&jp|~Av%)6-7)E?_FbNUw2%}Ct
zhr}f9Q6MvNoR^q!_+uB#0nnH8mWzM8NwA}gVi0`~T_2b)v@I7rg41Bm^my)%GpK`K
z0&$TcO&X$p7AujWT#OfPR#kAQfYt_F=`i@V$+w#eiLr3t>{G-O44kIberrvNVA>AN
z%&Sbwl*@XK%jZw$v2I|-#|yDQBssj8<N+*ru{z|D6@7dcK7)S)sLLEuVA6lAv6G!<
z)xt<VV?E3$K=-N(t=okA>>(AzhaL>zn1|W^VVBSPM5QrX<QF3m4=Jd4B`@@OUKltm
zX<Q4`OkC*-1gxqdXp7-kx$gaQeawgM894G;H@lpygE3za+tb5{6lmUC$8&1KUs4;(
ztOu9f1GR@)=NKuKZ5WWv#XWz^p#HGQQcxl1A}(Ui`b?59<45M_6UGArMg7l|_f_#g
zc<>ob5s=7gd8#TSKZ<S*aw86~%Cu$9Q*nipX|Vlp3}xh}ckQFSmH{W=v<+%`(Pz8M
zJ{v$F^c{o<|99D!T=rCfPXkH`Dy$%oR)E}vfD+u7muehRj>Jx;fWLpnwFFDGeL?Wd
zs{~Jf8o?8<5<LIk7d-tM!PCD(@ZP7wd?H&kFHssvW!?E85Ry@)5c}4k&jo=?0YYLu
zS|8Y0Mk#pQ=5=&$kFnGlswx?NeC$R}yi4ct_FWR%CaUpAH5;Kh!9;IOUv0xD)jZuN
zpq-l=ZnHG7i09+X_7Z=e<Rv~imZrU+QZNI@;QVsth=pbjNJ!&@Kl|gEW4}_$Qh7OX
z)UQe$_5VcTXePEmj}yn=6%bDBMQhI^&=)g>&m{-4i%=iYjGNQurAzK8f>6>YKjtCU
z4PhhxI^-(^-Im3#$570Hco)m}9l*Y#j;0F*p<Vt|9H@%wP4a(2A5uvmhJ*x50yWZD
zu>JwdPoZdu6-x^NU|SeY^(6F`AR<=95^Oy{Tw;H9L(iz;JhUn4tuq$9t%uV`1Je&@
zXBB1qRcA%nZPWjcbyon$n?wJzE7PCUFlCA{a6I-m5hXE1NhT|@N+NlTHT)hR*2XIQ
zdW=wmwg`Mt`AmPbwdl^;P2n~zU?4zS(i8nj=kgxmH83yM3|lHl$o>LC9}8;mfC0V3
z@8=o8w3cb#Y4FXe_)r{{-{o0d&gDLk3QO1bK9)+U<aXOuqVYhGvmClYb(>ABvfVD9
zP0E1!qPK^KI%G^)F}JypEbUXsf6cls)n|A?0A?iLFr9y@SFZxh0F=_Cm|e?*f#Bm-
zP#TJi<MIf0E1ZiJ+i!zdx=0(>w#aBK(9I_~rhZLM#pwcw_e}q@T?BlWt?=KrD03PJ
z9E&pJlICz+7kS(uRLXI%RV>$vpzB1&gJt7M+&w$n;t-aDy^rEHR6!k*Rl0uAHA{5c
z=V)~ipz?p%;lS9dsrLDy*ypj+5?hy{zY#~k!l5t`c~qM-7l3sqVz46G%^-^oo6?jo
z2R@vjuRYrKLCHi~pg%pa(ORsCLKY0!s)(%{8i?v?lP&2weeCi5NW7qtmZG-n;^O<=
zaH*nP?T%41Eyf(-IV9(BnOMyT{FvHa0Ro0alURQqAsr8->q$=LFHq-CEnQcu{DEpi
z@Kqmou@Oer8XMQ)SmAJYIAkKVbQGqpvj%HDL}T+*?6X6@JJV<HP*)-i3F!!Raj6Og
z5hiNcXu?-7?SiqI5#vd6zL!6k<RSK*Oo9WyuoVcOuy#C4U2GGX5HkP10~ri@CS=nu
z9SDDdMa(islPHuv^(4};)I{Bg_Ca=HsG^w4k(kPfZ`D(~Clmt<AMGB-CHrX8qHDWY
zeiR%dSvOwg!+bE)Z7l@e<CsiZZYoklk*w;fX_hg&7#I==r*`)hxcBXzKG8sR?w*et
z$b8VUmz=uEF61=Se-=ZXbti;|;1e_2fEj-_<ZR!NQ)!ajXTb}ow)+S#?{41E%OGNa
zuW&1*IqqVtfd8qJJ)9d20lc+_K+FEN364{AgD(&|X7%h901ZA`q$wm0wr$<^^sqyE
z*l~B=#`|)?!fmr?T?3vMZU<P*2gG>C7ebQ(!V*JlXU5ZuO*Hp;4yG6BU@X!>#3+By
zwox9k<B-=Sh{ned1@BJVunOB3E@a$vg{&16If@I_FfNqQ36Qi@hXyq1A%;h-1hE;S
zVExjCJlpw`ydjjE<E2OCZ1Q!Hp>!M7c7CG+uCsMsLC!7NOZS-|AGSKQ-b3**++GPB
zguD^`xqytK!<R-Mw?*p#@$Z`cqFaB*zu$j%L*pSTz>(<xnsawexW<H*pJo(_tB?xe
z9i=`M0`C;=Y+wS!b-Xr-pe2bV4X-i4X;#Okp>MkJ2~c-LGCd$!*ry{I!J|@!@$K0=
zGF&<a*tCuLa1-Z{KVf>8`rusg+5Vub2gmBz(0J_FfOf}$=13={uPlzn>mW1KyfPrZ
z-^5cQ1xR3Pq>Av#J}c=K!D#<@fbO2{;3f!Nw{V_GqX#at1ATLQ^=~=tX5N!L9w!1c
zGLr!-DSz!6%W~Yt@vg7f+`zII#=P*bOP4s3laySQSgcA?wh!PgNRG7&PykjGeek!Z
zd%6d}fU~5^QKfRof!x77d;0yDY;&{8Hs8I-`1g;OFTQ#$@=cXixy(10*Bg;$5{9yk
z*a)T4A}=?&DN?D-=JL4tF|qB1ERykLN87c-#eY^6BB|RwHe_P=P0f9KJM2!w(D2A@
zb0Lzxp^y47bky3_eg8lUI^6dI{%d;b8U3U~Jq-4zOzB;Xr}Ah$yc>8wIREF%zXHfB
zz_m$bNpN4)gWc7K1D>W7fN-G3_w^3XytQpZ9dDYtZ*6<y=iPL@9dD0AN=N9l(t5KM
zB7dz6pkxnQe+!6!oA<nee)cs0*o?U7tJkVpV7Sz2DG^k+_zlenX^^F&l77Yxw;#p{
zr0$x2$6E)nc1O`Jg<kAZ0o8QjcljsH(Ag$Wt3s6yk~GgH+$GhO4xmwD32$1I`DQWc
zW#m!-vJul^*O6@iapiRDpE8s2oaMz;T7SY7c+LY)C>9U{>Pizp{7YmT4d^SE5Pw6H
zj0`4e;0-?OW8|znHF^9j@zhyxHF4%@#vk|1#a84+bU$FMGN0gjd9js6nZ$4<id2|s
z;@7mJ`^o)R8zGbgP>W)wLzYE=z?SlJL}ZtMm^Tp?4knQ3Y?Ulh&?TYpya5?;?0<Y3
zh?_yyK~<cN%-rK`JqG*^jsDckLPpc<Ks`52+w}G5)DC=J_uX+JmO5!WJFp5kf_jl<
zvTqI%?hpVgsFBT9s<aX!<j4?J+06GZV|02}G7Mnjv0wz?1cLD~vIjD2zSu;--K0#u
zh+QpDVSJt|il_*c!pzL|n7~_60)J*CU`Uu0?&l`{xd9dbcr(DEiJcg1SfnbSPW0q#
zyZ$&UZLle0Z?8@hdGA3p6J3u@-BRDLP2VxeTb?$I(5(ACt4m<7=7S|u5&^a-QnL^c
zrFfDKbVFu<D7j!RmQfMVc?DcQ&tem@u&^4GtB@O0WGR3~mg+C=B3aU%dVgwwX|W$J
zj7hxp!@+C%WubtMes9~_8B_Q=fE^yGIlMpBeKXL<u^DfX?Tchj9rw=Q@;QhYG_!59
zpLXZ->h=-rY-EioQ&m(1dwqA;J461_9w%tolH)HyD?toE#M|8&5_mf<3z3Sl3Xiys
z8O#|hS%SP}dAP=UIGX-AP=AvUC_P9bxvTpc(9`roOlEj@oyk;!%?x%2hSpC5rG-q5
zUZIM(bAc3x|H0n0QD?^J94wZi#em%2evBy!nd)q9^h2;f7}exl+}~R8f6C;RB&9hb
z@va<)uzF<KV{X)>HVagnU#>N@Wz^}wGe90Ul|S&e2+-RsEvsT}_kW+$vLS&`%jWaF
zw_r8T(FHWSRH1Fk(yA)gLa9)K*?qhuBAHKP6Dc$2S_NC20cUx%RPdGQa9!Q=(Zt0D
zobSFR)L(Ze;vxey17bZ;ySp3h(f-OFqwySOoR(xWOKO1^H+|prkGRfcAbJuwS1{Ls
z(p<ToL$^hz8X1?h>wm5rm>aIo3wQx?0IjV;!h0_|LiQppmHGtJXjTlrHo#Jd+4XwH
z6`aYmGO(49D+3EW_L8txrX}o{r-felOfOztzIgv);)z_P8vJo^6jK41-5p>2_;a?|
z!yxjX0w%b3Q;+BnAUKE3+ZTVQ5G}M&ekX_|ijq%kBOWY31%C@g1wl{|%nG6aaMh%V
zIE10SlQLg9ggOvmz;c1Exdhm&kQhslU>2;~f)^?UPqUHpB)A`w%6OT@{bW()_)~aS
z@-qY#CIqPpyfmXWi#capXTU=~M_s=~-~nxym(K#EqX3EEUJ7(#0MQhAEOAK!0!)Wu
zyl_@lk|qJj?SEnxYgv8XmwgrC0}yPmdS#jw6P9E=Y66FFyPWsD-}?{?ra@2Y$IPo&
zv1yq<Yk$h8><{q2355UC(X8bopAZ~pGvh}^c*Fqz2l$D);1J%Y6$dFL_TkL8;NXEN
zxc8cxjI)MoL2q}W?p@6mq{qC>eQ(Dp-14q>JwzA)qJJV|ocO9N|F3d^d5#OzXW@e4
zGjM@+T%bRN3-U)?u(mnX5<hHhq%&wwTN|xY34nkvQ$e`@9~c|nNL3KK%g~C1Brlc0
z*iu$csDI;fNm2D5rkrw!ubghzpRU`l<EpeMgCx#5-7dl;&cs9cDlZ~e>Td<O%BBfF
zMz|1s`F}aC{@q-tC)@b*89vOx3dsl`u6D{xMM|G%eL)N+fY#;H45rXQ%s&T*O*{QB
zu-G=QqDK37WG4U5XwNa)3dcd#@u|d1<g`fQWijjPr^9m}@o7o^jSF^OeY`u=F67b%
z!enPgb<Uv=bopF@rmzZwVUuw@UATA}MmkqjfPZC{r%ubStgDp35G4ZKA+fcF*V56k
z7)@s-2$19$?ZUv8y{+3FoAfCS#wD34gtJ_+z>-dh_qpqv(PeFhIVv_Q9H|P3nK6$3
zH1DYAD3h-SE$}i#!ZGF=Fu+A$@0-pMnlr!DKLw7iml8F``!CBZ1SN0X=&n279h#4J
zq<=7&p?7XJp-KYOYNH@zaysV70?$C>PB(qy+7%yJ`Wscroja^MJha`>)&~mzaA3Y*
zHO~fgR>|pzkX*TiR2rK1^=a#AW2V6HC5<!1rr}|*SCj)Ssdg6Q0#(2jD<({ctg#0u
zftNWUuwaR?@zFZMG*Nw>3Q;-6ZGuxF7k{NmT%u8OW>03x@nQ={yu+I}G*Z(K<`ZvD
zE`<Tp279!Jx?ipp$!mL)(y#=Pr9Ou$`2BM|jJ83%NS$2TtKru6b!%OR;d{Ws*A6Dz
z?2jmk>sxRJO=&7+xmyg5*a9UIvlrCvo9m`;fI*QHg?8kF3mrPV*F6BKK^*^uXMZ_I
z8deOq-RZEW*|yysPJ7<UIoJYGT^}9>C&JL{t|h#zh)Gt2r-^WaQcQy4#jsf4Mde@(
znIhL+ze@CYy5g0F(_tJO>?NZ5<kLxvY?Zsjqa8rqIpG<ovuh6zXLqfA+6iQM;wzJU
z|MnHjFzB*FbCwVi-Ll;0X7?8Bx_`7s1_0ln?2BXwGu;OYYreX*H>e<iAYRX89-J(|
z9bOqU@=PvFcU!9a0>gG#X5t+A!M?-In**oM+is*IZkjR7f3I%>zIY{>49)w~T&oyZ
zLv<skR(GfVgFEC8WQ^p}KxW{um{epOI{xiKNw@}_4bg3~M}6}(xF&^4et)q0n})i7
zK#c{e0Tm^F(=U??J6>N@c>>b%4tL|eS-0XB)lYB{>{NpW1uCTR9Hon${MDgX77(N6
zoEZGaMG2uKa?!}@bg6(am5iioX=ryLTyD8aoUFhw^KUuJBDDFYIPR+!pgCDXHUN0T
zo1X^eozkbpT{r1eLU}*1Ie!6A`dPQ1sc2w+t@0)gH$Rce%JYQEFb7wEgwK70c89X%
zgJqiq{f;kjoNPF>ERT>ZK?7{<2iu)^V)L=-cXr^hj!*NEJeFwO|E_;WBhtg3tu351
zSX^CLI8t<KdUb)U=raEG-AE{uiAweIZGmL9%a01G&}*8GlsWii+JA&cI)OnUJfv}_
zYhGSfsQ{-sSpFr~A%G7+n*19R=)tQ1n9AE;IU6w=f)!fhDE>Khr$LU~a}waM<6P}I
zqm(3sAC<SCAPig<TIg>)ldDIRieFZbW^y`V6A4q4dk{uu^F_sHK4#@%48aw}WLh41
zWgd80rkQ3O&azCJe1B|+;H`asY6g$MJh?h~4Tw(#_yx{AjWMij8_-8X!|1#e>i3f1
zYF2D9uAivosUk}?uu=f&Pnk4lN+M=jhw)yxj5YhWAbVUK@m{oR0KtBZd&&}P-(}QC
z>fJ@qRMNLEEu*L04gB7qyBfjV`RcRfiZcM=SJm6yg8kl(Cx2f)n;+}oIYjNI-8bwE
zt0Y<{Q35=Bv#9Gw(S^xE8P>BR!HvuyAX1D$hD>J#s|9S4;mArdtRmS0jrHuZyD1Bo
z-7Owy#j>93I#7||1Y(n(M1woD)1dtAC&q#s<ZMxapW=OmsrVL)`tWgy`2<ugo~1Y{
z0UI3~deugxSbs8xnjNUeR(oqeBjx#7+g)j_r#stS0O{Y<gF|(-vfUNi?#kQlDq_2<
zWn6?a*mhU5?FJge%{Op^SgJel<b?r{#Z=sM`VBB}=K#l+?*S-T?kzhv(CK{p5)LNY
z*VMd-ymsa!x8JN59e%wz>v(?KxR?iP&L`fZozxKikbe|Hi;#E5VhgQO;hm9w!YvTk
z6cW7?k#LF|J}2J++7Awx6Mxak?tLK`NnhvYu^s^}HA#ITeWn~|)`N>sO6B+CevG<!
zxB1EA$tjaUJ>`K2B~qC&bv)VY19%ut*XCl<wE;g)$zObK0HPjj^LA(7*>UThH$L-(
zq9{^Ps()}f8n7PSog=EPJ!?6mv*bHVP8I?t9c#ePj=Vv9a!DPWd|(&JU*NI0W3|#S
zCFd7kw;Y?Rr7t)H&wR9hdHKr;9LM9rSj|r+ml>Au1sFRLN{_?pwb}f}jGje64CG{N
zdP57b7-(3m!c=KlmZSxAFy+<di+=-U294GVWwY8UNdy7rvm`Im1O&VM?V7XuGf)Eo
zF0)KGMg|0FoPWr(mOW7d0<Pw>=|7MM1gLO$D6@S>tptBe>&xy6&}A8*RB??FiRivc
zRlb6X3r^!Gog?6LW1p;FcSeu>f^g#Q-|(#Im@9An06Y&XNUQKD1Cj*Dnh)gd{{gg*
z#{g~^xe>T2hUrYrR{<-27_dLgxNKv;h?&&TO&2-Sreq)Qj5BA_(SW+TTx`Ja5NaJ)
zkU1=ii?Dx0bw8IP?&C7p<ha;3p&HUp3@p62{@3Y|#?E@<LfI%0#*w7Z_d!<C&dnJ}
zE`McGCF6YlDZ@xcSW|C3H%f}z1_@P)Do;0XOEo%LpdggzB7>>d!He-a#iJDaVKf0D
z_BQQePspNExc!lj=iS`3G@-Wy6fZV?NnOSHshl~!2sE9$0nAvGv_n=OM(zlo!Y*@g
z^Z!OO8Vzm9cJY@2e?c)c80$pVaiTa3sw<s$4{!bfDYVdIgJDgFVNC&tVNC+JVNC;`
z%mFmE8U&{U6eUY-+dOTNoQIK;zAeX=2yGslFPFL<1Xh1TS!!6qO$_X?Zi#6fb}g~4
z%P#oeR^Gt(nl-FjF!P=ZWL6eP-*bWV_6wxf0Ey6~^avrgk(EA{m5SRixC<x(7Ngby
zx@j>w3t$oC=yeN#AjYgQ6Nr<HU1NGE=y}kBS#@*4Y;`}ksna~{!yKnEzVv|4Rj#CH
z*Z5ExPoaO7FO}cXaZcfN>fU5_QGXu~n@K$>ucqhI^1S&f?(Y_vj0o5eQ<WTtJCG7v
z)`6V*-!nt-*2=B7!?I}W1S6cF+)7qep^Px}JKCUyaeo9_x)nQO!#$(9w6$E0>e8&7
zdk{%gtV)27qmC{kx}1|#o6_6e8PJ=2AABvF$+>^%olJ$a73@(&9Q^>2T8~I35<VcO
zqLW>TQx|3o>cqdq6ct!(af-1KVI~aY9K;IlQn}ryNb%)#JQ<A6v5or8r8pmplcxD1
zq7M~<*kFS7)x$Bhp(@cEonFgGpQA*%{0~f+HthCp(XNYIyu|rKZzpTPonW?;1s`yF
z-%3Z;+Ny+O+akzb?|JWOQoLL_AI7ariVrtTic9nG=0Qe8gGSzYm?U2YyWl^L*`QU}
z`FM3&e;o`b*X7`BkdKOwKy~M^R1}xe9|R{;e{&3ZH>;N&=)%d~lkH*UHYC{lvE8RF
z9MbednpH@FZ%BHChwX$p+t_^~{?Mht0o58?TBCQjtEWwWTJ}c0elt4lv0;7qPB6~H
zk>Vh36edTPOCSVJe>mH~b(B(u`=JTKPn*fvu=&Q5e1zFKS(43jrNd!Nd(18BozWX#
z&zed9;<4sruqo&sIGED$(*RW}cMrtA)RW${95gr17r||9ftN+I1y=C?W-C!^5iSM4
zm8=N}u#Md2WZpfNY#S?l(%>IFu!Q?~P@4KlsV9G4c(=$Wf0<lD>j-3_7a|f{y29hc
z?8x_Xz=f7xQz^^o025fyQxl+p7Ea+5T!h%^iF-}$^9Wc9dUOGPv?C-*PGh8gx+PBP
zQs=qbs%mp~(Xxw@aTUK7pHl6jtcwMeDoZ*9$w=_5y2x>VN|G!cI|~i&!5e*Yqzm3H
zou@uI8)*jge~3m=55x;N2(-`$Ju7-(gWhR>&}}cy5WXTLE4XitSwbrt6l+Y<4%nhr
zk)+cVX$w0Eq-CJhvLx@*tW_Wx3iFHTZEg=2#R$=gNF0c6CR#?qj#Fr{)2rJyM3Z}1
z<Q!&%&yMSX6jF)EjGh}|?g1f=Ea)bk9x2f}M7p}qe~3E`+Y3D;@q<@cN~4O1s9eLL
zCkkom29$FMBp5OnXwVF`D5*%&F|}nCNi;GF&Kp7m4d8RQq?;~~G0bp>m2UC3l9^cX
zDqE_xjx@2x5bJqOO=Ze-MIZQ#=;^UyNe_-R&CtUVSQtl7grJV$s3k8Qx@46NHUm*W
za#-u3f2#zDbTbQLl`!z17*KkTim4b05zgCS*XFlLrPN{GY}dwHb)DZf*L25Tu32ij
zT;_Spr1X|XS`iD|Hb5}sR92F9SbpPR)6iB-X_CosXmYqY?I@APMg-1gk|_eZzyJm^
zaSoIxva{(7v>|Y8qV-4r*v5ejMet-=Jhm<we-ok*=;)ctFb<ksy{0J=pFBk)WOG!q
zx&QzAx4eL`tl$VNuLxTju)yS9n({%BQB%>9i@@TjoMxF$t+uqLDW7T|lBRxAnnvlS
zwj#ry@;(x)y;bbmJ)2;5^NP%$W3w0##@H;W1C{fchQ9PcoOPNx%S>mP+%xx(^MRZW
zf7_=#N**_+trg``firQgL|38g%~@s%D(;vzgj&jLW#(oxm1nS1xvNUwf|QnyWxDc7
z#z?~p80X8_txXE0Jg2nx*r(ku)k@z;13jkxl%g<|;<6NxA!fLl+a{G21snObeX}h(
z<z#(nJKmzCBfR+Rl6U8Ht*G)gOQx7He>uS%C4_9E6H`)}Bv@6Bx^mi?WE#!Kz>bvT
zEXBsxlPR}~B_+GyNS-{vpIi3oHU|*iA#Mod_oZUL9bn;U&ya0*k)}bR4JL{p!tz5F
z#~sjJw&f{M&>TY=io!zY*l#gayQ;29JCGSe!PRP-(I;EQAz@B=PC4Z*vzpozf24BS
zE!UdM6VQo-DUF^t2&G<={wYl`=CspiH}wdsx-VJr<ntT*znpekIriDBX<(FBQ<?Tq
zeoOW)rxc!2V0t^=E0nuE&GNbL`52v53yh_dX$ceTIi^h;A;rA6eyX@ZWKyOsQm>m*
zNyH%M#XE!I*~IF_KABsVNME5zf8j+ex@IV)7%lfn!YS94(u_h%ZD${I9ZReEP}{!L
z$}D_4mpMc87AG#H7%-{jNDCL}sFaVY9B-2x#Zi*O_*PHsPDE}l*XJ^Bn|n{(PL3({
zKj*HR4{uw$wg7CQa#smfk2}k&(aGy%H0BZ+z*MRheT44DoJ%JO&6`I@f9@|WNT)U_
z(A}sDPIPTJJ2-BJnAW#;rXsg&G}$*3$$XmR&aBSI*VdM}ijP*&WNGW^&`pmJ$lOvH
zBn>;*Xww?duCtOOXT?1`uH{Hu@evE{wWHN^97)}99h|HHeNVovbG}wr97k1r<fh{{
zaVjV}x_ek7#ox7oEetN7e-t!*-;<pbK^!>bNz96#4!)t{s4MAY=z5!3XIU-<A;IvO
zLCq~{<dq0~thR~a9IthEU`iLLB<oKnO!{VOq}`*rNA_Tb*VzP`sqsFqqfS<aFAzu$
zWQ>6WmB8m#y4v6ni{oNrC<Zi^j_StIB}Cq5p4T$%K0aL$1BgB8e;k`_e8-h{?bl*e
zG!Ek7ziNS2qoKdJ1zISO8`|nGZ-H*kPr5lM_C{>i1vo*Ku2htrzm`Ic4v<|qyXjK@
zowi*!-a2z8TUyA3UYG}9S)WyGZ|d2Sb7eM_%QSD1?BLl=yOi&?@1VLm7u6;2jbN<%
zN)bUqql<GlO$e-}e`$3(xNkgAbauhINzgLpAZXJjO*358wwtbo5;vK)w&zv`hD!1M
zwl-vHQ$B6rhoq^Wl=>Js{9UAmFPw6T=Tm5MHHL0T>FfKB>(UK!sM2YM<g`!MiIQ8B
zTT<*eRNGH#*B#OFr2*RI-jf`s^nHHns#D~+^DYcIpSBUpe=LE7j!imragMydrZ3dY
zYcsr5MvpD?=P;kvLNe8xN^vwozr^6H*m+G-AmurwOZjva+c`HxJ{zT!udyjO4&<2@
zWhs}gvgv}`E?p8YuDMg4W>1##yI!ETY3i?=T8=1kNxDSmOaH)+0Rm<m^VPWLOa93B
zhLMa=GAnE-e`3XP6`K*diJ^OTa-wOM2ELr+e4xv&`OyjzY*W*~Ql|4xzRG9Xu!&gc
zoO8*KyEw<j+7|crVxRM%Gu?{G_EAAHstJ!2306LTPHU0d&1I=RUC|r9icc%Es7okO
z4lnRP^J5duGgbM~)5Fr-Bd2+l7LPyD4WLhvU)thye->p~$J<Qv!?21Ul>us!<|s_k
zML1nqF>igkL31p|0)d}4(33Ws^Fi}YSLr&L7Raz14J_%$KPj71N57M3F<79H*3HC4
zopJmqFF7rKPmkb?=8G^Ui<+a5N~2~@P)X$2pWDr4Q@yciKT!*_km9<8P|9=qx$`)h
z9;rU9f2aM)U_5Ik=@IKMy-U<Qefr|J{o@}Fj}AV_3KhfN`Lr-e;()(J^2EC0iQ;EF
z^eZk4yDs*peflnmi;z6*jbGM-^9y_oZcorKRQE(u<*Qk5IOy+>&WCk@?vH2n<!|&n
zYEOK~jo@m~$3<^Kk8uAcW`p5r{r5Ef^8s4PfAn4J<;Uxr^?CE^r>9R}L7$^n2b&a<
zHY=p0dx%0$NRL(}y-}qP;-sGSL_MAkhRsNd@8Wt{!LW;hM=zf~J$(Da>!Y{LWp5;_
zCkM^&biJPR>$fNLD+WoM8B^39O-Z%KUAld7AfAfn;)Qr6eipCAk$5A16>r6{cn1tV
ze~O+s5q;5b00Pp>%bqwDwKx-J_#1phZYa*0tBE)l7vkc2d{K|YKztEHaVbWk5zVL;
zV=<<0F->TAF|H?r=2ZMCCSoe4!`}2l%*5<sQrF^2d==kTfKETS*%s*M2m9}iewhQU
zZUfpA&>_+fGZ|>#1^Nc1%yvLmaU;+Ne*pQ_62L9A3s4Eff%xA!exm{4Ye8hd>aPG8
zfPHH+T-V}X^`u!wOznc$%a<Qszj+SCj(W3;XW!|2JA;NO`nEp58uljZ!Bd7v>MHt4
zA;oNJft{(k&{ZAaaF*Txoa^GIWf!-TNw5E<o(=1>*~#F%{m5jZ&rGehvisq@f5W<Y
zUHj==Os@bteH-U=FdYwj*NX!B;oCBvW*0nvJ=*{0(P73jDSH=~T4mE#nQ(_`(_5N0
zS8W9JP&@+(e`zU5RxWX{URb-c4iSgF%ahX{*f?%CFEep)jwbB}UfHxeng4_1m+#;I
zwgj2K9WmKDHwxn)1u(BR0!BcLf0otT$@stg`Qy(Wn{)>mjIL;F(axe@4okIUmg2H@
zaVxWQ%kCoce+UQgj9mbI*X^h527c){g7@M#@%!8zbav#@b|hmqnDnnM&xZB)6_>%o
zi~7&2-cZ!v`(Rb;MQ(8<ttq_h^(W|U$s#t1eB;HgU*l)S!>cKr&7W7zf2>Xw3o|+$
ze81s=IF8B2^=folPp185QnzkN{A<<L>YcpPt2ak4KK$A;;SP?Qpos{Ie(g<P)KTWa
zMDUx5kQRn;CG}VJ+|8#Py?On|JlqL)=*gif>1iSp-acGUcMIVUIo@eWo(O(RoWtB)
zA?6zY+(tG;wbPT~ngAyee=Boo1>v2%%rD1>Z(o0C5sr5tY+5E6{wD=0JQ(4)*}pu^
zrsF*@?ay6H@+fUsn4+0kieQ#u*vWJJ^Y^2p*S~j6YntpPqJ?Z(b#M<NTE}nL%~VTj
zlhT<Q%VCAy$%(vp_TuPG>m!e^FHf4`^v)nH7Ed@vqvM$O9d+R|fA8le&E|D~&CZ~n
zyM=6TPp*c;8is#ZU!r$$IT&3{mvAsUnb5<Z-hVuLwv5AF!2(W>k^a$&vWl(@9s`DW
z3k+*i+{t?Ge|-1ykJrm8-eE&2X5fD!A_Luw#>Y@`GqK6b59@w2zMd>o+U#V*j@~~1
zBZYm<sJSyN!U8_gf5c<W|CYopRJ3tv@AAjFib|>bL`BI-Bp0!K9PUjil9+y3>o1ZQ
z-RL!Loa{zdmnVoZ2In_U%4=d7Qf4`(SsT%;aW~t=F=i*T`*#12-=6*S!_o1<(Z>5H
zY&wIpyV7tzQFTIe<IF}=yXkzmWOkhHuGO<;J=Pf2XQO^|e>xbQW0=9&Sq)g}msg*Z
zZ^O@Fp>-RR&fWK7aSH9*B`9?4BENRE<^<t&O5Df{$P1X3iPm5qaHSPv#xQp{Hsv)<
zaWnIU*;Y82XjV9TSNL88bu)r5qU}(kh(1R?>n|Y4Q~#NL_0@tug)$+>-=DX_x>|Ei
zpP$TcKj9ZLUEFrAKQBVHe0DEFSUKHqH%qDbIE;XMeiAN}%m?5CeBncHVMDu+ytuR=
z3`OC*ID}ad&tQUS2{QZ)GTAJGm=5_1km>(Qi4nyjVocX`32M6ZBA|KJ{|8_~irtf8
zO%s>xE(DAOGdMFdmwhh;9|SWvGc&iKF9b&hQaMCLH8(XxI7TxwI5b5>Ml&-uGe$5l
zL^(1!H$p;0G(I3aIYdP@H#I~!Ml&-wG(|*4Gcz_bMldi$IWjpnLPA6|K3xhgOl59o
zbZ8(nF*G2Pz&$8`CDY4Oj$s(b@#}X#Pu{5H6-7G8S;8AS2su?kg;GxCREu_YhJ^)N
z3vJwEjD?jMGyV;Kfj_{`&VOKrzCWAK{GR8!@9VkWx$ZY2;%|ybp|br~kOT<|s3@Y@
zg%ax4$%e;V4;v^zZiE(SfX%QCDzF8%vv?EKD;i(8^F|7Pe#es3pr|gnK)=*P)im=+
z)8y0!?pIZG&3eaH))`Kp`5n1R!KdOwi36oB=zvy5U&VWOLObk$U9cMtLZ@P&*X#Dc
zUf2iSupfHh033!I98!#a_qHR@2fc6<j=^z`)=D}7C*c$fz#t64S?E_xesjNJI1M9k
z2A;q<7=`nHaDn!o>P3?GUV>$~3}Y}36L1A4;VE2&Yj7Q=U>auNhGJ&TO>Zim{c*ab
zc=yI>R!Xm(=A_jXr`yu=A*Xri!zZT&>C3Xy9i`ewr@Klo-#gt?`aSG)U&Yk4(*qT+
z+ME_uy#4>RB^CLn(?b=X7o8rdtp0F%tg_#GGdCw6Gvv(7Gc(VOJ98s)CvZV>x5`@6
f`d?%4Yy1NS^Fh>?{5J$13pOz`Hwq;sMNdWw)p-p|

delta 14794
zcmaiZLy#tn5@g%9jcMDqZJX1!zt*&E+qP}nwr%(R`(p2VsK`3xH7hDJM<0-i?~vli
zfLRiS*r))<+A_|EBgj3M8hI}$bzOKy`W{s@kel}%)axGE$qA~~jgVo%qHAQtpn2pc
ztEZRez~F*NQr^B&#hSyO3ccQ5CCK3JLTEnj$k~5-dwSY@DhmWDgr)ulUh1eO$7vzy
z&_qcGU7Fo7edX7fJJSskMv6sB8;gvKH6K6Jz8eEV?$z4q+79*NS2vx!U%g(=KV}1@
zMJrSpCCXWPPaDQkF9*L~BU?Hg86z^h6IOLAB+Z+~!QkJZQJ0pEE<$g^@UTruBGqJ<
zrfh>d5KbsiwY4F>OS`7)Ju3!lTi(5Pli-_W)9h2XuE&({xz5hU_E$Cmdeaup*!p_E
z;NE}@nI`xN;FLZQa4N&mm`ZinWU@Vfv~*csoMB)6)=Xm_6)j00Hts6B_Aeu`Y(d-k
zu9KWz`$>bc+Oo_B*Puc4+f3Kvl8U>}3cs{+)+x^o*2j?JV;$_*wUi+eW%jhcbPJ*^
z9U8-YEvwwBO#SmGL^$fiissc$vZ*6bnSp>yM>xF)kYB~EWwFJi?3p8lF86A34Mxh{
zt=jZ;OZx~nv5nFJv%x-yG~urD#9lbEIyPoTZgg$XGYG+2jkvr9tG4k;y5$5$297PU
ztilGpX;VoL#lsv`{4Nu<(j2>@Y7gJVn-p(>qt-3eX2?d`jV@j1?^JA|%1P_hyCy)#
zC~zohB825+2)vo797<_`Qoa8I6W6g@{SF<zy($W@Gf3vxoo}?fVtk_ZO4(IAq1R>;
z@9V{>9>Uqy@9miQ?XF&a9Pft6F&jGjmm>Ww3<2I=bW0+-eIa~N+fHv-7weT6)VsUq
zBBy|{H&4X+7S^OwK|H@g8)fY`k|Q8<!cF6IqhKR0Vz*B|b?Z@4ID<e`A=Xl@KCivf
z_WiNOGI&r^c#<bdjvA#L1eMANWf<$%Ii;eG2#(xMzqVamb#+MHq(C0GjL@T!%xBMU
zA;r#Va@pHhTeJFg(^}=^(wvj&v|wLrxS%K*_QqyBgV6HTx`%e3Da`r(0~-LJbIBbU
zOwJ&}x^>eMxe`7jYSrf7)JNk6t@6Y2c6QQDK5^;ui3&{jGk13}sPCH!f`w$9C5~i#
zPU)1+Rq%1CdugObZ@G!iq0w&OGMI$9-Q=X@F`ZGsB=s-4JIgIPQ9$+4NNmy-E>xSf
z+MNVBy?`r?#5Jido{o4P!xcc8+V$K+b<w3Xbl|-~X|SutrfLHUp0T;coa@F0S5HBC
z*CC=2LnkzdJ`$fQ<+s^kB0rp$Q}_0dI1j$+G1GOu3Qz3Jfn7d1ngiBQzp9+(-V%Yr
z^HC}b|8HsRkSc*0zMG*Tj=bLMmfOlfR57Vw40N6N&<e{Iw#o)JYBm63ZVUg_k><4<
z_CZ{2af%pZ?V&^0&1S^#e6m6xgT+a=Mhru!xHb5rrGmpxg_`7H{xt?OCj89bB!I(_
z(&|87`KiB?M|s2)i~L|}GJj*)r^vO7ZYA1TGDOyZr^RTg5uTDji*D{Xgu^$ZKlm)@
zvy^)>_(aXw-mT{Mk{Mus@|#vHPyBd2Fl&A3bH1-O$4$qbn_)(Z34S3O7<6;)U-)bf
zZPc>H`sACv{r;E%KN~s5NxHnjtRZ(M98%DwqqNT4BIwt;9}&sI__@L|9XwpvnnlM%
z@OQ0zz6hsZ?@((+c2X!XjrE1ULOYRt;N%;*HgN};K`evVsR2L|?eoX+(xQBaHu4S>
zX^Ruz9u-%K(WHYb*5IHj3N<EH+*^vq#m1|%U4<F`cdUr0N3aS{1d3OP2k>_NJUojD
z$($%#CTIp#a7e3ZL(lIQ5l9{)S%d~E&=#KAw>MWLcDZ+PV2>AzsI(ug45mC1#DV}D
zyhKDl(tAtc$O%x6x~UzFN^+49r8qxRnwX9HU?MD_5UJMzn4`(nK^j$`FDhKkP`qVF
zER8_=DFjXxJcUmu_UTq<7G}|UB!EL`2?yvyN=3H8-Y+}H|3N3lik?N0CIdPRQ%#Z=
zEjSL=6-N{P%d+HU>=hszJiAcRn^OKMRf6+(*0AD7FAb1^6>t0nS8^e%c{-_YY8{eU
z%DgDM40;2b2x5Mu+PP@OCsoRJu2|Y@h)dAuJrL3yvz>kKUUx94o@+6Bq06&kv02Zj
za(p!w1BnZN^#=PUT`c?_sNU~g!pxJe&JB#N{7v6dk*|&u!F%M(ZdOHg6KjF{_vJ5F
zYwl5&X9M7b=`@6w35Ubxfy*K-16`TWC_MkO9ZxGGszi>C=>ktB^HLH<u9i2ft|>fM
zODEUB>0(G$Ea_ZuIJ|HBxG_eyH}RisdQZYPfFNW1^vyjz9BO|RTbaw(`~X7qrkf$<
zcbP>6u#yg47zm~C_!LE&SYtiTs*sDHeh=%$YX|_xPaoI9`HuQ_x`1mI&KA2oCtbQ-
zqFnw7Aw`?n$JPTy=;Ii1WX#Rrc-Rgf|ImnWo}Jk>oq=3tkx-k3qLNd17N0W=e4^jX
zbd7<0r|paYz&nomBxjSzBZ5IjFSkLCK>9m@Tv|N%Iv?n*?{<y{ke8qDb{5}s+bT^8
zpab&U1{u({G%;?<lMJZpODall+Oq);r44ukxOy6py5f@L&Ah>nsJu5ypl@>l^EZSR
z&bZ<d4hs<o?ku{`$Iy}vN8$MQa(>)1HbXV^URbmb_RP22lb~<z9s$MTiycEllT+@u
zQh&~yNiKI1^*z*8VP4a3`nla}1N=X4A^_B0cS!P}-)JGTLwT)wj<L}87R*<C{Y8#z
zE8f3pVeg$Q_JhM`zG;;2uO>L|_L~g*bm$h=ygLjs2~Aq$h3K)U!sj9ZvvO6~B3cBX
ziE$>AAu=zNeS-Had?1aqUp#wmT6ZQ|1-zBX027ho#Phc>uMJ%@CK7C(t*=8S7Qhrz
z!#YzTf;s1^;~CEVzVwp2=A3Uwp)wf~;<eLR#Gi!RSMzB7;tw5j5_!MJkXX>DU5C@m
z8rLLe?f?PdTYu!-n8<{jQx+B!f$nq>bv^ZeOPiUrUXRT(2>ByYH3wp9v3vAI9P?s)
z?=4xyUqIR9gS?6<cZ;qExiwUdMu3Go8i7^5!vbBk%)mvuX9akBO3z?}EhQyJOI?ib
zN~&3{x_vofy2T}YXT6R<lxCZ^kpZ!!pJ-g)$p)*-2lG>^{<=p4EUSo>+1^ux6+aQG
z25w5;7x-2#N*j?`OuvHBz{As+a8~0c)38hESwkX&F~_!(HVZ~=XmCUUO@Pw?3s$rw
z34Z;&StrZb7)w8@7C|u4wI+;%w9=8<=&9_U7J-zI?qOW8wtG8o#@*peHXD864Ywvr
zgv!{Gkp4$J3wyW%yJt<IZ`P$C_DBYtqj$h1WMI?3e>8l$E?L)c35sVpwClNniPBjE
zgKYf6(&v=s1Ngpaf(RQdp@3PFisWCap(HURocdv3CB(-n?r{VHVqfl$v|FspVQsrD
z@mBc{OJ3-EtB*6mV|O2%TDw?g=?JVd1V&yXh#^Xtb%?+6X~2A`dmh-dTxJl+2e11u
z59R5ioYGeP9cRpuLy|ADs8^Jt`GQKR5SI@+Nf1w=wmVLvA!MDG3xMXOLC-w;<1m=5
z%G&!@RH?Q<DV-ay*Ela--{?P!AQcS)V#S}FeZ-Do4JFJHdn_K;sPTgee>!yW$P2&R
z@g}ziNCmOEa2%-&n|wAdJ$;0*+o}~Ns(nV~(m&0LT=_r4&SuR{Ub|&v43?fabDrpx
zCgj6bBUA}a!sa<d>j6cdMkrFR10{QT@LAVe(nig_U9=vvYjEbm&8m#CdzL(V3!sEH
z;WS+n2po{<X6sq}7sT=dq1_LJ()k;oiW6@Cp)jHJyaf7JJ{mUyp>+JwNnJ#gllzw~
z5@>#G=?9^_kiC-c^GWjOLAmYiuC6{mFK?KhR=BOJl+;QCpo}4M8sj&!Yy0H-__`;*
zc@4Dk+6;Q0Zw6m-5_%DM$k>}wodsXXTr(u|nWjUYXy#3^Dl)4V59ox)M%no})4r@}
zU&C&GpEU<yVB1^Ho>`#p;1N?UE?QNDyp4;3xE88^>7-<nX-u?eU(4_VyNhnb3}a?z
z>g?iVW@HP~=m-Y_4#LHp7zg(UaIL)^d)R^u*zFyN)h03@vS$_|z}c&@ivPQC1!5I1
z15-^``>(M$+I07(=g9sRF7%`6YP%>D1pI?694BFL-{;GG%*C-+5);rPefW5}Svk84
z({ne4YlP9737(pub%-JEM6rC}*8N8H{-Yg1R*ASO^N-<HWxBUbKVyj;u)BO&Fm++^
z@o1aF7JqiKYNCGnHj`LU)H29kn%cU$L*}aP*XL-md7@c^ASUyniP!R^ek#{up#SUh
zsbT0b&h@x~q=DF3&}H#<`q;V!m)fTcJait&&~`i}-O0Gca&E}wefD<P47(_uJ|NnO
z-vyn!b!w(>5t`LgVS;avn1%ueSaMYnONb)ah5?aKhU1F3lC(%~FrS*){7Wf?GeFMe
zQ*O$*;J_yRm*U!#tq#{JYhgyVoOKRH;O6n^_&EVLlXv!>J8hQQyo)zIb{2}yY^PZm
z5?$HHNLiD13eW2wJl`yNXmYW7?5Av->UmwEiLUW&9*nU$zDP)xVTgh_;P>3CJW(d$
zkl#%B)3KJpAZ<KoC=35}U%((GRT3Q4ZYwnYWw&4)s<LE)ZM2p#WPE%nbMgv?Mv!Yn
zPk)Zm2>x(Of2+K4(Q_v0kn*uRPhmi(?qggW4IV~3QAG$kVE=3<YD775;fpdsRpMjS
zrt;8EF6c-Bk*Cfa;hTa2u-YMV&8BI`8ep@kYWT=qyk2}F)uJ!*8px|@`6nF2@93|W
zEdsiHYSmn^O}`oj&4-e@SmG!!&4u6E0Z4tZd(@k&P2YL+$;iz!|0eRON=vqBhD3aI
z<cT7t`zVW75&Bu6wcLe0i_W>$1y803U%Cp~D4xfeB&!8I4EuaD0PJh(C05*sva<lM
zQJ0pr&D}P{uA4p(@|Eh0<P5mo@$F)}?QuNf#R#9uk$e=Ql_3m6+v!f-Q*H>G?V2TS
z8<h>6Jm!VzkM{Q5&K}+&L7fWY$+u&w7y-ZjU-%AZ-dfYWO1X_2^`43s#DMt<D|DF)
zQxy$w9~G&Yn{^8^K&K;={k1r@gJGhPOjNrB78b#5YEUWAaXsTh8DOopPg0YfH&)0!
zAlD)ZMhY(@RG6T?yZYtbgs578^@$}WE%`4AGTLqWs~<IqS^7NlhLKFH@j&+)7-`I-
z>60abV9Ij5lLQj^R~Qob?&TMOzxW(*0*ebn<F+$u6;h8jp!1TV(rwnPnV}Eeuu?R;
z4JjBogp=ju=N$ZOp+?ghYM{p8zW5I{=y};BZ01z7q4mb;8{?C>C9$$CQvAMzTJ0ka
zZ-u<vL(894RQsk1X)=S;W9pW?_O_$ARYM?Z;_!XWOsUz>s+kwAIX2AuD6@r4rrukX
z8hQp5Y*?d7Kvd^;J_i3Rk36sfx;qR!au)@azi3EghW>cbB}vdQd#;iWLNvs6Rv(!w
z(XwZ1XdRGTaO`Tmgr0_@;FH*`-OiJ~BW>^!Bcv{!*+kar5&o{wm7K2=D9=Fx2MdA*
zkTt8*Su1eY=TT}pG?{Y?EU^DYak}e1ubDDUT&+73K<uqp<ji-|o&}XwrB5OVTYhv!
z_79U`N+GV}xJHv47A8;}Gwvyra)+0*_w%J1WP=K{PBjLMa7&V3{f5E!lXHjWpu!Us
zirX(!+>KEU5VKokW@WDmJ!^_MPXY@ZYm~XoH69S|BL)FlbrJ9#B;!m>84`rMYQ#H#
zPme+;;H(^yb+7*4b)X5JC;~WCL#sSP3W&$6=xZ!?ia*Ry8(J1L!FWXaJDBrU>!WY-
z%r%<@o+><n`egvjG)-)dT{SLrU;j=}#GhJm+UUG)bDX$ue@mi@a1@+nqu|B~ieDs&
z!gR7ke-8;f*Ap%@3awa>?D)g3p_u(xJz7XnfWeG6{SAeo1okkHFQN!x8^aSFCQ0Dp
z-po_mEh0;mh%Sh0Jh2>3gI@|we1U!xq@B7NR9W8&gz!0%&*l+BCTw`U1Jaes5h`9*
z$-LiRy$Hh%gzc$H-3gb}=PmkNgY40c0Bi;Z!+<|d9!%q)9(TsBV~6TbZ}dJPZxE`f
z0C&j%nZ^-$^9(jv()Xy>I#T}|HOT7C;Lj@(^gR(o1c$EB^&g|u*%8rVCY>Y9_^?nk
z)h;K7r}u8gI?aZ*l(O{T1@u3fYM_>p4pq9Z_wD>AgLiN|4_d38Sy4f03}x5n#-cFH
zx7~+ajr$Dg(Pk?mV>jU&!&T0r(q^!rfCjFuMz|BiL|SlkC;A7Hldf5^cV@pMKYn5Y
zEUXKFc~|WK=zhNit(YK5<#Bm=)WM$+wy|i=IqJm}E`?=u^^*SALFT0RT>hrm6K#kb
z!%|zu$w;P6y_j`DkU4iC<R8KYhJRP(z_}pO&=gfbk~D@lAbG)e$Hh1z6Y(440JOJN
z&J@G)7KT4Lp2-L23WXjAY7wWupAn@7g&rU7e|A1q(-e*(PuQdj%c^(~L5Q){VA}2R
zKk%S*VH8m==Rf;~n63K{*=tQayrYU|6xGJ>4oy~_Q1m*&T5l8P*p;FMDzt)6z)-t6
ze47z-RUch){oRc0IHwo-h;m|C0Gl$OWDM7dCRQ6X)B)I<5T~gu611$&hQc_62kV5h
zc%SjgDHD_7g??%WWxtdt5<K_dURj(!^n#rCnAHNYq6SlF{Hb_D^uoTPEOzpU40=uy
zP{LBVUvZFCxQ3)Z+($HI+bP4fql#0&U6IG1F@^^7;U%!(`FcO9cx4e60J(<&uA&4Y
zI7KBYAsj1OO+f5%sbbSmL@_wf8b*i;U%(EE>QIGdU^Qfv46niV9E&z5;%L&-3K#6o
z%HVlnz)fglcLeiKMPtEL$@%YS2MyBEbpG#p^0|U_8nYN2Uh21;sT7=6!gxeif+T?n
zd@p5Kb3IHi0qOFdCRk4yz=vW(+vW7S#3eI@5h{=FJlwC?*)#SIhn-;sPKQcz><2`I
z5SDun7}6_rca|l>if*x|aW7NtKA)4ece=47YtokpJ*7tAQ<J(lVbfM8d<0x*#_0@Z
z*_OS2!mZPXoeK0ka)*%MB96&yc<V^Gp-F-GmkqZ=HFqBzJZ;_wFz8aD&j(`zQI|w9
ze5FZ%2S!K;UYtM1Ngums_0PH9_%k&xg??hBS;BM)<LD$2Po<Zl^uPspz`NvkogSrU
zQvJ<dCy9$K7(bHxFyX{O7t=F~NOhvVo{e_3BJWalB*@}WDvOWDdx)485<s|YYZ=H*
zR_XGc$3r`R%_Z>{5Gj~Eii$9#FV{SaILseQNq2w7oMk_RapzDbOGdDKG6n9iU5*tM
z=ErhYFeqX*kvhnqRaeXR@gg)XsvRf5cc0g+86Ql`*`S6TYga~`Q7YGBdihj8`bv5Y
z-rR)JyiG;lvkM<eJ+XX)MYl+D0yUe@89F#*Exs08ev6(5Fw|rXKSpl2C+g5YI{FE8
z`B|E9ZJr=3V+S*McEm&+7}GiIchywa+`?y1|3inLT+X8vR8LUQ=t=#uPQ07VWpS&i
zK>`nb9_H#;4g(H_T5YqmUzutyRAoNYmY5|lN$B$6?@O6&*v}vfK0#T9|Ei1`8!qcV
z)oU*VAv7xru=ywk2i6>-bzTb3d$J;n4=JYD5Pw2!P!U02R*hLwl{c=;`QXdP@_0LY
zMM}LYREW&KHhx!f5N)vIoe1@0T$G~&+c`{g(YB*z7LBjEsIW?5>crwi@mJ~mgc<mB
zQ^-Bb>u{KrfE@F2FlFG%rs#}YoO^G?+Nk0|X7yJ9Oj&|jco129f^sEOtovdV_**Lw
z?jLVY<9cBeLA=4V3TGVB8-sr-P_7U5%p>>dTtYQn*=HK^<m37PpQSCAf~&&=YZp<O
z!ng3XUf0Ls7!$6;l9mMQH{j*x(02U5KfwFUY$3#HJ~-8alSV-AB4gralukK2RND{)
z@=)Ldj;m&J4et~?qA{V88^>40LNbh1y(4gfw6d!oVIDNb^*eCE+R2YLw+1Q}G&6sg
zpun6MqoP6P-USWUKHBIuECT_nPnaTPKi@Z-xYmfs7&^wZRx{h1eJoFkVp#>w!)lT=
zcRdQZ>i(sOuY`>0d))d$q6-PCIT@=^7#YiedaJWcI|Yzs{BT)BQH+ywQSqm;y?kHq
zmmcWVw55laIG!zQt|Nss>SBH{YM&N?Gpt`td|d0;QbI$j`h2bh>2sRM>U(Q&d@yXv
z1p5POB@jX?aN%;O+^fa}`6G!(18sj7xFPI%h<Z<25}AdT^ayll|0mj(>PVtx-k$)A
zc~2B(6dG+0F#F1?k-ybz`4O$YaVpkaO`@G}>LF$+YL`cM_A9}WbS<w+D!)VjpgH!>
zaD-RTDlq^P7y#*k+?K&yr(glITbWEl((P=0NoHbZLljvJf<W*I<nheSN&sjOe&}?f
zAOR&P6D#Nc^F1=oTmN++j~X_1YYpyEo}f<WC=v~a3&i!B$wfB!eR^@h&9I6dVMT~P
z*XRGGCy|2^C~DN9c#=SmSLZ(;T$y$??B^$+jY&BGUV+aOI*!b&WAU`<_v4VtC@p$+
zaUF`Hoh(9j0m@xAR}K9C?ogsK1vo!~n1PE~@3#Zs=a9o*ja!FQg%%Z${Vs8LKkS&E
z>aZP(^u_MIruKqHx4nJH0-)D{dE7dnR+Uqo02)3KqV^-7bQf*Zo}6A9j7`eUGoxCA
zu_Vb<nz}w^V;-b0fdEeSm*36}TT8s27gyn1ra}eB#KMGZvI@rJ#!8UA5KA-l9DKM}
zJVgo6*JC4p-^wXDIZQ;}P`9dQ$}rdzeoLOS0UI+y$d!Sx>D#&OXi|9TIK#}c@a)xz
z$}VHvb&%r7hv4&%<|7o^z*STcM6%f}K_o*w%)y>HYCm2LW1uTS(1Xdp9XK-^EhXRp
zNQP^YjcS%>4jJ(@UxXD?ul0D`fyCi;Rje1#$J5m$!bOFRP|0o6PJP)x_wnGLn`-Nv
zQC=*4QF3upf4l_b+<2r&Kzg<(NR#T+h?;Qh13gAwx&ZxBYhvuB&cw=in6FP;rgb)~
zX9^#x#-bBEI3Fz86Olc|c$Pcjgu9#<bqTG5c-(o2%_t5O1Qg1}I1&4mZ!<Y&-6sOT
z-*l=kFk6VyfpA=r^QMlGNLT}&GStz7EA+6h{J}Z)qK?Gb^rP7f5!GL;s|j5#;EQVq
z^GMFEEU3-k7O!fSokbIs#wo&VAzmgzAA}wzsQ{Tc^8c_xy8yN*Z@fMaiM<g$%}Y*S
z%gQxe%Ka8vD8knE0bW)1-LcLl8eao=hASw;SXv^g={3E+xza|~jIQ1+DmIyg5|%df
zy4AIg1LNAY-pR{&jPwPl=@E=9TlE(B33et;1r{%`TWD(*b+@$x-wY<WFIyF*Q2L4}
z6sh|Y_(!aMN{IGZ!5VQLYV~X#FrP3|Hi$yNq-GpvQO_y0lp+|!2av=KC1wH&==}N2
zHiwf!)HTWtL3~b(U4fdsGX4yMFIo>%i=IYN<^1ZoidE@9u5;?5i2JwOx?(n>1(v&9
zKc$)mFi$YGwRP9L=+p(SmQfv#L*EXN22YWm6XYx9uQ$!u_;t6;sTt~ZAf&qxuz&Fh
zKhc|M*{%Bha<O!x#j!M+bH4%*x=u`-fG6pX7C4G&XB(T9t>G*F#_gZ}A^T=dgNpx{
zv#uW^J(_C8K<QC#&Vt;C)FF*<T|l&ruIjac5RHdO52(Aj?`}V@RC&e~i{d7d^`?=s
z!<Yt<qu*??BFvL7&GUk}O5K-)#+p_;TW9O!TeE=FZK(MyA!XI&uh|2zo1pBtEkN^&
z&Lh|Bq)5Gk?3QWx8+aS`zoU8ap~A=;YkRI`MJ9)WM|F$Wn9(lAp)*#RGcB%XJ~k>?
zkWxgM!W=kGv!U)8zLYlR3KV66x^RC<0rgiCdO0Bnke;@A0$wy^g;NNqtC9u;i56}(
zB}g>$q=k=iSmM`*%ftW}K=~PS@%|3yx512;{to#AeP)-!&cIC;GDcHUfl4sS2w}qD
zv*W80FQ`a>D2hW@oY+K9yRAX%=i<Z&YLo|S#>`1h0UT~n!l@+JZ-RWts_gL`a>-#R
z?JKf$uuQC3-jgm-@BrR_m11=b3#`D3>(k=@Xm9TgU*y0b^VR@+&&(+s@iHViY(j!<
znX%sdcipk%i%6#Bp?k}fTdXXzAGudlzo2icu*U$d%w!)VRN<qlX~~8tLs8Hl%&+i~
zy||svHE&9yA4Kg}0uWvY@Ya6iLT7acT1#69n;py}p}y9;qn7Fqrfryf_jxfOAgy>e
zX@n5U>EFsfbF+YaK|f@)H3ol>V<jSkK#)%$@V~i<v+p81;vS(dO|421&_qK}gg)Pp
zGa(oYd(OiBl`#>#L3Wb#Vn4y}t34@bY`>EUi2+{|n{MAun||0F#q1&~6}c_fJ}kJb
zMwY;U-xg*7&43v<!W&nhaM*F_(8sQLK928UsDy)MA1ojo3PFP=UAADpRLOw2kCTPl
z8^>?~JnTe<$VT@Z0Bp>YD3Sfriz50(XvqM!5iuHBaj5byb<FO?Xy1(eVtMi0bL=f?
zEPb5n+RNG9nbv2bn6+qpCz@KM9Fg^(cHltM1o<F$vCqbb4~|zx$B-2|JbtNxs#vI=
zCe&T#&<VggBT{G7i4`2XT8&s@(^|8kI>JkV-j_BAzrs$yWjk#^&BX*bG<XhZ#(6QN
zyLQ7xMZId3S6Zq6%N&rQuBwtEB@x6C#_S=?b&|ojC#HipF#?-G-Mhw55z~VX%f>y#
z@JJ1YVefm}!?le!*;z5lnpXB|#NHt7-!mCKB?O2kM}%&d54UVxP}BK%&1#t%2!A*l
zv0uTvy*yDxmyrF{9d+sR!GXV;O58vC(|(6~Z*8f~kyJhUPYPsvbxi>Ux3jaRqYvlL
z6MiQ}1W39gA;o(kUHWpXAh}^W3<!(lliAn`zFgtbYB7?PpnchsOV6~+!DUn+j5mp#
z{3ZbZ>TVvtYbXbVuqMlc?F|M3?y^F~m3HG@H>az*2Fne0PSfTwEASLiY*nM!BvNtQ
z({)}?!&XV7M`)4B@IXgFMU6mS9vRf-x+0K*EqF47b1C^4q=g=>hjtTXPRfH~9iJ%+
z0tNkSPoO2e;13R33pW8M9uG`o?DA-Jt~j7Mts_6CvH@{w)Z*O=s@OK$MOFQ8T^_tY
z!N#m#zChkJ3a_#SjsnD;eE-RN;<p(M_#d1rh;k(-2SZ<Qzv+pi(hCPQ<3|OA8-JQo
zvm4x8(_G!bO<OyZy=RG5oMyR*^Pal58;X8u@G`=_$YEGr{yPLP5t=c$M>2`qT|9sc
zzAzQ^IL4W)tK{e}PUj2bj3YJVzu}w768Mda3B@;gO4slFBY10x3pJ2^ayovOA~dbC
z@(*)WNi@ODcsdZ}jd4a^WrS2K2rc;V0Xlx7y=bYr5WrKa#+~P%?0MmGcXVuC%UAQc
z=R#2b)(9`8iaU^y59eh#*19g5O<%yWl7UF)Ki0YB*YDqw?=U`+xtyBm!m<9-Qfube
zB65&vgp!yZUEQsxG2wW@CHCGljH0HnU|PM7Ek_xnzH*xGt%b_4mO9)>?<*^>6cz;J
zb>oj!)gcHxe>+c!5W1T`^lfk<I<go@orWRGHLKsAY8rVGF|Krbm#li|Rqy~ZDLvB(
zNSQy+2cqsUVtRPDJX2-h?;P^3Tkp$*Wb(B(yu{(kk%jSw{GIe^En4lRc8;R0yFZ+G
z6QT5cNcP0}Q{*TD=M9-~X*5cEuFqC~I{O3@b>RffW383Ei^1%+jS1(izoNyD@w!;5
zc7~0-=fD8_0dmvFsF*?vC9;5ZAE|SO^10Zh#b?f0a@a{L`3|HqU&s!L<Y%P7cTA);
z`+G(iEXlvRXdiR^Ek8tq^^~i~H{^P0%PuZj+qzGCL4(U?Vynr5qG@VZFD!4}x1kQN
z%-;)x*Fg*Pmm*T}NwNiEZsRN^&sXZ46Z9HZD3g<MlBTqKR4fNXN+keg8T0<`{6k5W
zs2`68Q0z*ooktP#`<&dZ{it`dJbQ<bA$Y;GrZSwRye{zKTSi6<Y;RMULz&N<tNRm*
z6y`PqiZjeQ>z7Oj26F%TaiL8LG>Pc|;%s&<)w(dm-FBB%ECbm|Cu+zUANQIqv#Qr^
z4AI}3Gv3O_<lC8X_ftT>`zi8n_H2+fN1%^8|GQnB(3USK&fC_OP(i=5U`q48`P3%I
zUSBljR`)-Qyo9Ho0!2t1U(>Yyk+aW<KnkNN3CV{`IrI5Pj<&{-k=0_vwC#V#f`69^
zv7Hc?ggrz~_zvnlA<oN@`oF+;0{xrO>AClzmnHU{qwx2>cTfQ=bC@?*O*kagzHMa1
zJuApCMlq{iG{Wi2(X`xzyXMSQ64&&mrORXQ16}>Ns5H3|f_-$%CIhxy;oY3V!TYw=
z`-RofBS{0>Y9cQ^nkU~qADYjd?q>5Vvs=A7lbo6n!!D<N+L$+4S?k<!$&1u-J++Ku
zPQlQK@E%u8I90|k!^pUQoQsjD@b&^aoHZg3$i~l4g;$f-#yx4*==?|v24-LU@QECF
zr?ZxyygNO;dcNL~z)_T+w=YVi7ZH>0<nb5iYE2Fm;D}uyVQDz>4rQsSA1j9AI8H5o
z!9E?wZX2IzWWa#m8-*Ab!NBBvuM--v*c3s5!yBzQ#Ua3EkOBUUjr>X=AX2A|5JGW~
zU{s?XK#dua)nI_BBQtmKvj#ZDinft3Q5S<^Y6GER@KVR58RWb6dWDr*p2AT-Ax_6v
zdyf^DspM8apqx<jA~laBU?~vOX~>6{U!aXq6CjSLEvPNDL6?$~N`G_#xHUp^-IjWT
zAzACQlE^G8m4;aNk#v@GOF=z7diFhcsOA`&F}##vF-<^0N3IO=t0T6JlkQcfBK$ce
z1jdfe@8y%EiAA?#zD*L;p``NJCsC#4=?NTJYA<tnf+PQLk`WncRe5U;F8Q!+P%yPv
zaw%?GF%CYRBGQC&eE-tr-mpjZfl|abV}?JHccxoCrRhA?*3a1Jquq^m#+(jF1ZHMW
zQO&`ZsyvqgSjNmYkU6dN25}`Kse7=3Q7Tv4JD?CdI;6$4^Gzn8HxCOZ%9Z4jQwa}Z
zIeDftl;anuU1&}uRjC9RH7H&_G+rL~zX55Dmj{iU0%rWbj;~b=7YqQ3T3l_s$!n@R
zvq*7oOW+%#4xvMqx<&w|4uI86xLh0RL%2*yTQSl|b(Ogqkk$6YvawyMv+JMorjOIo
zBJ!ODU)W@b(}(S#qN#Jm)hb4dYTOO!*At{-RJsJ(uEW{}_y5c2H4Q@K$84URE;whh
znnnQQM$lxazOBosvps8j!HW3FIBFHez_;g&&JI#Qekfwa`%k+^72qc^%Jpxx+&ghy
z?Aw>uhT^%XD<el?^6{jytyjmCnZ!^;R?P|TMo=1z0pV!rHwQbgRo$)|KSfN35^Epq
zSiPYF#;mK|^`=D#_V5R)$%V#kD8yUb+3IgQ-g2=P*DFy%g=R7kp#;93xC<9$IeT$y
zCrcj3JO7(Dd!q9^4<H@Js!~%O+3^Wf(wdJXUGmL8D?;9lTwX7YsSNcOMMc%Hewb=R
zJUE>(${C5KP3G7wmt;6~@!ph?dzAfxBlW<CRzBk-GK^mp$~XvmMFVkkDWpQw-o=ZK
z)Hep@_(6ctXrpekQ>>HT?jGe7u{*m8+&yqz9>b^a+?Ph{uUMt>hi%~1bIxlH#qebE
zqtVtx(VN3yQ5!3wJ%OnxtL0I4sMLd59qi9x8oc(u>{yM`=hc1(Cq0B2i1)HLC{=d^
z88f8R2O=;^WrN|*Mz8e%`m8Iybpmxek-%=kP2Y_HVmW~#t$%25Aw?1%Yw~H0(c6><
zh<Z49iq-YSY+ml2UK>xPX7)>$9EFPnxGSNzV28YTR0szU7T{v3HpF1^fRXuOdn&Uw
z!VIFXrt~cMOncsOiL~rQ@^xMfZoHAE(XE0V?C$&Ls`MuYg1_<*nSnO}K~(qx@QTGw
zuYy-HeCKkLLe7Ui$XhEI9$KtAhTy;UEhx<x`^tAZ;~D-yn`5@}C)+0DE%gJG@IU>a
zGHtPt#pSt4UJs}L+#`Jw0ORCMF@hLg#3c2_G`#XtF+Kv2x|Zi`$qiWui~-<$Y`{2{
z`)Kn0Vqw`iZL3TCvHrz*5}11uJi-Y_tl}4?ItNAct~s6673>z6be(^&HSYi3Q!Ud^
zbZ9D;Hx!A)MS;y$MB=@cl8rCaH_q)5bj?#rbnI;}cwQtn#+wVq0xZ#<1-tx%1Yx7+
z8VexV&$3@2w_gt6_GyG00zXgE$eS93>PN_~{pqaHrMUItTj~-8wIQ_WWFGjZwZRke
z8GFNaPF@_8qL8Y?h62va3(Ux_IR9brba&W_u(pSa%@~QlU9}NN5+{dEowXyM5Fj0l
z1==l~$D>q>6jgpk0ooSpy5~et1U!<9Su4WACfM-23gSxa1i3;51BrtIAf^=~6C%j)
zRdm$jMM(?yx`;*RfcQgYI?h3`2jSc`&oKZr>EY<L4aM4*lBj>->T_0?R7+$vo&sjW
zatLTrp7tac;rM6+{y9F0LS10RXg@6E%c!p#vLJzzg*Fp(0HpmTIU0%7p+SGcfs0?u
z(jwxK_PgNSZ2dA-rSVUu-E6zmR^M%%<=n51Gm1+l&2n00rR}YX(Ch1LLBT}i&B(1n
zk2w(7;cbVLDAM6VDMB+BV^FMYME7hc(n<1Q;S7mTPX#?mcrs@MaY05Yv;j}Cws>QT
zU<xI}CYK3J0AiScn4EUF!@Tw7zZoR=<0`De>=NXepTD($Nm!^!n3%qYh`M56SlOLo
zDnc}DHgvjnDB73_tCaI4Yni_p59Ke!Ru`1TR!Mf&G|XRAx1{wOI>R;FjnL*CwPRmg
z?2*7mQPT<Bid?x?XEKCCTBYfYQ!Xjve9nkmgk$G*0N;tiBS(`BZ6$hiPSjIr7xamp
zEVGongBUB3vf?gvb31kwPY9mm&Wa~GNvm_s5}I7oDAq22$DXL|jUvSfXKC+P-;HnO
z>ZfSdIg<+oDi{?idKqMF<B-gDTV*Y!x_E<4j?DrU`IXd`Lpo{hfyh~UPj~s%2x|LA
z(+KQD0P2*I@E8R*6M3;zpxSX6O=-4N$-EETl_b}UvW*X4x>Kd5aszD4(KzT=yIHNB
zGtfT%L5QHA0rkr!Fh<S`+`1d46%N{lVF45gt$?ZH7Vx&2P7Vn8G*eOjA6l1~Zw!^@
z%8QI9&~zp=-Rcr--v%{2l#`^>B$XZJ%5pg&0CnlMQ*}EpaJFRH!kK4~;sx7z?gE&J
zJddxNN^D)lX9Btce{cQoq*ZIONp5!rgt#YLvK8WStLd$i^gb1MN$X*^aQo&WZSH4I
z?7UWUpQdzr3S)nc$wm{h@|5>wX#}n)h1mtka|czjD4dgi7t32XU87gM!buAGf`+9s
zzymq523wj+e9||;VZybTwJ=V3bA42IbitZ`sfD4L8FqKZY-X-&I9!GbLe`pVVE~+)
z`nZsKbdyZNG3ki0XBo8<H9m9lxjkz0WCyi1(M0-tw!QLXaC4=G4oIhLOD%BuaOT0<
z<$)VJx?L24F}XsFFu#-ObU{k^@nk+8@F@ekP%a1K!f_FZs)sv+5R*w<@?4!=K1PQt
z_cATYoJ-NpQu%n_+)Q3`P^iYLwc*9>vPcTX+^Py8HHg5z(FD<SS#>m~<v1Jjmuo@Y
zXGqxl+L8@lLYYGsl9EPnH_`qwt9BWER0*B1{rpWX3nn7whG0eV-rd#^HfWL(0RC|y
zYDy`F1c~5@(pXR$WM9U4A<d25v17iXl_Db~$>?LQ=p2dTNe)6-X=}_m`ga>YZ2_Em
z{9_S1u_qnJ>ecclo-b|Rl|8`Rx@-2bEHQZ>7Ayf2+hk9W{H{H@hMjZhpluB+0yw(3
zu*1;>I{r2Nv{|+@RzYJ3$XJ0JFlvsw>v*j9*{q2k9pv#=tDU!OdB$KDfB|z9((*B>
zlb89JXonEdjlAg!EU#EhgKl<OTdY_JYKl2y*RcdBwOx&Foyi(c!C=nEJOS2Ru|VI<
z^lCZMO_w`np4K7b@o-zDKkmHdDXW<ga~=04#k%ZK7LA)bJJ^y#1*x<GXlEnb;3A@B
z>($wzG(~YiZ8S+O!f7;LmsI8t$(2-Jw<ibWEBo|T!>4Z~Dd4{SPwf2S2ms9Teb`dn
zgLkwH-ROK=<&Te^!w#1qyb6~LVy^2&*Mg&!qZBkAgyomS4OfXKA1i;i_E)k|D&O4P
zjtexdn=V+9nNr&!W`GP)zDXj2UKZ(X?Kv#l;|3f<^1R997uQs2bDZpRi85aF90SuE
zwVQq-nxv=vxiY`ru-goL{H!?D@n0L57*Nj=F|CQ>N}G<o7Mawc$iFwLrR#~sUp1I#
zwizFnB`Ks4x(W32?+XHVCjS1|qdsgm?+F4A-Tyt7NmDSm5dcj`-C^@!JL9vQ>@3TT
z;s6FpF2b{hzkW@DV4F*<bxG%4iM0>u>^3AA+^4z^hx9Y`^)18Q-D{`3XR@uv4Zf;z
zQPu;bVu8uCxr^FIZq8a23K~x#wFiq6v==DSNrSMwb6;zskFsihPfvBMuVYsF#lwfO
zE}&;BN2x8G*#I%Em3#KZFUG0^2{ZqUA}*Lj7i4)2^(mcPyE(kbdcnO?_@|=j6Q8-R
zMYRt35^!c(2~e4jdxeRTrTJSjZ8!|eB3BkiB?~-(2^smw-9=K)!grFW@q&{-MGIx=
zsH!MjFQ;zK>#FuP8wyJUv_)<7#7PfXemA@^B@z`1mjEjOGuCW|l-JOQ%q<F=LT=BS
z?eAOYU||qqgYa;jJ3pOl0_UE6Ja2uCN5R`V!gs{vKuzSgRc?>W0SOtR_hIVu-WLdA
z&?YJ?+%jLh>bvy{F2cuW;iHjW-sho~?>jDE>;@{K6KNaHljD5m9O+hX_`0Eo_Ag)b
zuaCbvGr)J%>;3-2+;^wz<NSP4WA5~9xu4=+(KHOzIYGf3>O!HKl)dVC5NEjspHggT
zkP%D5L1&x3I^$Kr-uz^#0^;t^-N%`&^E&SM#kdd)uU;+X_s2E#cZAS_`3Rb;go^UD
zW2&72URdcty9cp;c<%6DN3xwTgJB&<+8yAaSU>@#yr{twuD}1&Lj$F}7~RYe_78+X
zJZ?nei8aa0@V~Y;0)r(|CM1I)Mcs&aOHr1JVa=#>9a&Ty{iw_FJT7!VM0i@nc^yYL
z5(`sj+EQ~HiotLzdfp<UT?_cfd;d>op}~n@<);4&m`m8xpe)3Odq9s6RU`hK+7{Fk
zFhJ0+8O#Cp8blSuk?6-gX4ewvO&bZj?t=yh!f-bOZngPMgU>1jS*4Tc<>5GV)jkKb
zzHsO2VBP&=XC{w0qNjScHl%yp>n@2np(DaiS&}JJ+t8wd-bD?PGpm>b=(MBFYJK?H
zbm0<EZZY^_<HpQsFTk9B#-dVLedWh_3#hZxU-je0T%_rf-yPw?mmMA4c`oqxduxd*
zT?d~vh<*9>bV<9Ho!9{{QD1LTqu4UqcuFamR>KA6Lz@*GFwllcNS48?!d$tj5*aeO
zeK=}@WRKaLo|MJQDwOIB&|troKl$BzJN|s_r4u@_mym6li5_^90G_Ih_D4o41eiU$
zG5;_Ce%_o%DYk_m3TZa9n;Oqj2bJm4i`%O7wkBw|8#>3n2nT?9O)v#_T|cQh0}Qxh
z_PXEXzWCiCv+>)T@W<2ODZVQw8;!nwht2Np3Vt3dIMAM7`8Ail;&t$({=zVLe?xoe
zG=#~K?0VE+uD+$;-&w<0Jf5nt04i062cqO5ckMlhVoVwwyK%Bt60LgEm#TG8ezlvM
zy(n*=E~e)r5qn!?+ky@$EJ*?j47?XHU6VW+k-yAIivAJnN`0&Gxj#}#*z5l?A-WIX
zdpVU!du5CTw<4}CZVSJO#JFoGPzLUzoHBA~gtd<dw8q0LT|b$&QS?a-0VZ^{g*+%3
z9w+bjH&+R@gAehcGE<~Nex(9n5XCv|UKkV_&pi-jKb=<-c%^YMN-WJ&NYc!(+#c|L
zzrN4To^SCNGblR4Es1s1@wbA)R_<}#Se6quid;%YoHX;@ok$pRx;ecTKgJF(Clk#%
zyICY7hf!nZ&p1zfCG_5=0H5trjh@>EjVxuScEp>TiIw|xA^U_Ysb_k+2;v%x<akr^
z6i;`huY|EVbcYTab^eJaMS!M)ssfszQGenckU<=D_lBHL!rkt#!{>$w_cg@`?6=&o
zF+q;PdnVN_S(}H4!H(UiTFS{JcAoXR!q(YeNh1Rm>l?VpI)4RLz`f>UFS#Qux~=X;
z=a*wmf${~vl9Ch3xEL*AsJBW4S&6~QN92(hhgUQ%W#|0zF|=`1|9f=eLxyHZl2(k>
zia6`vt!X#MsP*xyp6%bmRtte|2NxfQru!f0%$dEb8rCcY)oi3~&c^vAJ8nYybo|oJ
zzt1*Wv25bc>t(xH00i+<EOQ4pbx`$><7LX@%{ymJ%k7aO=htq{Ea6W#Xc+v?v0A-q
zE+LO>6b@$42apo80=6_rU8~_VrYZawIZrv2ttlU)7Ib-v4D^{D^d3y1?Wny-OP-`Z
z1+($9AFxmlO8|2N-{w7qeA%)4*VF2{nw6}@Q_9Cpg|~ljdac=)vyqB)x!aLQ>hiCf
z)5_HR2&8*CxxujHkHBp}25)mZYjgBu4D@tRB1)VL2uA8dENO<?fp7vrvh0yWB_cl{
zrXRh^#-jCMSc`0Wff@AOP?kB*zre^b5jp?ME>aN~!J@#JSs0m8a~Q!0|F^aFGlEG#
zD06ag3W>0D2y=-siHZxcGmCL>uyU|5igB^93W+d_G4d1f{(mWYFbuM0b`~y{M666K
zME~o-$rs1vRfW(-?jM)CVPK%H{Xsw?_WYr$1DzxD7AB5(5%<vC?V_N=ne5<Ub+zdA
z(%-=N{dsZ+y7!^C{`}5PGwgPkUTj@tU4(`8Qjz?dh{+W~A^Vdjgj`LZRA(1$<1rOo
zmjiNSz6K&!2i+nJ+SJdEVIM3m<~*Ivp?JpwNGYNaE1m`aD8{HLNtVc<D81hELl>Lv
za<?(h-hWEpji*8Lq2^EGRJ;hto~yz7y6(oGgXYPDb?fIpQ;%x#&?CGG@H_hlyaw;R
z52R|;TktZyi7<c!H$aRPf*i9V&kZ7Bgf_B=g%`@=psDnG=0~|lmKC^o2WIVnF|v;b
zxaM1iPI#3cVBti>+#`_{!l&<7X@gs21Z`x5d~^_Fdfkw#rPlv;>L}6l+;`Dia4yAN
znCw`TJC`e6q&t7~F5O&Mr7Go-CcVwNp{Aqb+E&rNcIv1#F1dIT9NrX=M^vNz{oNo%
z&-LTvQ@=OcvJx`ebT1W4UM5yPIWY|($V&G#KS_7AuoJz3pqIQ+<Lc^oS;hK9`UQG?
Y71B#3-~uCnW#eS!gdryvQxJ#wKYC}(RsaA1

diff --git a/doc/Projects/2019/Project1/pdf/Project1.tex b/doc/Projects/2019/Project1/pdf/Project1.tex
index 57ee02c4e..d24675702 100644
--- a/doc/Projects/2019/Project1/pdf/Project1.tex
+++ b/doc/Projects/2019/Project1/pdf/Project1.tex
@@ -145,7 +145,7 @@ function}.  This
 is a function which has been widely used when testing various
 interpolation and fitting algorithms. Furthermore, after having
 established the model and the method, we will employ resamling
-techniques such as cross-validation in order to perform a
+techniques such as cross-validation and/or bootstrap in order to perform a
 proper assessment of our models. We will also study in detail the
 so-called Bias-Variance trade off.
 
@@ -159,7 +159,7 @@ f(x,y) &= \frac{3}{4}\exp{\left(-\frac{(9x-2)^2}{4} - \frac{(9y-2)^2}{4}\right)}
 The function will be defined for $x,y\in [0,1]$.  Our first step will
 be to perform an OLS regression analysis of this function, trying out
 a polynomial fit with an $x$ and $y$ dependence of the form $[x, y,
-x^2, y^2, xy, \dots]$. We will also include cross-validation as
+x^2, y^2, xy, \dots]$. We will also include cross-validation (or bootstrap) as
 resampling technique.  As in homeworks 1 and 2, we can use a uniform
 distribution to set up the arrays of values for $x$ and $y$, or as in
 the example below just a set of fixed 
@@ -270,7 +270,7 @@ approximately $2/3$ to $4/5$ of the data as training data.
 Implement the $k$-fold cross-validation algorithm (write your own
 code) and evaluate again the MSE function resulting
 from the test data. You can compare your own code with that from
-\textbf{Scikit-Learn} if needed.
+\textbf{Scikit-Learn} if needed. You can alternatively write your own bootstrap code.
 
 
 
@@ -321,7 +321,7 @@ of your model complexity (the degree of the polynomial) and the number
 of data points, and possibly also your training and test data.
 
 Try to make a figure similar to Fig.~2.11 of Hastie, Tibshirani, and
-Friedman, see the references below. You will most likely not get an
+Friedman, see the references below. You should include an analysis of the bias and variance for the test results. Figure 2.11 displays only the test and training MSEs while indicating regions of low/high bias and variance. You will most likely not get an
 equally smooth curve!
 
 \paragraph{Part d): Ridge Regression on the Franke function  with resampling.}
diff --git a/doc/Projects/2019/Project1/pdf/Project1.tex~ b/doc/Projects/2019/Project1/pdf/Project1.tex~
index 0c4ffdbbd..9253df903 100644
--- a/doc/Projects/2019/Project1/pdf/Project1.tex~
+++ b/doc/Projects/2019/Project1/pdf/Project1.tex~
@@ -145,7 +145,7 @@ function}.  This
 is a function which has been widely used when testing various
 interpolation and fitting algorithms. Furthermore, after having
 established the model and the method, we will employ resamling
-techniques such as cross-validation in order to perform a
+techniques such as cross-validation and/or bootstrap in order to perform a
 proper assessment of our models. We will also study in detail the
 so-called Bias-Variance trade off.
 
@@ -159,7 +159,7 @@ f(x,y) &= \frac{3}{4}\exp{\left(-\frac{(9x-2)^2}{4} - \frac{(9y-2)^2}{4}\right)}
 The function will be defined for $x,y\in [0,1]$.  Our first step will
 be to perform an OLS regression analysis of this function, trying out
 a polynomial fit with an $x$ and $y$ dependence of the form $[x, y,
-x^2, y^2, xy, \dots]$. We will also include cross-validation as
+x^2, y^2, xy, \dots]$. We will also include cross-validation (or bootstrap) as
 resampling technique.  As in homeworks 1 and 2, we can use a uniform
 distribution to set up the arrays of values for $x$ and $y$, or as in
 the example below just a set of fixed 
@@ -270,7 +270,7 @@ approximately $2/3$ to $4/5$ of the data as training data.
 Implement the $k$-fold cross-validation algorithm (write your own
 code) and evaluate again the MSE function resulting
 from the test data. You can compare your own code with that from
-\textbf{Scikit-Learn} if needed.
+\textbf{Scikit-Learn} if needed. You can alternatively write your own bootstrap code.
 
 
 
@@ -321,7 +321,7 @@ of your model complexity (the degree of the polynomial) and the number
 of data points, and possibly also your training and test data.
 
 Try to make a figure similar to Fig.~2.11 of Hastie, Tibshirani, and
-Friedman, see the references below. You will most likely not get an
+Friedman, see the references below. You should include an analysis of the bias and variance for the test results. Figure 2.11 displays only the test and training MSEs while indicating regions of low/high bias and variance. You will most likely not get an
 equally smooth curve!
 
 \paragraph{Part d): Ridge Regression on the Franke function  with resampling.}
diff --git a/doc/src/Projects/2019/Project1/Project1.do.txt b/doc/src/Projects/2019/Project1/Project1.do.txt
index 237f1d6bf..b1accc914 100644
--- a/doc/src/Projects/2019/Project1/Project1.do.txt
+++ b/doc/src/Projects/2019/Project1/Project1.do.txt
@@ -17,7 +17,7 @@ function":"http://www.dtic.mil/dtic/tr/fulltext/u2/a081688.pdf".  This
 is a function which has been widely used when testing various
 interpolation and fitting algorithms. Furthermore, after having
 established the model and the method, we will employ resamling
-techniques such as cross-validation in order to perform a
+techniques such as cross-validation and/or bootstrap in order to perform a
 proper assessment of our models. We will also study in detail the
 so-called Bias-Variance trade off.
 
@@ -33,7 +33,7 @@ f(x,y) &= \frac{3}{4}\exp{\left(-\frac{(9x-2)^2}{4} - \frac{(9y-2)^2}{4}\right)}
 The function will be defined for $x,y\in [0,1]$.  Our first step will
 be to perform an OLS regression analysis of this function, trying out
 a polynomial fit with an $x$ and $y$ dependence of the form $[x, y,
-x^2, y^2, xy, \dots]$. We will also include cross-validation as
+x^2, y^2, xy, \dots]$. We will also include cross-validation (or bootstrap) as
 resampling technique.  As in homeworks 1 and 2, we can use a uniform
 distribution to set up the arrays of values for $x$ and $y$, or as in
 the example below just a set of fixed 
@@ -153,7 +153,7 @@ approximately $2/3$ to $4/5$ of the data as training data.
 Implement the $k$-fold cross-validation algorithm (write your own
 code) and evaluate again the MSE function resulting
 from the test data. You can compare your own code with that from
-_Scikit-Learn_ if needed.
+_Scikit-Learn_ if needed. You can alternatively write your own bootstrap code.
 
 
 
@@ -211,7 +211,7 @@ of your model complexity (the degree of the polynomial) and the number
 of data points, and possibly also your training and test data.
 
 Try to make a figure similar to Fig. 2.11 of Hastie, Tibshirani, and
-Friedman, see the references below. You will most likely not get an
+Friedman, see the references below. You should include an analysis of the bias and variance for the test results. Figure 2.11 displays only the test and training MSEs while indicating regions of low/high bias and variance. You will most likely not get an
 equally smooth curve!
 
 === Part d): Ridge Regression on the Franke function  with resampling ===