From fd5b8827fad3e26c9443b7b9ba8df266bf9c94e5 Mon Sep 17 00:00:00 2001 From: jakobrunge Date: Tue, 20 Sep 2022 20:33:43 +0200 Subject: [PATCH 1/2] fixes regarding typos in tutorials --- tigramite/causal_effects.py | 261 ++++++++----- tigramite/models.py | 2 +- tutorials/figures/tsg_graph.png | Bin 0 -> 64063 bytes tutorials/figures/xyz_graph.png | Bin 0 -> 7246 bytes tutorials/figures/xyzl_graph.png | Bin 0 -> 9421 bytes tutorials/figures/xyzl_graph_bi.png | Bin 0 -> 8870 bytes ...orial_general_causal_effect_analysis.ipynb | 357 +++++++----------- tutorials/tigramite_tutorial_pcmciplus.ipynb | 119 +++--- tutorials/tigramite_tutorial_prediction.ipynb | 6 +- 9 files changed, 346 insertions(+), 399 deletions(-) create mode 100644 tutorials/figures/tsg_graph.png create mode 100644 tutorials/figures/xyz_graph.png create mode 100644 tutorials/figures/xyzl_graph.png create mode 100644 tutorials/figures/xyzl_graph_bi.png diff --git a/tigramite/causal_effects.py b/tigramite/causal_effects.py index 101a43a5..f4f56584 100644 --- a/tigramite/causal_effects.py +++ b/tigramite/causal_effects.py @@ -268,6 +268,13 @@ def _construct_graph(self, graph, graph_type, hidden_variables): self.tau_max = maxlag_XYS + statgraph_tau_max + ########################### + # # self.graph = graph + # self.graph = self._get_latent_projection_graph(stationary=True) + # self.graph_type = 'tsg_admg' + ########################### + + stat_graph = deepcopy(graph) allowed_edges = ["-->", "<--"] @@ -305,8 +312,6 @@ def _construct_graph(self, graph, graph_type, hidden_variables): raise ValueError("Invalid graph edge %s. " %(edge) + "For graph_type = %s only %s are allowed." %(graph_type, str(allowed_edges))) - - # elif stat_graph[i, j, tau] == '<--': # graph[i, j, taui, tauj] = "<--" # graph[j, i, tauj, taui] = "-->" @@ -2148,14 +2153,17 @@ def fit_wright_effect(self, coeffs[medy][par] = fit_res[medy]['model'].coef_[0] elif method == 'parents': - if 'dag' not in self.graph_type: - raise ValueError("method == 'parents' only possible for DAGs") - coeffs = {} for medy in [med for med in mediators] + [y for y in self.listY]: coeffs[medy] = {} # mediator_parents = self._get_all_parents([medy]).intersection(mediators.union(self.X)) - set([medy]) all_parents = self._get_all_parents([medy]) - set([medy]) + if 'dag' not in self.graph_type: + spouses = self._get_all_spouses([medy]) - set([medy]) + if len(spouses) != 0: + raise ValueError("method == 'parents' only possible for " + "causal paths without adjacent bi-directed links!") + # print(j, all_parents[j]) # if len(all_parents[j]) > 0: fit_res = self.model.get_general_fitted_model( @@ -2532,89 +2540,140 @@ def _get_minmax_lag(links): import tigramite import tigramite.toymodels.structural_causal_processes as toys import tigramite.data_processing as pp + import tigramite.plotting as tp + from matplotlib import pyplot as plt import sklearn from sklearn.linear_model import LinearRegression from sklearn.preprocessing import StandardScaler - T = 100 - def lin_f(x): return x - auto_coeff = 0.3 - coeff = 2. - links = { - 0: [((0, -1), auto_coeff, lin_f)], - 1: [((1, -1), auto_coeff, lin_f), ((0, -1), coeff, lin_f)], - # 2: [((2, -1), auto_coeff, lin_f), ((1, 0), coeff, lin_f)], - } - data, nonstat = toys.structural_causal_process(links, T=T, - noises=None, seed=7) - - # Create some missing values - data[-10:,:] = 999. - var_names = range(2) - dataframe = pp.DataFrame(data, var_names=var_names, - missing_flag=999.) - - - # Construct expert knowledge graph from links here - # links = {0: [(0, -1)], - # 1: [(1, -1), (0, -1)], - # 2: [(2, -1), (1, 0),], - # } - # Use staticmethod to get graph - graph = CausalEffects.get_graph_from_dict(links, tau_max=None) - - # We are interested in lagged total effect of X on Y - X = [(0, -1)] - Y = [(1, 0)] - - # Initialize class as `stationary_dag` - causal_effects = CausalEffects(graph, graph_type='stationary_dag', - X=X, Y=Y, S=None, - hidden_variables=None, - verbosity=0) - - # print(data) - # Optimal adjustment set (is used by default) - # print(causal_effects.get_optimal_set()) - - # # Fit causal effect model from observational data - causal_effects.fit_total_effect( - dataframe=dataframe, - # mask_type='y', - estimator=LinearRegression(), - ) - - - # # Fit causal effect model from observational data - causal_effects.fit_bootstrap_of( - method='fit_total_effect', - method_args={'dataframe':dataframe, - # mask_type='y', - 'estimator':LinearRegression() - }, - seed=4 - ) - - - # Predict effect of interventions do(X=0.), ..., do(X=1.) in one go - dox_vals = np.array([1.]) #np.linspace(0., 1., 1) - intervention_data = dox_vals.reshape(len(dox_vals), len(X)) - pred_Y = causal_effects.predict_total_effect( - intervention_data=intervention_data) - print(pred_Y) - + graph = np.array([[['', '-->', ''], + ['', '', ''], + ['', '', '']], + [['', '-->', ''], + ['', '-->', ''], + ['-->', '', '-->']], + [['', '', ''], + ['<--', '', ''], + ['', '-->', '']]], dtype='Q^~c{^i3>482n2$=dyoXT7I#Xae_9HZ;sh=3P@|blWJ8@=?> zOJRzVUshHoWQsNuCQMjlkwp}J$RUUP;~)PJoXuv>IOB|S&N=7YbI(2Pw9^m)M&bn2 z*4A3CSa_NS_~n;h{_eZ)7F~2v#eV+z=bbur>f5)kiBq((DiB!}U=l_Z0CW;Rmr^G2wYmoH+4>6Hd71nrlLJ!Aq4t|M}0q{q1l6_rL!wvBVPFZo4fS zi!ntTOKcKA3aW@KR0T};?%h!ZAS@ec1`<`kKl$X7tFF50cH3C#&QARj52MB z`GANEvODj*bM)xZr<`(%S(?zs8koU=q!Toa06&RxVVElds;jF*OcUB{v&}Y-KKiH_ z5RKgp95^sv8p4N97#&2*x88bd%{A9F5&&%=AWBK()n>EJHhcEjXRXRwYpu2S-g{H3 zFw=0#%ge`&8~5IO?`^WlCdk@;`|Tfo_+ccTb=Fz$yz|bIOD>6>9&yAGfBoxUVZrt* zDk^Tg@y37t^PgXT{qd$ds{$skHmD+fDU8oP`;2TU$PCxI7Ucp-@d)2&-K z;mJA)CuzWMx#gCR#nZGlNkPU~iZKNM5r>)$88W0}$Bv)72 zG|lws(>rwNU@`#8CJX|nXtVOlD_?o#l`!wPC%Pn z28EVdYAKv4WT?XLx#ym{>Z+?q{);cZXy3m5O*h>{yUS&_qT;{<54`l!OP_h>ng9If zKYQ%42ddn2&pouSS+izA(1hJs6-Xa{{BfQ%5T5){g(u3$ZgR&N~l}Z-+_F=VIB6A3wfFj~*^v`V+TyBVT#t6~nEm%`Law zf@u#OI@HobRg8>)()#d-*+Klu4L98I>8GC>v;Y448!jH8`as;706Pwsrq++JY?m5 zBa*s@m?T-i63DPv@P^uz}@Kc{nhz{>|eCi(_#5DH4z zoZ{JBhv#2tYtTIT_3bBTt|9Ggk5w-p{HVxIdv2HUTixE{iM`7=yY})Wb`TP=leW35 z+heej#~SVgr6O8 z_hw7%25`^tE0131ijd;FX8C#XQ)7^j(D5#2`s9k!>lxOVHWnK3-u=m6 zPOFDj4Unt$+DNf8?_T5cEw1fS*<*>W%L|j!232a?<+`_2RltWW8+f*zt)QpW5c2RnGtuArU3sx?NtYdFGIdyZ5a5@Y1t4 zxdXg@Ic?PL)vRXYoBls^k=~bgSt=~*Tx;)CMl)#=S<|7{&<-`1Z+uIXOimm95OMoQ zzu$Ys<7OlB{tw-Oh;n+=D{O@X`YedKeFYtuq^EEeEdES6>~Bi}2cOuMLCdidAuh zsG}fqEl~u*02>esK>;DXF!`Q^R56d`8WRDL9#L!*%%@x-RY-t(z3#f}Buff~YK0gU zN?5W8Q$73avss;B84x6tTWqn#(@#Gw*w2tYMHMEVm>W(Ot7M>$6xLV9jvdPa0*fw4 zN9g63Uk;wd2WIeyV?_t^WJ(4L%OXmNrS&t2+KLFj9_qH?%QF99a#1G z!yo=&ZDLheYhhjK^wV=g08L0)>hba?dVp8v0Hg;1dhn##^F&3KCRsyNMX-JYQh4bB zF*88DX2@uGr=5196crS&Kf3G7y`_fMp73LKU%v%puJB6l+gl_*pHd3NlJeWI`%7#5AD_zJK!M$rufb(-<c`|rO;177TG zzWL_=`q#f0J7Z5QM24%(DXKVsW5$eO%^3GTc<^ABIs5nTk091NPzcLsrSrrSPo%}t zQLsRdMSt_nH_ThXKk>v9;fpl%`1{}g{_9`=8Uk7(G-%VOPam{tqEv`h8KQ@Pr=EHW zV8nY#xL z9|cWTDo{w3Qf0M}e%`6VDlok_gOGe0&n;DGKD4QDmdP@GJ|%>29aR|k@oJgw!fXG< zEF(smHB%LG%&I@#lf*Fqbw3v+P<*hNs_@dL_%w>7Y^Exb&Wn}<4mhC5l$s=Kkt(gI z2a3k3ZYw21VOT%N2t@ps1ayJyPVXKH3v?C{v-v*WCRr|-lh zku2z)T)zUE?YYL~gCCyFo%^3t1qSmsxunZd#lS^YtK9pp&nCRP%xp=agxRc5UVx5vikl_ z$_L+as>++cA6h!Ch+JIJa53m-D6i-S**SN3HWG!3~RBv!j~xKb%0o@1SN59)vq+ z0ud|8B$-8p^hSn&V4esGi3RgiRLGKwguuVJiX71-5Qz@yrI0X!c^Uu@%+qC|{Dc_s z`WewHsMruTJ*z9o+-|z7aD)L^E|}9Vd)py3Dju zm}H(|VTlr@+1zmP#TQG+z4zYBY7x_8mXO|m|9z&~Fp@o{?9A8SeDh5fHX&feF%pX$ zh#6_Kw~QS(=&9Vo9u|u8Pb?CC^{Zboe`fiVO>+_$J&P@u)IvcP=}{<5L0FD^@WBWF z```aE1JYbt8t56O3X5h&>^O2PYix$er|2mZ02X=4GXRBR8UxNe1QZHS6@dwi0pO>t zz3iSvIK@r5KvNM-t zO_3eU$wGm?rR5SP*}5?57S18b6+@G%l~!3l_~3)oSdh&xmr$n=mbyHhD`FK4$&4f_ z`BN--v4o5faZ*k`7UL}ehh-&=6i`&Tu+f-ph+0e67D$-o^V!nKiln70ZTt(a_oI(K zT5-h{BW(!6YkJ_2fHzT>TyhC}FmcJecan;Jniqg&N%k4i9C&-i_BgyX+aD!L-Wq=l z9$t1|zWVB`;G-?}i-Vc{g+H zvBwgH+hRXGc-YZFsVh0X_QGirO#KyGF3K$hOhdIG-iD$RE)!?Wke3Ys9PvPJUZkNS zgFTRF%)28>N}Ds;-OEcP>>0nnt5X(Y2M->M(s*d{!`ozeE66A{`aa$}Ly56U2;GC@Iwj(SZ~UBw&92BuJmqFC8{K@P#lZR>5F-ua7#CvO=Ij~qGD)}dn*L~;5@9X3&p z1)nSwkj0jZlQe)hdQAtZJ$??d1;B7VAXJEc)MUKkG7JD&R0L6M7&L6*o@#pIFj3SK z-B$P}){Fw0UoMfLA65{1q^oB0#V)(-lB)SaQ7*W1Z3AJ5ywXZ5;jnbrhLPTnxm?&> zfjeVh&9p;qJ2C+Vkg55{wMUc=F6j1xqa{n^nFRsdu=NfBwt3L0hnON^j*8`@T+qvf z`BVYz^W~~Yyu0co{eoUDNsU_(^8;>2tyl_4Q4DR8NMYuwy&QGaQH%gJJZ1b|VEBM? ztc2rD!_#m1@=!b_gIy12@=2QyRviFv+O&N=&NJc%CHp-;tbZy5vN`n2Dz z{o+pkqHRp2QsUiKd3?^=1;US*y`zs52X(1 zu|+`+_Wa~0KhXh=WT8OZAaK8i9Wr7s-{#&(&YX0!oK4OxMDhD>u}zi`6g&%4-umIC z`>u4N$Qvm49$t0(au!fIyFTW~aa=u+D;M{ISHE)#8}}MU7meu(hbWZ71ddf&AB}r` z#h&XL36R@r_4ZJJULE_Ly5WtL`+f~fWukXfDCfWQvvD)NB^K99uarvp>bH-w!;E$Z zS;Mcqqmmv!R78O$FNk84C-6FH;8vZxfi$cFz<3a&29J>1X*S&jsYAcOI6b=r3 zX2)oomb8f~iX29m+IoA@9x#&Nn5d3$pSICWQHr8G=f(Zl^+y!t6^SxENtb2IF5Ti0 zVe*j?+o)X3FNO^l-w{FfRpo1*TDHoOrpX6zs7N=lh(Lux?@(8NX#b+oUtCl@ji{1hhKC=T)db$h8d~s3LRHCZy zUB6m3r{Zioehm4u|kKQ$JUr;mb&y{!yvI_`s&;bm?f!IfL%( zT-nV*fI{U$G;I>KC=}rAVw+bC`FF@1!YY&y156*J;imEKXfl;uD|US9az``F$>+~n2YzE?N)jfP;mZUvggoTS@z>%Q}I`I45XR@irF6@N@X6ZTR#t9lEg zHxj^pHDSDwA%M(~((J%?rQ7?IHjC`MShZn;5GR`1gh27}8bG;)of}Bm<_nUo4&)P-uprp2#-Y_e{_&4&kp<5?Gejvl zD0QryGqT!Eq~y4Vie0s=khZ8?)KJ+AmTy-CcjG8MX~WKpaEfFVMZ=SA96Wi~#WjDt z31x2=M{F@D5wJz$_rL!=FeSnh_CNjUPXL_l)wVBo{jvZ?`C<;d%;#MiKMp5r)+!dB z;g3S0MsWhXy2ti;C&e}l-bsZt5!|%B=(fRpg@QPaQ&J%-tgr&th_Y`*J0F~kvj+w^ zEGD)F>dryoi2~4Gg@{tp$UHQ+-F6#63Pl1H1%1BCmndBW%sq~3;K`Yb02v`)E%X(!WAEKk%* z#-?IIfrS|@+;T!*BSSz0ZvJ7E;pUf%t3|mDwE+_%fr%AH35RgVXtl&bFr*BT<}!hj z@G}x6GI0r}f3QFk8)$yHSTh3_8etSnMT*7x88EQ{C=DAhQdtf0kuqVFVFM;M(EM^y z9;A@%77B4UEgFBh2EBTRppPkNYP zWhR+?NTXO`q&gT0Pm-k)=Oc^Kg$%KjON0-DrIibp4>QISKVt{q_7h4p%xh!4NR_&w z(jg4Jg8+!-!f+qiag@X=$dkb&!$k&)+>8~`urHr2k8Jt0CSekmldn2O&~&-b2Q#u_ z)WqVp3k7B_DPep?Z>L*lT*D08IOD_El&hI37vn>P1oI?IB2eDg)XG*^j*&%5J7|DH z;Q$fun^6EHhYuL4NeF`!bXF2;ue~-KMImtNPMY&$mS&`Zrptx9gBi9d$xx8qy0U|m zkvMqjhvQ%O-FF{n(Wr9`=SfDxP%sQw83(i3m$5u7z8lP{Fi>V>sM!C(ITtVfelR(qu z0*?wM1T2W_Y=~uY$Z#IR;bI=5xPzHD)&ST!RY$J4WP?+FI@ZRB7DdS?17-GqaX}l| zGF^vLzrS9GDfOgcdi+|klgO2#5Au3 zsmTF)XqZnO0yA|?3Iro{Wl+Er3P=i(6FC$&%7k3xR`eiAl-8l3?#~k?RYIwuKQkjH zVj}RK4}!^}X}VmbN3u|mC5gIlOmLWn2~f^>0A)ZLJ7~bf#s~nLtg{hvC1sjSToBWz^5hp1D^?gM zXQVOW;;C6a4NX0?FN}<2D%kvTNflQZ>5N(Fh+cS;5etTAh)HD$kx3zvBOw+T-slvv z3!W?GqO_BVN#i7bzF_87xj2&s3TH1WECur!f@OvT`B);!scM9BL@>AJ0Zz$4szM+| z5Rw72?8sL{#NsGAW?h?QOenJ$B&*2&&_>7`!jj{n52Fm1){atd+~-Irg948AvXT~k zriEl!7NbW*34V&~vP7)uxxtXLw=55{ zHq59JJd+QV3v(1kMy8Zl*y;YIDod#PAW%qSxao317NbKnV6+xaI<#WSaFdY&Up`_r zn+0bEI}EfKq_VP22H@pJ2`-6X=)nz(jLMZTB?u)&13UV0J{wA+SlQ$WtgCog0Opu2 zuM;=}4KHIQ4Wmec({Eh*NI5bBWU$AfEJWclm>f|zjNcefG7I_SlTSEB%g~(DMjTUP z=*Y;E@vg3>l!h*g5NNtwU}ii{s<3gpB`S)au_+}&@{~GL6^5a_1|VU^1q?KFg#?X- zfh--QcARqrkBI^wVjwdJVUSL48R^l~$q(u>Hs>rJcS$gcgP7HPR*y&pu?*_52###A zFpe}14080(7qU{sj1n;f%8P<#j)pYe(r}gzeV8IJZ{X-MB?~(@407gxC@O$oXn-!7 zrppCV6$*e(FpHTyW0SmXKm+D%4F7olfc7kXGXg&P=%bm%P@oVq$KzuytXeYyWW-4! za-I%6Q$(_j5HzEVIBZX&qbB*V3>Y%clQu|Z;DVkzMKC%B(-c#5LmifaWkJkWjiAIH znVOJis){)mlQjyIrEg7BFh;B!;rR*$2}P9)bjcqhNL~_YL4Xr5nw5|klzcaj*$0O& z#ii>(^~o=q06*c?STxGgm7~nCFrUBtnW%xn{w;vgqnj|!}*pa^(^=P9d1 z2WFDO(4xU3kW}oC5w1cpdlG1Vxgg6^47pnsiqdeQ5KJl|j2Bs;QRxWpf|DpF^a7Ms z!O=6Bga~kLa9OyJ*0zAa=9deqM9Mjsj4B2r_9Ugn76wAbd~hToB#WY&UoMfLA6c-i zT+%DFty~u3a$yX`O$prL%`k%3{^=#r7`sr|%i#K7yj*_^J0YcU`9F~ zIy8~*rlNmf<-){=F*n0mMwQH-7~nHy^G9Nv%4`YoKd^FPHxaL;_?3{wW|1yf<-*oD zd?XWJCR5l#vvSlrO!Ko`m_~BT8q-cD1bTnMN(H~x$yv*!%+qpV_ROg*-jL~b2^N#s z6xlNU=N9)oEf-!*=#%qvD>tntcb=9@>#050CAF1HdJ)gJa$)ZX!?0APqLW=LYq9j5 zO36wkTc;gWK&EII`{}i}S@Hq%x?C88vHr>x4HOFN_$>D)HDf80wN72-%sMH%9g;M{ zu!_uTyeSK;Gl{;)LV7;Xc={#}f5J^%;gdoMDZ>T8kgOT-ye^k! zs5oC(nkX0Fz;W?;<4%8ndXp-0exX36)qeX-&6E2{Nco_@l@0i5K7n)2+-1ru_xmaz z8TdJrvyY0W64#D;vgWA+6;=3Fk_MVoDb=;JD(WxW;QEbg#$|rlGt<6a1D~{xxL7Dv z+fFSMj=1XBp$i4}n&%HsmW#duLS+Qp{PvkL{@ z>W8;hMu@#|vj@1pBoacfUa+z8+>37fXarm0KZAq%(Mb|m%?Cm>zSC{F_0pxDWN6cYRcU^}V3*jI=XT1ql zz4P~wAm7>5n#Xso9i4mereEF6F_-W7_Uurj|4mYI|A(%+sCxU#kW)#Y`s^+jY<6#m zH$2B)FcUtpA_i+b@}52GzI`f3T~=AK)01XP0(z&F#Dk$wu7CH8_s70;;&u=8dSqwO zRBU-$l33SlP+1pMbj(!e+*w)g^^(^(fN!v%MwIK8Z*yNHlR`QDx!t&c<9A!#*6E3T z6kD<7?E>-@3bAv_g>OT$Z< z@3_n_ESAi7OS@Jml`{M)lqK-Rve&6k;nf)4!{_XPrgs2^l+Tz~lSEP`#upV0d@M zsa-aUp|gexc%u-n}-`H}s*4m@%6r|nliwdZ1{%UAX$5Ws0!_BnCfmzBzJhlgTiw2`Q# zW=x8N5QK8Sl}=jo$%DhF>N{3Qek~D7#=Gl2{i~2N`F+RKq&J!}T3o5IifU*_I}e}2 z73H5_+9xw>!lhfho!V!I8bmO6z-y05R4CkDQ&~UCPF_@m1ny3<9QCFI*Ibk_cB5am zLvMfGFO1;!wYo6OMw|92l&Lq4Szta(tSH~~nwsZ+dDQRvDG6QI6xD6x)Kfwf3Tw9* zOBAPzUZqxdv-zLdB3mYi+{i^EL0Txxh)m>|LYEaL}*_2;b-MXZa3uh0Qj#xLgcW5=?wHy9AGL3uGit`DC3pWQU@r|!7D|$zLa)AO-3(VU zPWp7|!FMn}kjdz7x;gaBjQ9AU>c3lO+bvSQ$=}|Y`pJ#297L?ImfYk-{eYD zLU6#!r>xU=OH<-Et8QH?+jZHx$)994y^0ppy?t-)qe?9}xN_~NtpB$BQS4CXqjLf+ z1S)oYJ(tRkUm{17CJE$|hWeBs{v~H+Tq*j~S4K|xZ1FD3teIKXS|zXH5@r3!cv(^2 z@cud9P5X*II8wE8-!F4X@AP$Ly&c{0(I2?SF<+J&{EI}8JlBV<)N}nqRy#Wq+FC*M zKvYtj@MgU~ynX~9=(I{<()uo>Fur;#qWt-)zYAFmD`{M(?jz%7%L12MHYyiHaD5nG znoR0#=_LWkIdT1K!f(=2^i2xaeQc276kSH#StzGq1u!f5uX@T{SrdiN3eC{O02gom z;LvAw@+ECyR1%KM1qwdDDeZx6z)8+xFtAhe(m3_HMGJw3l`vFiPM{GJ7qFFQ8>tuF z=*K9ILPr8w|VL&F#dDRp>I0WXQ@>*Bzi$w($*ma@|Fwt<6gMg z{m5GSz>$C6sr^s)>$S^H{qKG7n~>Qs16)o|u;rU$0eoymyNMlp#T!Q}+a7chl`I;N z=WTiy&)Q*strsV=o^usvd(zlPA{qr7T@*+mj(q(1iO)ooq_TSNK?gL(I%xpBSaPmY zuw3egs#kqqF{3PV^rDRg2O;+CbF{-t-$6`+dm`-ydl}w885?36?&7{}=}c|a9Q>ru zF5fA(Q>~Ea+IY{_|!hoLlJ zC!Os&ux|3_4YBQe32>sa&VBMTGCM}v@% zOfo}a!vH2W(mDVXAP{{jS5ksbW`XBbIH&bF^&T=Rm5?h5H`R=i9Ib?rqR9d+EI!O$ zNLi#D$dJG<{tVi0!ShwOcWF^ z$S~unG^PZ=P=-hap9Nh|Nb3;UP|93bRY+FqKe9h?g8bKFPsO3RVemyp;=7^hr!G zA#YL>Vf@)#dZq9V!~(pLsMawThW^~vL#m8>wL6()*tN~cOiuEK-*i;+Q0lBXcOOB+ z@p0c01A)phS=M}fk%Kr$vzvqeojz#N8G8V~@mTJ)1 zTpa3w^*;^@v%k%}N`w@Hx0k+C<(wkOCT3|?>OK;GV7DB7za5$-h=OFqadtSr|aXJd6Cbj=7K+-$o3z1T%Q>-~vkMqApnY zq$?+y*=)(NRXm4;z^EKb0#$_(aH*7$0?0o|Xu_fqi{(%gazbEq>kxH78)e5H=TR^? z=((tgD{MKGjENJ}H~;YV>_q3}3OhV;nX*Ex69S#9S8)?`>dc$@U@I7ml1qfR7ywv5 zYd#y1*)vAl3P~6jj-K*h|A}>4f{j?MA);Q;!bLfVc7U2b>=AADiq5PbBQFz`2onZr{ zO9pt7Pfp|EEg1#@(@#kyDETGJ&I%J7!67D!QV?l47+_KQ3EoIE3J7JIK3^33%K0gl zyy*HlZf(*17RpN~5Xwkq5);xJHb93P-ae^2u@Qo;@brC*v`^j8Y(`>>`OP15DWDmJ zh^;_C`KFSoE>olC9+F~h^>jnuTza1kEN68AI498VFL9)cfRk-fOXi-tncz^=Z z#FdX%jCxv*B&Q<7L{s7UY%WoOI@<!05Vj-Q6^3TNv=>rl0L#D zldwR^F@XU~5GIK?K>+|RVx5HH0xYRf^UYk?%F1#X^XJs>-6ZQk8Fn6$1R+Lnh)FUg zW0M$jF->DK%&J`h8WRIR7AdS)vjcD;Jd@sV8;lquN(Y%RqJV;aN}$CW9|0jj2%9Wb zTgMuhi{`@Re0;lzc_VA~PEP>$JhHgNJ(~KQQlu;S2{8m&hGM~+k_AtYOB-2((v8_6 zrGzKhMHAs6PQ`#%D(2$8e%38m-C_|>YeFz8EvGI~5nggcsppU(Ls*yN1qBN^y!v5* zjx(Tq6iKm3eZYg{#|kyeca$4e8?M@M3lQtqEOm;Kr&SWhI)r$l{6o0mCV^Hn7cP6` z4FKMPFD>b#Lf8uLIJC+c>cM@dsHpo~cmc_pRdo3kiWhR{dFJXC05vneiHd1^|~zA7JO5cjkqL7LA-u5ro3IrjfZTR!%$Nr386YMWPAW zfss{w?iIz?vPMVh;q~#@EANKjWje##0H3Di=1h*2pK-<++-b>VN2!O2f+5IiFETMu z762&$-e9mk$^O33S4m;(pATp?bD`9@8iCh-vK6JLJ3XVcd)5DhWGR}rKQS|&aGD_bqz&!Fd_1i8!>L!@pfT&`v1 zbr^MkWMIBS$g(WTV;0=1z>>7t=|nJ1T^TE62vg;prl8HRc+ILS1YA3S`=LDXD0kd( z2gY&INhj&nRMMpt;b!<)2fmi(M3c~mK6tIcO9K|G=;A3cT@b6}Fn7MWK}$xOth~v0 zx>h~{TE|?drz@_wg3B~`X)%{IFySmv#&d zKj>lf6O7crL6TFP3MLDZ`+-^1=L#r@33BrjLEIZ79vsF78*E_vkSJ*MBS=>Qj6v{J zTCC7h6nyl7AomV}=N=`|^W@@Y9D{YF}5N~l`JrCNf;Rrq;tVZP*RE_ho&?PMg=4qRwu(KVl7s91&x$I z3Be7}tc8JC>*WaKov@Ln2SA8q^4djPaPcK!D&~R`f5S^t+B5{o)I+gJSTd9Nd>GSX zm&yexVywKW=00=1mKP1ze&d#irITf+2j)_$daYwF`EpVKA1X?RO;C4RM=Z&*7A9ne z1tge|R~V5gDk&7E%Mg>yjAIR?N>i+909@-vno&^>nm~Y`QL&oL znoEMEMflWUBqR!^iYhE8lAxeEZmEZ71sUZfl$i?8A9G1nr7)7b@hd=+j7?%-7qUQ9 zfg}`_%4^9@vKJ;u6%NuClS)unrepCgg5_O+OO-Ys%*7TRmV%MagD}dE3X7)z=N|w} z7>wv8&0ML(W>>6B&_S_>yh$2sLX;^jp>PsXGASWSB~(LEVL}l`DM}eV!Ps12)VNJD zMO@5tKA4MhBEWejsEnoNAr)C5he@$2(2zwmE=@-{EvCv$-II)_jDV!5kRVhk!~>^l zA~m>+n!z0<1|Y6xgGbD@Q?d6?Bn+6t=|NT+%*7jlv`7uM9VZk9eN zQ>u_fZpfJXu^h=_21^3GW}^OBs=`*-ZBI?G8cVtJDPfiq*Z>PVIj3>c@}MN=tkxrn zD_JRDtc3Rl2xPy$-{Gy5Fbus8Pl&vB7-O#Y&CPiO871yt%^J=Nqy|uA7HZ{Gzbn` zP^PpGmR4zPQ~)giJj7V#O>2c^u5Dqb*A z*{u9x!8)(Ro-`KJ(HUX{Qco;}@m7KABvC96`ItDKiMnSIl4X99gBPzrb>htM6ZAfq zCMOwqVk@v9jVJbjp3OkB2ErgkMiqwgl?qXKHNj0jF)VY1#GzuGO(7Lw=wX$#&iI9?gQxez ztC?5c#lcX10k{>+rGWI*C}qh}`GOHXNpAuJ&=?lrpxKi&VN_)9X4hy;ltzY(Ad^RH z^HG}71O|u-VN(illmXKxG3E`UOwS-e_$by~sla?NmsCC4kn^Afd=h}Ppji@8@LP?544#} z@ny9DQEJhajekt(@fqyaV_|}uB2ZCx;z?I*1wd1z&{ja|@(z&||7&8@QL;0V z6+hN&SU6^bEh~*2lxL}s-JDWF3;e8H6XZkzPZmf^(yy&EDyg{U8!?N{teCNwrrp;x zeU_Rq1zu}t*$~ZGX=IHUfEN@zb@Ilx0;c7^a2!*-v^MWjd|5S3N!_!y%d#xXuk;kW zk!H;rx6ewq`v`0WLD&jUbraYMux$l+f&I0aO9@rOD_>UGse78c#*gfsW>uP3cf8F& z5J+Bw1F*q>XYmWTB`7gpN!!e2PVw`?nfDB`1Kw2dT8cMKY$woL3+Qy*P+fx@gWh1s zR^|x$@j?rWxv*t~ci_w;cu7gqXKM?uc6nu}tA|v-*b0E`KrhmHo%`dh`sU>9M{6$Z zEMY^8i-Z?C{J35a498b_vMIu`Z9tJF@S`=C&{m2pi#D<*OW;RqE=}gVji4DN&}J^p zD71x_<;Q6*tdX;Mks&Sf-qgBfXI$0C3@ZX0MJSc4F#V6yT*xyQpXAFq zXF-T6-c}?hpYajPSq;&Y`YgO)(-PN$ChQb0Ri7s5Tftmv$Guzg=#H7XhL%^p>d|Fu zoYZ8(^NOIZ>W&pN)l(aitL$94>zkR1hU;6JrvqQCt(*R>c(YxXuGr~WKAzS*QRZqa zb6~|z&$Lix)m>|5W{zo)-+jf(9iL#$*~v8M^OrC4+p)houJ39eP58&m z_IDrLjDcdU>pYhStBTYNs(=puon zz`Ds_R2g;zw2gULvRDo}YwPSW{bFUjIGGP@lqZ z3Fh**w@w^B;Vr~%w#07wI$+HQe<|zPp08N8|G)7w|Iz!W^+_Fbde7Yd+N$a&_NdtE z&U#$24~~v9MCEW7K(Rct>G-muz*Tx}^na_Ln^J1CDKoV*s~_5;YUkYPg^m?njE(>c zKWB{uT`tyxk3nbxEnV=PdWQdmj8U0JzL!zBzzIO&-k zB&Ty__ft2#p+gZ)w`p^E{+s*m`s51h2k_#^2YAzQij4OcM+18to{ST*h&F{ zhNy}TW2cX5nyj9Jj}{jtv$A86GRd_XaA5`N%Gya_r{Dvj*vc6j->gsA7iK3muetCE zXg+TZ#aWx&KB&JON27B(M3rxNdCkkmAJ#FG*`fE*QBv8mif!(VaQT9q2V<7~Tgu}|rwfqnv0CM}5A+Ip%!oH6&IhNP%c;-qGP7n{QXF%$ zP*B-2`|?*e{@vt02A76pO~M=o)6I@hvU5x?`I_WP68!VMv)&!^ zyzn%jbNBrUxaPTE)_(oa_$~vwP9Ig-ZKaCs9>_HQpm|b`F*gl6@2&5jR*)|~pO6z+ zSN-4Z>Lz`b>D)gvYg{(lu7XajqLW#g1pf2!pI#jKfXJ5XvDV>{%$eh|Wu?jd+uJ7& zAODsFF0;t$NACMII#xflRo%3anNIyPGe&3Ib*tFv=}cK}i$x26Dduv+`{$uk(@3Z9F}uZDp-RxA@sArz^$sO|L1>WG;ScnBf0^#9 z7pcb7s)=VlzZYFq$I8yrtEMs-!r|BRW-GRT6o5g)aa-Kozsqt(3WV&1kq_MY@t*+r zNIlN`^o?%r(LuBQX3CsAYy8(U-i_8Ll|JvcbJN74CvnPStx|*|J+;_^X2}GT= z{ozUU=9@4XHAEbm*JURxW$>SEHuprVJ;pe@(s`wE$B=tiHk)Sin}W zVC)zceA3yQ++p5)Kru3#r)1JrCe9o^X8NeHGe(V@K5F}c2j^2M6(5_+*ONa^r7Mga z`pk|HY}$XA&Zn&JgQnJRPAta0bOurJ+B^OR(=68V)Y-Z=V;NyDUAsT#u7 zU@oZ6&NA8AB(HG zwH-Q&YC1S;Le-tCifkyE`n-P*&X%3L7`e=ByY$aW);SP(0s=pi!19w9%e{MAx9`^H zZ+Z_vX}?+HtL|J)EJ>MLEwx|1J~3}9Ge%e4y_RBsgOVTLTMw6=+%K2w!AtJB)}n*o z9rN4~Yh5mylAbwp$u0S1>Hy9^b=1M<_xs?6F@q28O~gO>bL&PfpIv+_VlRC8 z=VPXiJbt~a7Vomm_tU=qm0qSZm8Z|82v2VrF2hBmIL9xL9?oc=WofN{l znw-F%YgVj`O*Zi>S?0iB6qbmn7!T)x@tp*ZsGySoScY81(Rj zmldGH*1R}DoEMQZ7ugE`PCLJEhl9t=xS@AFk}o_~Zq+{X?5juL+r3`{CX>10{d0PD z=)2EKCtR?Z%+@h21-P$^GdP^a8=oy7wDbX+uX&1j z7$|jby@{E{3gA*WZW`l;BM#|+VGBMRO(GfxS;q{QG8s28HEmbkUIaDM_No2uITJ@f z-K0-0-{H#$83oULVK0U|Uhb?Ozj@uHPa_%!>!y5t`HsUK%m>6W$F)=@6CZ|96|>h= zszHJh~*qZxg zyRA^Z`SsbZ0~OSBZrBQ=6SjY49z2K^`0vwX|j$|ihJ{pik_ zni(NJ+i&wzGGlA$e4p9|pRxVY2f6VJqDWojxx5Y`B(K*ah z+w|?c)OJfBy_N-<3DO5$rPAXsFG|YCtrzGY(2-TGkQJg=^Yr2Gj`j* zv8{xfwIa;jOy2g^ylZMwmB}u$a^>C){(!Rn)fyMA*mGSdFKtYXLw@qB+jtt?t7E?{ zm)s|5M%jKtD);$JWC%QNy}$SGvMj-9(D?S~XCC?D#_y)i@m^H)jy*|jfxbKeD<-DiE6DE|HJlRg>$=5N;iW0SV> z!pQp<@3QRD-BvL>{qiMSJnSIA?>4@LHM%j=zd31x>-u&Y;HcvJpxZvYjH#s=>Gk-d zmtlEA>lBJ2cMpAjJ^M6Fi2+gU1G}$gi~zhKZZAZ#@F=rptP|NvNLDIbl(|^wxr6Sa zd%{-0Jo?2=vudXQq<$WsnKGjKwNo-RQ_GhBdD(znOy6>VW2#J-xJWovlnPo=)wNaU zzI5P(dRLET3P?GRa1I>a${zV zQ^HMjCTIpwnz@ibv=SO^->Uik<(lV@cx6WB`#Wdu_Q05u6yoM!jvjK!1$UpYQ-Te3 zT#e8Ch)a(1nM)|YqO9_a#>@Oww=Tu2e9?5nY^IK=x@(PGFq^I1cSOh=!t}YGqGKp@ z>Z|qMWX(mlU@n#06Pc+M@ghJIKCJ!pU%6nmyt00?grcJ7hW&Qn&qx_x`|^$qSEp1> zY=7tK-juTlGRGtH-#OxJStVN$5t7tw7lyeR zo6C~;N&stg-txkCOT$lt6p)}&>(^zO zWZ4uNXEuetu7;KI%Kh@1pjjl7p$R=>Q!;DGLaN1B8h`|P4`CrCYiAP6qKVnXOIaPW z?q3;n8(#oufw|OHS#Fo!71td*>V+Hn3A1E2li9dO=BxWMrI-u%q;u#7AHy{nY{8T$ zzSrwiB7lDVg>UYUkfWx2_Rwe7I0dfe}GN zu=@(9KetI zGT(%zF`@c zpM!n;3KBL%?l|xl{N=T*BV<#7iXHz~{m_=FRM~FJ zl&x^cdCUDQHMRhq~ z%)Y6yMIu{ZF0AqW^p$=;&3v=kirGxrq5Ea{%XVJ8u6kPe`j?h1IeTrmNJUzmtX+2; zZ|%qbsCnnYOckTTI<`n=yDgWg>{_wq?afFn%W0kczNvZr^x7}($;=v`lR$&%u^M8V zClCP~Q_9JF^X%HsZqL+jp<-ibw$~b&Y-PpPcQr@J-s|QBzIasasHZXwZv)D+z1Pf? zcc|F%j%EmqSSmb9h$7m8Z6y$uK#N>UB&*0^Xq~#&s{mt?Ac=uJA5xN?Vm>kv5DUFw zfK>WO+lVF<5>XP2h(%DCB~^%#p|(aQ0j0}~fsc+5%jZ~<2eYL{lGm8|U@J;H(#%2C z5HtpqPcWu2AW&KLYmh>V*tz;rfxidf00-ii4szP zD1{HXfSrmdP3Clo(%rkF7fZxTu)uG=*@`nG<-(w01I_>#7oU*;q70u#nm#EllAvM@ z$;@kM2BNei3Zz*Kq*7Xr82yZh6%2{_h8y2rGrfT%rJ{@{$n655UvZQHOG^djo2{g3 zQy8f%g&6}@1s7lp=|vV|^1(^U0!f5X7HA-hZ6VC3#mF!2p}?II1(J}VK#=T!3UE1U zF4_t=+w8bw%7jRT=sG57*lM*pFmF<6b+sdboaW?Xf85W@KGZ0xm|)@YN|O@iqOIiH z%@kUOVE|R9EXjHcuV9kK?4}O^$4YW2DkMw7Vl1FZ$P$VTQ-mgr%EuynYKjpHZI3UD zLdkbWc@4r9GX8*Plu#739F14Vmwv_<3AB!_urqSVkRjaWg%W(@fP0AZ-IW!p#U4sN zOdEb95cZG&IKuG#43>1cx}JM`_^vGnr$vubrQm@104ATuJpJ_3`EVm2OXYhCT!2H_ zB}Iub^%*59np9*YHlm4wPAkN#UPFw@LVz90d=G*z9E0TIXYTgpbI2@q>pr50JqlV5 zjJhtGd#||F%ykO%F)j*eC4- zGLeOF$jIk-fAW)`aI+jYyKtdCM!@G;*|*C!UcPmH=fEZ%mY8NE}5;AWu%oaFGBXmK6pv zwxY7*o$rEjhcu_hDHft6%3Milk3RaSV)<~5$pks*ZA67R4L*6x zkG=cG2;iH(w32`R^PlL)mv?!|%vJ<8R!ZHd>|hawuQ#)A7dvpp%#JH}}yld56HxUyh#GAxXFyX>Dxx#KZH2XTG$|XIrHe*M(pk&h~P=*Xyspj>%H)d?<$Y&4yxLkHb#eg$4_jSR{~&0L@flu7>ww1d}_Sj<&qCy%|0`QG;ToJ<^ zU_L60+EPZEh_djr)og{{RNu*PdPysBl-i}V#JtD@-&QA>)Lt}P9tkf_LK1|8nDUJ< zL@8Eyik>^%=(eZ^Dp+3+XCnlk@(nL=7ct*=_`@Imz(basLCc^p#IP z{j|vV1c<&0;e?)i@=3lJKr5%|;{$lT#aDRwVvH(YWO_y!(~lq z%59NYZt)xR$tQVGZbyt5!RHuUd0~M9Gx%rU0_POPLY6`T=TJz+hUEpGI-`sng@>R% zs8OzTXLQB60&b+jtJ8Kj*kA)b1t6ISIp>^nurpi;t=6d%YqumyV)-dM*cJdiES?PG z3ssFYIs6c}r{Wv%$~=*SLGx)VK1U$oJe9OXS&j&lE=uD3cxlh=hYV}TGUGX1 zk=Vf>aKHhtz4jW;m<-Yk)RHCH2^9uf!B!Z5=u`1zk9j`^UN}j`l0yJJX-~SCl!7x- zWK`tmpMS0cts#qI!>3{+b|XVdU`D=}3uhuZ`$&Jm=iYG@T%8)xGfKc$<1nauM+tB+ zEN&yYT9>+^R>07Z@SPM9eHNC)w`*{#45zp>6&<877x?l6MiZ{Fp)^8;qZFNrX`}A`b78p4RJ&6j*h{e<3+d*26 z00+Y;!^BcIv|YYy#c+f=!w3cs9{k1|Z%~EI<8U5~av+0+pEH6#fZoClK$C=x5j!4* zRB(heIWVZL5hS~U=yxC{QL)hc>XV%#6MCYA%=c^sTgfN7NxTaPO2&qa1Ma{7e)iig zzx?t{!WjTDVpSE2()m#^%$tfpDTZPr9YkgVxoFc!pko2TB!qWDsbq@PlbXX^EL=Wc z19>AY(-;Av@F*%GP|#ZGn~P@0Wn!9)_BcSE%4kX?q$U(1=clcNdH~}{crp`VSm*T; zXGX+fFoFq284d}|o+MKg&s2KHT0(>g86_cAj&w^#iQSc*6mCqC*BAjjqY_}K%g-A6 zgRRMLD<*lJu7UYuE0Ja@DLpY1l%q#i{;jBie8uWw%10mxCbPg0FmDnF$s*Xq;6vDC zAO#Buvt~>xqa{R=Eifr6DaXXTPSSXVu8QeGD7+YW{@9AOmx{_TPv0I7}s(R$;LV?p~(? z8BRuF*C6!_W0S=Z#ukPP7hIAznR-Z*FE*)R5);Zvg_~pv34jrzXR5S(NID?pg5|4- zrWk4S$FEpX6+x&dnGq5i;iI7O08#yk(od&iz)@l(C?Nt&W+1|wE17OmPADAuq^OVp zc7h@cQxQ`7nBM9ch@wJKiV`Jn>oi^giMd2vmX;5g&$bd3y0i?742)@9h$f42WlC|m z$V?`JP)uYy>#n2#*kB3>!#tDT-$@%5nfi zCe1^u)mPl zh&?susmN5x`H;N4#a!6HR;sFzKx&AM#%$#TpoveMI58qnFigeDvr&O<=j;kiE~BLi zNk!VPU~@K?RLWH3+_Dw6FR?pITVj(5^RE1Z4?Z~Jh$Gl#&7QQVbQH850I>ZS4_JgszIMRSkUu+*F5A1fM2(T3hB3;8ElG*bo zA>g^%pHgPHbNJzhCmEx(91Ls1!U0s20>WZt_c(iOG0>ZDzM1`29AV*m0_;?>Km(E% z@+!(HniBhFS0(gp_}zEkec6}Jcc;zH)^3u(-C89WB?QOD_W>~oxzURO@SDPc8Zf8IpJlM6|x3w?)R1hiZ4=> zAZLH*$k{!~E>WBZ#c{h~iv^G<+Ah0j*;B{%ur6J?sOSj- z<9M-Mwo}m-!zwYaq`^+z(C?rLdp;g{HDp09+3KA5}LMMg2bug3WPUAxqjUJbdHDFTESU95r zN`%cBxCE9HXmdm{;i6TsX@i=+;f5Pv!KSspl~~m}06GA3{)jrGcVKrub!oHa3?F3(riLiB& z-G&^qV!JU#q>Nh`CzlUsHCtgTGB#^l;kv+lN-*HHHq4nB3Q+pW82Q92FwmjCp`oVP zhDa+XmJ_dJmCE4k0#UdJtcq=<_#XD^LO=}>%SKvuf?uRvC@J$ zL$v??`)f;)vPrYSF+k>A8vw~B0k#6tW`smZY#QH#*P$@!nCPt-pDC8qNgBUGTcVdo zXSAX`u~&9w^CcC?s0H$lkC*;PzyY(95&%yL0Vk?-dM3;uaH@t%SI^e56{>_vB8Mmx zZDI`*ONDhtl_<)ZHvP#bpXgMna>6zX0k~+AV5>cDpIp%*nN!eDbDRgmV)HXKM{h+B zNK-@Ry6dh>$+CZ)mL$i+Sb-9h4pMQ%H6`bIOK+ zIZYWerKf-f77Q8-gN4B+{l$R?9*7!5A)Fo2vSvw$WYQmTR>R&46JfMRzEB2A1oA!O z@K_o~j7=koPK~FeVZ_M6VGJKW9EU`&gqy*Uuvd}>w#61(;I3#VXsX#ogFoPm8J0(*hm4M&5j4BS=(#|u;1DH1Piuu`U?L2da7aoHD^#o~ zFWO5FR$qa;vvd*EP&KM-=sZLxY zvCP*LiyW(=gpyoQS_+0p!v>^=MA0!Z`liib$`~v2DUOTN4l!sMD`a?L?Tf>>@#5A2 zf&8qP#YR9&ZN;aHkvpV*`5{W6a1j02fP+h?E1{d0-&V1Ps*@As7~_lyNPUL%(n0JN zr6V&ACWVKZD+w@8%dtsDXgHM5;s;S?gx-`%USTX;lmty__E2ER1&np71EmY2R5<}I zq|>Eu63pNx6lfTrz^|pYQb2U^Q7p+o93DgEK1#Y3pWr-3%P`6mM2P+0Wffu z8rqocD9bDi%x7CcU1={Zm1jK_1uP&WOJ*sR%o4<#q_N~sU;!@3LLo`6RE$%Jc*BGo zC8SwGfxt<8Dt$yo5c&MR`syo1F-4N(h%!np3eIO+Ap=QujRAnb8jH9vi-{v3lwUg;_$C3@2G4nWi)#mL#JIiUH=ctwh>8=*$~KWdFqn!Kc?74spq?S=CQew@h zU}{l{1|a|E^P^z9n`ptt5GpAA(P+-3GK`1J< z$gN>3svL_8PWZHZBaIQ45(=e-a1<++NHGT`*))-y zFeq4*f~^(7Fj3QRQx0o!1X-qI!8#-}jYOqB=bNoWg_2KK7oxOR5Mq~$faDW;5vOL` z>=qRg3-;4VCuN3QQEVvL5|mB2L@Yc=enOg3fS(W;4gqtSKq~nOv0wmRF^Y^A=^;K8 zSOA`Hwo*VRMJZPb0#4dUvw#U< z6F4DZEWyM|3pIF8Zlu{GF5{D8P01^Iy+z?oCoi`nN=wKG%r{$6(Wp33PC)V*>TP1v zJ2sV3NSBCdSm92-luycJp8%Ut*khnnoVo$yy$?$mKSRn8h9C+?EMY7!z z;diKFG`)e4-oklzC;|+{1deSIm~XZs&15ma0vxp7#v}nt00_${3uxh7QO1@L49JXZFk>In;K74=r3?!{O90@lD%V?@7ytqZ7m_`Z z?497v5qn~MO_#|;!^=No#Lfm3HcUA!Vyz)asd$ZZ}10GYu;Ghg?lw5)5T(lM5 zKCmT6Z`hL>NJ~92bG8m~&oMzZadCAx8+m9Yc>#-~FCkU`8I zId;+@gzW+BHDQZ|6%c`+07HfhVZ$C97#M7FX(e*lN6coIFj4CRAiTC%0~^B@8M;eT zBH2I4*y2R_fI%hpk+Jaw(S}Hq&5{(2z7ug}WqqVzpqT80nHCtxFq_*jaa5XY*Bz6JX~yv!ph}WKo}u*>v*>3 zAqVI0tv!&@eAz0?VicRiupzcyv+)rHFd(Xv(r5d${EfuIg15s8d9#myA_yJ_#x6Sm zmGp z_S-V>WM>vPqfi>`tszBzxDdAXvE7)_s82+&ISU~O%RT)Sx5C#(%V8b7iJm`c)?bSuEN zCSqyvcinXt*NF2Ti~VfuCc#6pGm=JzN2Ta#@nnXwW?v(FPIcS?3<9ZIa?YTSnxom# z?lDlt7St=6P*Uu6##Sg$)hkKEf_1VHlWq677MuZ5ylzD-I#a%Eqhy=3$_rweEMnO| zOOK!nd{GRSfRiIF6ebBeO4=)%6lvDzgFtAgY@L8XbF#)BHtdOx6;){=u%irJ9$t{X zfX#&TybPb&pwDI{Qp9B05RPRqB4Ljb2FGSi%!n?PoU#d$b0?SuhQba`IyjmrrwiCI z2|X90av+OsTu7rdurthv7NV_Zu&}F!X3Rny3CUbsDNPy*l&xbc2x4sqPwo>mv|VQN z7(`?!P_m;I$4jBGk&&&@v^L5Vw?Ve$x-mQ2G(|({<5Z|ZGE7^fh^YW7kH*B_M9PTG zf=Uq~>1=y#|s>JOQ0*{77TNQ@6pfN#Kh zurpjb{(xPTvK6kg<_HK=XyjlS+NR0Km&Q%i(g3I-Yyu&4Owh2Ml(8+QNuNRu;hbmy zP|_smF#%LhG!K4Mm4*~-K_^8#0o@S3hm5dQlwOQ+44GxfMW;+=Nd>dRHBxS{<4`FO z91m8DX;S74l_)VfL>w)I@qtE%EY%m?y157d|-O8u_6HmP7{eLlg_v zMnXxyqSOmNTgO(&5{-|BNXO}sW4?rGB~%r@0ndqhWVaubWS-ikI4E=)6Sl(Ok$Q3y zWIc-E8x0O~q4CiZ;hD8TA9etc3|W9N4%`xb1SUvfQO8V@F)Ydv>rv4&G@(+JF3VGz zYB7bOIw^6~LkBu62x!eDjc!yD_Dv0OE`+{GYvd5nSaCCI%?yS}7bFIU1nP`@;cj8U zb%Do?F@V7Mfc3MLD5#ywD1hN28Miu{FoWjg3;Q#M-WD?@361uEaawp#a-_LBeIl{Y z)2L`gq%K|(Rl39qu*?ELOKpW>;*bHhLK-+^DgfVrr58vN&xwkt&J!1c)dN6Jg<(LL z4kgO~1MiZnjp!=i0z@&(x-=TIPmZ1RP!)b^_ zAz6kZB(*5prITWu2$Bg1GaB4DE(&^nR3}Y>rb#S6Fo?yOVcZNL&=U;)p95?F5HRj` zeLyL^Fe0cIlY5v97(XrL9SgB1=O77UE8g!+kc^-eeud#6LDA?*cTaDCDhGg{Xw7*C`E?xU%Ltbdmgp&qp;slAcKNBSufVX4BqecP^1dL#m zqF4m;q;oOssAN5X2`SAI0u(;Z)fikdFfL8^svA!#g9i`B-_rn;Vl=RYGO2*s4Ybr& zET|*^&FE#!LQr|2BLkpr7%m8-rzjbYXe8^5g|SI#CUcYl{E89{CE&L`FSSYn81abL zc#8r9%z#~nqF^Z2EQ&It1xA=AgO)=<(Jh)@7R*!}n=ldpdVV0)NrAz*)gOTmRSmh4 z@cglrNDtJN(maVOtbJ;1m}e6RkW8hRN@iu~vzSdN7zxp^0HpEe5{3~*gh>XIywIc~ z%^L!y1Sn8YCNr-wjAn#n-9<=8nYaKjA8aM8yihaY^Tmdgsc-?tFe-#oYfEUTnI1H*(^p! z!}OS46KKrqSnJKZ^k|czqY40Rwo;&Q7ThSdP2nVX)ymr+7VLS$j2F=mM9cOV3ff~R zsd;FIp&+NtR#FRd!9uc!kS%s>^5Qi(8zgx1$9@{NP~k!_6m-QXM#99SR$AxwnyhLfo+0Dq7N2HA7CDtu&+1T9buAC~xxEpv3M; zSq7V}*x{hw#2gkh`oPo*5Nz@ z(q$PK3a~ccw@#7?Da|kxWVG2zq99up!IoS0ma}7!&A;r{W7jvkdRc1+W~UNKc`db7 zlWjs6n-aa3r3um?y3JOC!e}vwL9leez>^~l>QFeb!1hfx0kV;bo%E!@YDbIHX{mr_ z7z#H7!r)8^r_p$_`&h5#)S+mfFEiWbCo>l~8TxRZrc}Xgwo*cg zagc!bDV*xyC=*AKdBx9|gOf^hC@ha62m+o9nq?$&*bGBKNSm!BOR@pOK~#>T@(x93 zFYpsG5zanBp@VZeESOisf*QezXgp%6+O^qALy0l_U>`hZi#QF+@jlMq17Is0$AUr~ z3a4yJE#C!`z8Qu>Mix?AVY!pTT8uO3K{yo7{b?Ml;ouSn_?Vt(ft-N|1;m(i!L09= zXvNv6@K{=rq81igVQj(qYRygbWO$@Xy%}Y(@UUOYSF;GjE2mUVxX>^aUp$g+{oLi5Pk!-oO!?!q6>ffG&ww(?jpqWm z-5M(qtarsOg(d+ud{`hC>53i`;Vv;=HM2U*{RQh<^Rw{fnmxTE{WX)C^HBw3WIn^Iq%=_-n(G%z$@-aYY4jV*vIAQLVw9xuRK zOw|7S?_UgA@iae9TPZ&OZIs2AK%1=;Usek{Wh>Z9^?$do8~?5fGTW|u#ZO<(RCZlJ z%C+XL^J_o2GE>JTuQ`p%S3Ra|jTUW1s2zTD&8sIlcjcS>wQQ+f9c&)Z>N0CT_*>0e z=gn3oTfX`AW&O9E9c=#pmf1?ptg71whKaxP$>>+FU+I^dpH{Z=5h1>LFo!hRpwzsB zPo!9N>wrvMO)lx#?^A2u|6|RtD=PQRZJ2Gwuj;menc7()W6cXk*1R&ba<7jNdLeO(Yz0HP^x83(eqKK6+-}SL za_p5CFS^5k@>eHkYTmvuyZBbw?kk2VH3Q~MZ;`|J>D5y@R&)uIsJ?H*215z4z;!j% zcduQs+nW(O3eKpW)}f-&L5=DMHjNCWfu(v%^+Q`%?D%*jAsL)iGoxL3hY+7vW>rx# zXVuJXSKhvnz`NIOY$y$pbu-6RKe0!}R(F~>$?K?mK(V$`m#KSq^mA+U+024!-aH%l z#6x?Zb@15lU!C;QwTt$p`Qq*Q)zwx%v}NUf-_!?+eQ=R07 z`+UNCCvI?U-%bORUft9Y`4XcvUd3hz+e0BX?9IGua#7KZlga=Py<}XNj)M*M8h^*8YtCvu^6Q z*h%v9+45g>%juKt(oknsu1SFN-|W8OqdA=0Z&hz`RC{S52z@;O}KC zA6|&2KCY&=`kWW{tCx9G*U_&PqMtRX_QR{nR{C`z8gA35nSRy_d*iN7-SCFapZ})- zaox{|Y|UHWI+Eqzb{UPTpaD_=|+CWeVK$Nb^-BPz-&&#kSN zc|6g#L(Q<-lLj=}@6L_!O{SeB;C@703s-?Vin}BXN5JtA2|`R!e`-_GeZ z=I|#VdbhJ**!Q@~0euotA>zcZuF3G1_&xf?O`R&bZnng3iI_%+CTyj;w(8Jl z`nTlz*!8aL-(|VlFYc`Q>ft!mAbfJaJ<0~`!VssTtliid-~QpXoFgp08!BWg*S>rD z2V-AG-1>`cyVr`x4t;t@(zsx=`{cgL*E#FNpa1In)wE`KKM`jI-*- zz8NT46v+2z>3z7}swnDArdP*)rmQS$r%DwjyU0rQlOM(qDZlKZo0PA1v?8Jrv8n0Q zj$QXkx>{kfi*D?wB!Ix)tDTU^7Aj1*C_Pz8Yssw0R)$KJExkw0i@!~(*N|MFjry!O zr$@t)FjB!#^3f;pd9(Ms=l7giHF4d(TkpT}P_g~r>gTT9YeS1DTmImSjyheV+1Iac z7yv3ZyVz#ss~u;`CIJp~ad2Ei6W*_~ZJ3W(mUM`VAykw~+42X~esV*E%mvFTGt<9q zvR;3?&YvURvcSFc++j=FK-@$p&Y*ErOWkL%L#;rB4p*>pH}^6 zMS3FB0F*7ZMW*Z0**+VD$igt6i{vL7f9|p{OEUJuaj)V9Kb!dO$SI!@i(|O3$EtPX z-w!FXT?W?898%GVl2KkHg2mIcEAu$?Vvv%bAnAXJ$=w5;>sD5}BD3 zDt36H5wi3^imgnon)La^_qSMb&(KiPC6~9a+~WjDIQ>c@AV)UHcSd50cTN0zO2blJeYOj8*6>9U7C z{q>xg#RZ$)Tb`|`{qo+L7mmnwSu#`Bp>EPAWh?w&*(ygCmJ`Mvvf5eUko^43?q#%* zka=kFGWS~Hm{8!ko8FZZ_}Q&BFCU-nvJ_6VZqo4b6%Q+0;TNGA&4W{Hg~2ScD5)YP zmTz_~Qy?_ndD)|eEV*})TyukL<(}(*Fz!_(FV=bK&XwH&vc1->+~@P_U;X8`Q@^KLS6Xp0|{bSD+j;3B)D3gwx zv07FAYV!624{i{+%)VvI?0f37yUeVa)(j={Wh%v1=zTU{Vvl?b1@IlpI~9nXt5Fon zY0vIT?UvE>>#9#&|C%s4m0*Zz62AST-@iQSVUrc#a?BC|wRN>;zp(d|s)pWyMd6JX z-#MZwHTdrjFL~|T$HYY*w^{0dQYDl`AKS{;lRtw0C(Hh_Bq<{H5@$1bTf#!0{nE$> zk-%85Psb%5`|@A0v{Dv(m@5DA^&^q6WY-msS@$n^0vc{Cu*C3<@1Ohj=x1Qsb-5#0 z#`tE+r%dcil+lQsVcB2bJPy>di>z_PT9@NB7ir(85ye~x#J2KceW5cjEGaFPIH6~X zUlLc#iCG;v{kh$cvrPBZk6ior638l~B{Se)h^ipH~9h@=MK8F zLq%s8dvsXT(f}U#?Aj;4x~)iokYU7_GQbKw);Vm=i^bT#%d!IWEJk}+dW&rn$#HP!We&Bi z^zPWdfWYskeXWrn#d6jrcjSMWkSShfkxK5r^2xRpmWW!#P$+@eRwm6F$3)IWQzY=> z$oq?=Y(=s`{eL=q!dpLI^>>A3?y}tB_YVIXG(*?FK9%yb3Ge>(E&1~`k67!nRGJXU z`2ESRZ_f$b;JOfR7%$6;JHtLOEG~y@hNg*tCS{hYB*bm0O(yjDG{iK{yF-;l@0yke`cWi~Ogg17Pd@A*K zIe$J4e}3+!ck(JRB1jGHyzDQrm3&Tx7u1vnjLc>?T70Ke2}xuDnf0B{?xiXz*Ah9E zvtQhI(#)|={qn(&87>Bhiw~ZTm0gPAG>wu!sHygC`gsrRw)~MB_uDy2w@7~;aVdx! z*m0R(L{ZOrSAtxn*G8;yN3l%?Q)RY9NnFc*v)9~}I-T_iPi-ssfR#^fQl*Gs zj;+)V|98!6L$~kTqt}E7sz$w2x!psV;zqdVy>uW#2X-z~SRo>BA4S+y0N%C~ zY=_{%PPTow`+)wWrhJ z8L<=&R^PKu-SqEFQ&&BWjcxz!z2Z}q<%}QC*!Y%wLu|7of$DqK4&NKpRZm6cL%mmg zrm{S@@ng{4(ayJ|ga}i8_ZoFG=5#c5GsaZiz4q;W*Lb6>T2rUfp4;WTP49`giV5cY zO4V&k$xuRu4y$)PI`iHsPrrWI_UjtY!b3P#y6m;XMosyQS%crlnzB^@j`VS!uTP&o zx)`z~nKjQG+{jR7GjBR^M4$3iR@?4@*@$BQS3kL5Xeegee#}Sx7vFT9Ep9WWc%b^p zeM3VrK?!3lESAk)tK=Wosydxcvsjbcl zt*!2M;mk>=cHHC1+B>(*+0Zq&Ou6uYT{4*wz#fO^C^MP+KfUUy`gbLC+glCTPso-$ zx!Y~$opT+$*b1`p$QL9;^IRG^O+DNrk(xw^tiZGz!ete!d^mdd^K0%p zJ@+Q%)mukj_kQM;m-_#6-^{r$?muDXXp#9%RgVsR*|+CxheDeHckbNz(MKQMY_rYa zrN4ji#TO;y$Rm$@|NW+GBAYGa1pK*#5B*W&*m3T{mq%3=Sb_m{Ap~csp$Oi_&5IX-kGfAT(tQEb7ET?I^>W;xRlmX zUwiE}!%2VzjE=Jptv&JOExcrbMmqxQCJd{7Ki4`eQf1!$i7`i`CY%{dQ|U3}{T#MwR@uC>;Ana1adlJw7xxNC#pj77_3KK%JV zWF<Sm0NkP6mK z`q+j7hBps>mzheRL{@;R-21cSR>*9ZrCC3ySpVWAPpe}Xz4BciiRIWwk!OMmt5H7q z=Gr?~&CHt6h^)N*%)07Jd-l)NpBGuP&*r?8j$YC>q7MQUJ3XbN3jr>-dv{Vcb3w11 zKxRxE^xJ_Kf%*1(H)5<06d|*G@RECScxS@QF`VVQ;N^pKdLb!Fn2O!r@>1CxnzG6- zD=PoibFmK3?fTob|HNpWDCPll6fEJ)A!`cC*-EBfV&x86Y=49Ai3?R3F@MwEBAQ6_M>ZS-Z(Scb=gcs$GXX%mMyVUmyQ2)QSHoU zN8Izs7dNn{MB7RbiQVv0h8ao$@wpA7Z=PNI@jr6g6U#f+P5r8@@20&rz51+KmVo13^1!u$w{sccG4(w43}g$Nl$<8D4=ZbsNXVE_wA(c-a-j;0#eEy%MUomHL#uclV<&ZoJ}) zn=;?ubK#1Imbz$IAP@5;s%*tW`D13t3isvP50?&1Ak03AVAE~cnOou-(*v;T2_u6W}{?P$x z3!g2seqZJsnLQc2^9LDoG1W0JFNc``Ub4l*+S6$vAyB^YpUN|TYS=@${dxjp`yjjSTp56gbA#VYAj;)x8)-Y)CpPu=`UZ*^}b2xM|!H)@`F_2&W$P}|Y_S8N7 z^_^^F(OwpijG$?S(XO+){6)a~cGoeJlt;e2DfA62N71Ly@s;#?LA?exxAvl2QsqT< znyH%s7wxnpU5}R=ct6A5;nOy{DSw9M#H82?kzLv?f_Fj+v$E$$-j^Ih%@Ob;f;bh- zlv}4}m^?rJ<;_oib-TM2GK$^oIPj3pmECnJs>!T(Yu~F$yTMG0nJ$#nS9 zpTjY{mg&`DafY|KzYcR|GmYpHfpl9D862I9Khyso#b<8AFf%hEw_zBzxnFa?-{zKE zvYA`tmU2nuGIN{T+~s~Jilk7<{hn(VO64A*$)kJBoFL#yRTuF<`Zuyj`TC3bs%^T40N1~dOv z=JH3E^d3L>$RN33VDW)1ENwA#A8@@gNG;bl!%wR>T9HnBc#^&px-V>5@^`l6OGw+9 z5w))e*|Ed7pS^qV@9taqKZi6a8E3Y*NJX?Khv5vL;D5Q?Wpb-%CT)P>@88agS4#fU zQ<7Tr%)PGQPr=*o9(i8+XU`jW;g_OR@2%n6qvQtvoEe!ee`!uRn{fGY$z0#)Gued}hx08tn(>4w+r48l_wQK(!zZ_BbD)v$S%?kyD+WIim z_sHd*TYi-jJ8$S8+B6h@@&!w*&1Tii-4BEnvG*-QwUp2N)va%&TW<-6q+mFv-S--H zW;R?o9hRjU&e4n5Z!8(VqWWT`TW)dU!tn3ylAjUMCU>b)6Cz9f!4F+O^~o13ORU}a z{JY_mM+C>SOU@&kZgw>BaQp)Z>XoyZZ|D0&b)|m6Oj@l=?r)HLqqX@iSSI4w-BhTk zbEi)2-W%6=kg)pi56TNN=!y%RR^=y&;E)G&Uc&PC{fZAFzoD#sMNf(f+4}_kD~%fg zM99bd`fqo+f!*I&Q2oWm09C+?lJ)#qe$f!=k%{wgzSV==X6(o_3k@XHUszLzuO%pbm+{VkDQ zqw(-ePT9$6x#`{SzkT4YPxV*E{*@`WHm_|~UA;pW7oOfV?qoteccynz29yreIln*q zM_Gc=_l4H2A71Yk9|hhD_rm-iFFf=7dHq*S<9~#j-2ITROWq45($^RbquYZV z{v7FR6uU1MeNjU?N%6CuZ1Vq({jcBIGv@DS_fhP>#rna%l3zOJo6zGg3)}AP?--ZL z1MAt*XI)ELMGK7sRqy%#v-p1F%nRnX7xONCFI-T1#HD_JwBdZn`GL>=!H-3>qcCTn zi(I3B^XYZw6?a&$d51ISDjvI}jdR!hF-eEyCZE@7Gfrr~yY(_Ry2*2@x3593-Oq}=>n6S}D6t%Wh%?Rl)bu@4 z$CvkY!9%T&BHqb=eUq)WQZBOIPF)yIp?v)@c}^cdDO%iIq!e?Doe2 zpuasmhJO8}=b2*kb2~xt!1uk4=I#WlHmR?-YTxNy;Y{=kr*nS}C!Qv2?Zg7obqrZ{ z-(+xKhIljvHLA9wax)}{wV(w zYZy&^`uNQ}zGC{9%HZp?3znfkji z$CS?)j_gBPg}&7i^etbd*r$6TXD7iaoQK^CQEwuXQUZNnWkkR8=HFr{F(_zz-IRTl z5ca*WN7at}y5NVq>g(@MZifC$Sl-F;aQ`*Te^EVCC+}2<8W)=L9KUP^IGr56dA|i8 z(3kXKS^Mkl+w(-?@`n%80$)RVTAvvSEudoOlRv(`8RaSA__r@Vq)cxEucJEi+38cX zt;pZs#e<5IiuGcY+eT5sDZ@f<13Ha4>?`&>r|oM*CN=VqxgKA6(yFVy4*yu7P@8Y& zue~!!AbLF+yZGeBD|Hj-(<@aDUAZ1)HFwQr>} z2#xPb&JSGF$IV^-uDns3w%~EfTAS|ZaDI$=8`v7lG@#_EKduoO#=V?=yXT<2O=kZn ze+Cu%Pja2_2cpe!A+#uV^Wy8?f5}gPkHOlJ&p%v`l=(dvt#+FJa)y&N`XnkTq_;JX zySwwpD@og4V#Gve-D1e728VT$id_qs`MAhCQ--2jtwDO(vF;SA(0s30Am=%kGoLDK zjO*g(`=Kg|k=G@^9e>X9ordE+p7Wb3QHC{gyM76J@Tg0z`KLw**in5=d>OK-tEe@41Y-*rO_faozIosUq6?+?4c93y~y{c`|N78`Piw<}MzS|(+7dg0Y+HJ~PnKWaRT|ADKK$r*b^X1<+YQB~ zX-j3(_yaC_>!=;Q$hA1_NAV{CdYsdFl6vb|?7}lYExxn0Teht&7lzA4oZhtyJ#zEC z^l1`O%uC6?eEQM%{pQkL_s!sMj_g!HPUBQ*w^D)|N^CmF5{E^ z*|pcHN>Qu*3+*i!kE<`80`GQj_5S{Pdm`9kJghBRFFmGeOH$)mR~z?DLKDveTpByp z$XK73!z@E(UeDeJ!IuWrjg^WvjOiD%Mrq{n4iseHL(zb_#B_5_b8jnBm zYnN%Zkz^;vlbE)kHCG`SOT%aC^JRz3CfPQV+l*4*4?GVBz(+0Wu?13iCPTe;XG6U9 zNh%JR;?)51Yy%zWMJ$vGiH3sk>ZqBRS$)>x8v|rE zTkU-&#ZQlx!!5-mCp-u}^2z+-Gf*csAAUl$7fn<)crMwt!>GEta&s@IwawEI-+bLc z(;qL^o>Q;L;=Kkj;|mM`_|qJvWb_eA6edb;yO?T_tvQUcX~MUZ`bdf7M(;;6VBmU>WsSBW}lj34PP@n=6`maq{}LUwAEGp zedpqzG=gRP1K6gOM!qkABgrq!9N&BQf&@oleV|;vF066BFt+>@SuEX#GWJ%OtA(c6 z+vXIjQ)w}TJPb`{BHP4VT9SN3K!@rUVDRlmY{odgz}le$7-SP~mbxjxJXo7*F`+#2 zgeFSNK1Cy8!2x3SM`9eIMsdpf6jDpzY;;kHgRX?TSe}{7oYqyOGj37LtB`=e$8AZKul;ha8YT6MV+O#W% ztL5Go3lfybjSBz>$&p}S8AD>OKxj}3);`?6Mz7~`1C32USCTcv-ofGU-mO3B))hgW zRY~-6_Q7P~Yl~_hVh2}@aZX!J2d&h2(x*)5N@YHwvKA@SIU##z#oy(X3d>q-+otjv zmpJXo>))RK9Nsl-_uliPAbSCfWmB6zezSg7Bt%Go{;Al7AhWp>wVlp zl}8m1!`;C}-?BeMELaN#^nC64?0B2l4q>c{Z9s`uvw5%LINvTlzWUkZ5UXP-Jz`CN zQPRJctW8=LfB}mh&spH2Gxy#8WX;^cscb%UzhFBa8P%;w{qnXYNKGsEli>qL-OFzt zggpCGAcq&=y>Dal_?_?V%Lq+z56RDdy&I0dnvGu@F+plyYtri|+okXBZT5EjJNyq2l;b@qm{@6c$wbQ!a%)ZkILza=c~>h^5>?zy8`LpowX8S~+R?(s@u z$di2&ZN>9USI+Dy-a4`DO}V@5DyY}|9p#XTd)lEW3I-;V3WUTB1$pB(K0bhfD+;{g zTn0}N#+A~__1u!l7^a-Y>?G8vV2J1oX(P&kA);h;F!@TitnZUKlQT*9F2vX4zb7-d zdg-w{rKlypiW`llo`8i-uL|aA8*vB0dRd5g8PEK1c2o&J;s)p-*O>(dYhNv3) zp16)yhZ(4Pk5J35yql=fGbFdD4TUg`P1@dHpiWc;uCep~{{Gp=y?pu7R_8lRA@LN4h%akd{To$&rsfNkRZ+9fq z$CClw`NBStg>&tFOp~B)UG0lmy0u(I4XU>*#?gyVwAHVo{zt1(~ZQ1?#$5-NxEA zh6}~v8@RkSC2*`_81M~{oGcPV;IS>n`5bJFPYo@J0Tm@X2D!An%=kS*8SZ&wrs*BN z&hJ>B6oia#XPwaLApdHI%=GX+OLDxtnGe&6ukvmq1hDn-NoDFN znn;oz^1~y(FN~)a`Q>S?yK|vK%ahPyrZ}Ice}8@c@8U&v4OLOZe~zAw(Y=_xR`b~F z5A~uMV>x@&UO#JbN_&Bg-+(d#xOZyArZ?OeOlIOr0>Nq01^ZdIP7`zUxV886etvy7 zJe|uDa1Gb_;pXqTi-ej!@TkA94Od)0=oyAjW~H{>#^fO+uKxsx>3ID8mAl5{=pi0c zG=1afLfnjnt@klh7(C^Oj^N;`x4lUJ?o47Ivd-X% zC2BI&8uIbrggmCj)q#tL65n7!HguTA{Jm>TiIq^2qtU0&2mGk$1t zXy{L5b!dJY@=QlJA(%ptc!Q(X4;^a=M6ByE7|y*sf3fU>-@pQ=Xj*P|myH9v-FzPl zQ!eVzRq&ktrAwRn=j;1+>~=8Q(yH$w988qAy@-nhvBruzV}PFwG#@8UPec~ve}(#^vp*?R680|b!l^~VMIB|9@> zYpNK5`eHN?Fh0BI1K!Q&jfE+G@TXVCX8CUfD}hSY!^D3bk|3;l;~zo?In|9~yr|#l zKVB^fi0i+K8>kk0dN{2Go9-3;dr&%#c6mr(1{NYZ}Xo~*EhMfD5kJQ zPnHim}0i~}!E2k>lstYomqqWchx1M6~|3Okt4&8Se= z(GkUee}6t}U}*wwy*?S)vBM#&I_UoIbRy=5o7Q48Y&lDK%Ew=R>*%W;b^ZaPs2 z^N7>t;IQ1NNONQaXL6)12UkGDzhCmmn^`3Hk+W) znlLc>8l&muk}QvVd!ShXuR=OVC`QSM+@yN=rya|%h~}NWAjzw@oW{$k!Xu<>?F+ai zR>*^sEs!32ckL}_p=oYSW>J}|pe%Mka!eE?2W|J37XiOhKLk=P7$$SoG1DY+vHk;f zxQo8q@u8=>UrG&X^l@8t{fx<6b5wvc)H+X7&rGxnf%E?l`a|>HGKeM?Q+J9~)^}%< z$0fY~Pj->0$H?1nbXc4=YFStJd*x2qhM62`p3-Dpu?(-yiTLCkpL}!OUOWAf4D5VA z^|IaVziw7$OUF~aBEQ_UtAJ4WaW|osZxf3| z7Z^(9p4PGVr&3DD;Rf2fnY2?32pteJzp^ibU{rjww-vEtMRG|eIDIZ~4<_^&edx1F zYoCc@VDyaHk&G3QP-me<3B}5Z*pdcPJ`7dJ^3|P{6?$$cu++}Z64uxdHR#t!Ldn-M zDA2$bQ(2v{6f&yMC6y;RR>5`T5iVlwvvdTTUwO1tqQel)!K$$I{^E?gV@mNQY@V<7 z7-&STi{9eo1IakjS-K$kv8dh1S|{zSLj+_(HR`Y8sw)Pvx!wq^i7=YH$M(jOx$ zAthU6b9)u53|4HC&yJ9CMCP0Rv{AIMtFC0oeqh&?lk<0lZ?gf^Vr%qG^r+9qW;c@= z5|1hHMSjT@Y0bGfCTT(+i^mnHbqsCC)>^jq;!L$D;lObq^^qQL9M5Ba7)iWu3@BJC zff%_+y}Tbt{HRjA$j3w}6v-VLJ$oI@%6_kg2@=?6ZfXG9s=-QzgBi_Qa?;ouT$39ET3~(Cl_?8 zP}x;ekjy-`Zc&F;%M6SK(2(j+C#w{0kmrp8Vre{9^znW^!~Q%Ry-Z_HahV6p$;Mb@ z`VZ75ai`{n_syO+9BlS8lToi_-7@7o;P4)VW!Q@64sdvxq4EGh)VMtv4Y3ok1vbd_DRHo^^F&b}F882Rxb`KlD@-{=DK2TCgPhNw zUnXnFl;E@+k*LnYhx(jka6n?tYq+z776;5IA^|@Xk?`W0YSCll$Pmz;%|54q0<7Zd ztLLwDIalK%2Z3_Pz*nOwG(MmANYR*eLr)0_!Bs=jkTFcz;tfeGBS(r6LPxuZ>)B3% zbNWJ!5efJ-^8`q4o+$Mw9*2ZX4SjiXIL(V5MnV?7;p;o-Dj1ej73@6NK0=?kQQGMi zqjm87rv2v%L)rL&3uT17RXc_S%M2QV$A5EpyK?@%GNXCykiqPbx!%x<#NAq{EvGc| zm=0teTW%hNfYkV~Fa#8G>|q)#wMfZ^Bkm&vfKq19FQ69k#IDmsX}mj(H&CNP<~UTY zH{5fheS7k){|_dY3`ft=hfFRh+8o0h6LvW=2(g-Z#~?=rbGSE?x{JD4Zn34Fzg`v7 zXWtPL8c*a704KMf*vC^U7H|+TVOXV~qn%g?uG=gcxy}EHn@$Am&-Af&!`lqweHj33 z4Ceqh363E55*HY=4wc81oMUBa1BCNga1lCc#RnH-tI0o$nBh0K$618lyI8hh@KC%b zvSDc^X}F^iIm4<~{~uVv=nR zu=BLnM})j?p);ATu*tL7QVw^{s561MI6E60!oGk>5nV3ndL%2gmlxzh!q4IhAUsA; z`mkHxtZ}8d4%<>ZkeP?VopPOdGm>;L+ zd!Xp}8Io8!bDCv{{S;rjk$!Hf8nu&)MVz>PxGyeGrG;{&&f8;1sT*4Ug5Cgw0IKVN zX#1|T)CFnzm<~&?W|xILCN!+v*m$>rJw=>r5+9nv<2uBQ8!GY~Yug?~rQ_md7nN5e zZuGg-kHM7FA>MU0Df5Z>m^e9IGiHM(FBNR6{|B(v3k`LnsNQV5mQ7SF3%fr=D((Ng=q#|de$otd~aU8 z#ES$nDk0;a$MLhz7<3;x>>xxD{ds7Fjxkytd44M;rv1eQagrLvefKYsQ9m1AOtoZh_AQJ_q`86c@AdZ0kqUwLzpu{jP=ab zzctZP<(+boH=w-rlFZ^I18zyX=3Ri0c+^H)PDSK=5af}S&DpT!(hlJ6Q&p&5IQqGs z*#bCUdpn(-H&JEVS zW=%Bs1%kIrY9Z3@Z=|S!DMu^mgm|=y2K@;=Dqp)g+jJ(Eq-HXdL&m_>Y2NVMVm()g zC2i^`jXgETD`r)G!Gpyr034T-v_wys?K+!`p=7Nr+g`+_Ybg{RhyJ^p|FKA&b-C+q z|M_c1yK9$QKN_sMdUVOEC|p~OQn>zf2_kIF6S3K}d|{TSZvQ0C3`VLVLmo2G;8sEO z5^DyB5*;9#^GO#u6=~$bPC4?mwNFDs{XGHw<|CN=+S37#st3S4qi!bl+$10&UcHu^ zfHL+fKsiTAFxX$k#BY*m!g8V~Wc6ycP0(2!DaA%slihSs|Ht0LWRzlk5T94o9-3et zW>PML`s_{xYgh`>OL7jK5)6n)0wSdeOO}lfgYb)VOZ5s1nF!p^ny@`MVc;#dHYIK0% zz+6>#%|83pq>NJVmydhcxAWU0jlq$$9wOk<>x!R_23V9y7?XnLrRgf*1+)soM9jdqt9CW(9{H0HZ6fBH8U&QYfEAo0Y)FJsk zZyNBSly4TuK#(_`;_2YmUnXu-!3X&D&@QDLV1E^qYMF2=syvA!sOxX}+nF)xhJkK$ zdd}tCqyag#9+`rB5exE9mJjm0(-weOHajV%r@4-f{OE!UiblR)Vu}DaIw3YMFMiR+lGcOm=;m(`l zT34-B-z{5UG8sx;#y+k!X}fT@V-=24LDV56O|vT&nYH|_$@n0ok+Lrv!`vZ>S9Us6 z`scYY4%pw!eU8^$n1sHLwvNO*m@a3oi+w66cq2J#U@VJ=_1unt?3N?AZ{V_(M@#A; zF>ffP2-Rz#|HvxY?gdY_& z$Fb)zKab;J?e0rWKW|0h@S8Gl_!utyG>v1Ej6pl{z19u;nEjQ+&m#B|rT@vQaPAM? zlDO`v$Qj3H$(ji~`goMPOr{7fs$o1%jl4AsehEnN42Z|c%`Z^Xpd4^pLn7DM3)??M zW3W37d>DuYOX}wh1zb5DgHbg!hK}rt|LE(SIkTn{{atSr!?f~)s6ZIZy0UH*ZlpI) zYavJ3zdMeMk||E{RIZYmu4LG`C+cldf_e(N__Ck8NVd38Dg7x7EBW*9{uWfmu6tyd zRS|-mUGPOr=M`%JiLOG}S4o1e`ia9yC}ZQOQDEpF_A@SB9Kwjs6AJhyXT@*=he)`| zW#UO3^)?>~E~jAutWA)QRh>v-0jiy&c_FT6;rECf0m0;E9LYlCHp%MqiO6PW@Yp!DwD znOv*d810L9C(CH|ULl7D2h;=h!>HqUzGr%=Vcfx38_oa)LocZqkK4y|`8)hG-*fOP ziT9aR$Hw{s64v*#D)$Iwd7EQ)3|GqpJmK2v&(b7<5SSZ+<95B++78|L7o$%3iB0^J_Qp|e2S@$Y zTTP}*$WC*p3%`YZm4f;BYEzJA2LW2W`YXc31T?y?M6Q+SzZ0G#Ih}JNBbDY7{C10x z(NRQ2{H|RlLDIA>=^1?~*u^fgW>)+@MoGC7ifk`RP_+E#Tl>sJhM6^J5NB?T)bPvM zE+tOaDmJ#Egx3w0{Sp4TX>}oa&k3W-z9sOVf7ey~N0iUTuuv(z!q+ z*<2|GflR%~wYke_ai#WcR=;Td0D;FV7K&lf!+O?XxlyB?HY zQF!224>p?nJU?ArlwE3{M@cZ}(Ktq55|B1fSyeH(vG{5=9CMb9noul=Arg(u=g|f= zP@hTbCL}9^x4^R6p@%F;OaGQ{r_h%g$-v`8qF@EinR(=!i!Q(*^k|I4ME+cxuotmW zxOvu1UDtA`%P}p$HTvD|gkLQXu2jlq?$-#iMavU&tcv*acx%m=fqC;gW4i55T zkHrBG*M1|w*Zl7c&AUw{a8&+u{vOKHq{1}Fw@@TXec(bdP**WwkD$(`2p6R{34|y; zBHGw+55C}6)brix6G&&6{S+#s98d~c(8I0?w%vRqoqlB_!Iwg5Odjsb9C-4*Z+SoB z!KD&`%j1U9O=+8i4+!xaf{g5Md2UC z$D2OCMA>>A%dA&-?#PIga=g!F8L{MLT3Z>@$}9@Y=q0mh4PsBVtbuAqFO@(*6ILdB zqbRFaJNt7b5A|kjFJ*_hgtiD4{0hDyBPtl^qp-=WtzxWU7Z(+Jtq?o0@&^!wH(+7# z%$H*C%i@0&MKUo|4%a**#4Z;|l?!7B+ekW_OE73vsTSW}x!A~c%I`rcb<1=FX2t6# z@~KU@bZDmPX|jZC z-tNTn6G?Xf^gDq5-hByhn^PwY{(ZFbfa>cmc_#-V>zQx#yQ-URplZ7>UH%~U)sJle zmZQ4p>H2rRKn?bZA~K9pXU!rLV_o@@^=e?6k2UM+BB!cpZXZ`IOGhv`p~6I=2vmkB zH=tD!Y`7?*;h?o{#5V5VXl?Zj%j}g)C#ey2<+PEc{pp0;4g#O&kS>jDysF?N|CbSA z%Gco?0zz1_Zy2eenkwe zS`@vUZ^G2n4)rOG4cB0unFmqOKVgr%uE$ny*wFy$_fUMY2|1$pQv=SVeO`q@d4{cL zeyiT7JYl$D`heYbUY-agG-5grEzeLmb8qP+f`G472$m=gR{{_yR!}SZq5yhek(7K7 zNVJVeebBQJz|mnT$1E@M(Lk*cCGi|cX~B8|7tTUqM5^%y`SI}O#qXU_YW$mcW8Qkz(CN4$11jCG2b z0DYeYTj<1+|S3#$A~FD<(p~={ckK6>6;I@Zz+%Ww&A)fw#_$Jd9KfHgy z5L0fn>S_c@BN9Nh*gmQ&=#%uXl)f?N|K55)e$WFukdnR0?kO@^#%doZkqD76V|?Lb z?S1X`Hy&2iJ%fjAfm?%QB8&@5M&T`Z=WYdApk4To7 zni{VMLP`lts&wB%0YYAS!+QRVM4x_#`~7*9com-a`UM-V`srK2Q0PGd!QHFz3<5qm z7-2YGMb0)k-qN9q5>R+pe1fHXpe*=Fb(H;t*Kt!U$;AA&wq2>j|dNnU*Fd1J^@4{MU&6s&#h(Rg^ ziQ*#;BXW`qjGFvXSm7Z?Ud0e{TZ$nSsUA>IZs|?(Af}G#U78wDAD(yV^%*af^m|_u zzrlL`IF}|?FS%#1;p%Z&_FkKXM*@DegatSVIQ77?rvZ@t+3F%e3ZPTodkJD-DHV|Y zB>;$vA)Rwc4791`Pb7#xNn}AU6d@t=kk8O7Za@Q*9Am^nB6kgy)y}bqEMthnH{W*V*B}IQix=$n1XMgavL2SZJ3k6sGV`(E>BwZcuCV&6!kfSo zg6yv@Vf{~JO4?qlsYSUboO-md=*c=Wv!nHf`ZD?I5H$@6Yf0#8q%#xn6jofZGWdcs z{q-djf}Mvd#31Hl=7v|;ELyaToK1kZm|;H&Ozd9|oMc$>Q(%s!c*Dh_ zpAUVNeOP`l?{T5UmXa`>Ru7q7w*?9KF3vAjH1fK)p66Mw_$fte$}Z)ypyzwoBpM zkZ^FT3LP}pV6HYBGY6$4Oy`l+UkS3OMAap_*Pv^?KbKzs5`N}s;}04=PC z#b~FJ&O^k^#}RdWu~_%yFohr;pED$*Ai+4hI><%YCv*j=*mB3T+sCMYEHaNc9>-*LSofI)9O5;3rrVAjWS2jReIMx z^T14Y`5Y`5j%ys>s42 z0bBx{XWE*1e8Q&&f-hX_Z=1>elPg3WQYYCT-5N&COWR8L0DRl$vLsG`i zIjK}B71K=IP1gWLQs+38*1cjL6<&Y}2qqUfqq|^!MH+>=-JWqe2G?6ZeXmf^c zdmRGq2)0X)qI;x~A~%NKAFR-&F~_Fy=DK5p+CWXCo)m|if)y7Ay8Y_Wiykr$y#4m; z{ueHC0t{QgkYZYPZG>A|xB>CVp}o|7^#T9KB8|55dUGnMRW7ibj++6n%!MG#jtBk}8C8 zm>gK|ZcxqV)h&jEoL3MqlaYNNO@(2Krja=rIJg16B3pJ_w@W9EG2G+X$_9EVE|?`p zk6qeaMrF~je;5mTVLqJle#fn46J>FDsz@KB7XO(*D$$WF_ z#$0|^x=Mu~2$^>*Iub|OF&PYdY3%W+g2s^ElaY`-p%`J+LcIr?oz#`Q;l}h6QoX*&80b(E><(YL+im-)G~g=n%IMgmwUFI&?RZb6cvb`6FutO`9V@>~Fh@m(DtBhxVv{-v`Y%bWMw#ODXlW&#l_cU^7HEa2BA zB9J;rLs7p`4Y~xr`A{}E_n_6yta7{SnC0h?yy|kWadC}UA#KksPj&^!S|{GxC7dSl z%;Olqa9$igHZ(7mt85;c3n+Y78M^VVa+SYngUV|R-|qy@5{vP@o>?U7A=!ladq&6&)*iXz_$Q@lYPGIetIoIp2Y_VwcTztd!K3)Z+>fqi!%SBa=yRons z@l^ETa61=t?e0YdtLW{j4aC#4c}cx7$ z2{X4}?6enD4p-1b@^6&ADT_D65LYB{|25` z^PNpj92UCu5sr^lE}tyqQ_ljV)UE6P`_CYS<&PZI`=&+lh)I!L2U|7?|Ke7t*)YM5 zi}8b(5#GcokIwXfu*<^3#^_yCmQ;%$ex1tOy_Ik?>2RW8DiRjg6!^8%+sepB5*~o= zj1$6J{}Nk>^dN+q>6W7 zO|-V-q>3KTy~fTDPNRmT4rj`IK_3N8t(d&)EEugX?j_mntY%0$$-7MW)8(U{S6oM}mce1Q|7|pH18Fzp3s+ zE#fm3VwL9|Pf0>tNv7fmMuwgU%Iy1EZ5hgh97s`~-PE>lx{*OKtMzdgK(vn$H?(re zCb(zALRnk*uOPqqcxaQjFuHJh{^p(qRc~B9qxdUK0Vny)M8g5-2;oZ5;E?LSd~X9- zsBbvTf{HGMoh{u5w%V{HGA5k8O=zOAQ1w_OmbKEEZbm7m6}s40v4jaMzZq4rT>fH| z*f1}{>ECt0^~`X|Ua;#7sT%q>2N#cz%#wW;g;ic1usayzzAnmlnjUh_&NvtTmt$ri zsR%5xr>`8|%s8qaPF3GVx;!v|y&Pq~qs_6i_6Efip4Lt`XqAvG?oPVe0Vb1qJ{#+l zqNUs054N@*(E9^ zx9`HuU`>QJZ>d3*O)mm~Y9kii0qe~E zJx6_0SSJwmM$xq#j#6QHauX^uZp38%1qXt~e@gOXE@IyroxZZ;zj_3T!ZuTykqwa9 zW^=)x@gd?@Z`Rrv$LZ%S3+zTQ@bTL|k-otH${;#?AG+Ihp_!&im8isE&yE#cUsD`Y zIRXO&Rw@zefk*w;D>^dROT|K19kDF%^)NYVm0fu|*zoL8?b?{uX4Yk;B4TNR^uYXH zOqb0_Fx3ZT_>0jZ%S2zq+|*2(%rUY`tk@__J@b|EPQ5`fk|;)HYErq%ZD#kLjonZT z6k0%^-RQ!efOlOYR=b8sb^k8BRiL>d4gSGzvY9BO#a;$a;BW90I9c%D2R=Yqk@1Wd zcyob0*0MAf6W)RCCT%wr-Vh2~1k4f9T5`>J)bh7q%k(JRziyUy2(|}o%Z2MEd8Ug& zzc9#+!UcFjqj3&Mg{IoK{T6L(yH>mm#a(&HV7XC}K)%0=(c6)RYzGB7(u0J#6=zI@ zrsNy$hhtF;3qzr%fiB{9BVnok&q)J@;8eaS$2`T39Z-9CGkJ7I1CY)6yo=T#wfs^ie*53O(R zqaRpVuy_`$=rA?Y0@;gwDS~D-ivn-5BV`kQP=&~|2HQ~qvJWJeHJkc|N3;CpXBSS^ z>jge3nbur(xM(;30I)atLKp2-dN)F1EQ`nkd8o{j?NieDmYDq&vf}^b31ts|dHx^FH4&!_BwE+l)OJIE@ z)$6A^b1Z93@Gi?RVuH9t1tBmCC!fFmVMG6{R#>;Pv)Oq{XGF%Q;>M zw0a4}2M`qVwYd1{gyUh767j4}v40coJ`tHvQO#S=v`6WR5@MBIbxAegwJNU93^p5a zn0=0$*H=TX{A>?MV+#u*W(jn6h&xK+|dbnq1Lxhet(UN7fD8cCMVKsgdOC-cm_ zxH2CZxU*Nlr>EXUlfgy1vrpnvtLst92VWflrA(F&U8EkyyV@5W3$xtrlBLWX2Rc9Y zXWl4OAa8_T$^%D@ynzlZa{jK8BMu=?TyQe;35YjdU!p_=d6@6Il%pu4h2{2!)p`0l z{pq?!s*Ij=59JLmX8mZHGs@6is);dUu*rx4yvO5GA>6XHds%&uBAya3U{+|E?lU_{D@nrjsx5>)OLQ*N^x+}VL#X)g%`fMS7>eKo0Jw6Ga?}SJGN$QYXR5`{lRA@98IAjQ z%@4vvx`JZa!~LTs3d=XLev(7a{tuR_kAA96*hV6@I%L}b96^zsh#IER4T=qcgRF$e?v~k zL6vPqc)L-ohN~^a({FY?HQia?R_>>Mdov9-Y!t^Jz0Utcm+=A^7AM1P7`TM`A`lZ+ z%n$DN3e}f2tGuj4#&TYc)3{QUCXkP^rOF@yjGSMy!W26+a#GVNZXCiPK>`~V@nBR$Bdv7u(#Bi{2clPlQBI3>viJxCf(g19b>ouB^(1TFiM$Q7 zJF)t>>8zl&r5!K(adh@4XZnG<_G*ny*wL~7u52Eoq)#Td4~@fW4h!v%^iJ!N#cN84 zNu=SAk3DvaZ{{2zRjmq9#do{KHSn{f=K#7;+Ank>>C~`A8`~nFJop3vfkItu&O#95 z?1{a331xzcp5ddN-ZIoLRC{f+ejakK_L~n|MiuKJDdQZWoy)>azaJlDq&Iu6&H$ok zHYqj+y_%QiZ$%9uu5DFr^odg$_NcHu4}|ixWNo1yR*ZX(b=``a$0jf`bxKw^?nTwXs2U25)@eH!3;izB|^^%Hs()iS521KS{h%{UP z5HD*8Sg*-I0y77fNr!l=B$_bT5_qBTEN)B5k}43DBrYD46r02pFD)OXWQNatl+DBQ zhA5SB?&T#lIgn1VwK8jg$rKAjC5b~!le&cjPQ^f&%}{LfV8$$rs#tUj*_mXqy2p#~ z5IL9Ng~GGUrLt=i+g1V#YzZ)tPwK{d!z6MpVcI;)Tq@W$v27(VKTCjBVOEzSSHfF4 zmW|m7F%L9r^DOhAklT`ID}i}k0=&#lCdoF4Byuic+C0l#D%du$Z6z>2OCWX3!VVXH z90-^PnzebBc~Ho0Nwk%~ye@&%F$?>B=b0XDo@JgDbz4GhB{1(wAa%^b9tbwPa58Hi z=+Wj`=0PF1CDB#_^ST65$1K`zH*ZvF^DOhGoZC`qD}lLH0yq)QaI!Ck6AzsE=BOlx zJUGqBUPrFl2vg;45j%Q0NF5^Q65Qrl=2F47iES%^)+qt5V&HfoCnwmL&y^0Gp=Sw} z$&RKr9B}7gD(9Gp!n1I=j32S7X+xZfpBuBaiS8>!o6jTL?qg;kU99qDOn-&`4q*nnT4}#^Txll zd6uFqppB@l1e#F-+_Qnd;Cwbm%sHvdF>WsX;o=*%-h<~TIk(twmIv2>Phsml8{2d? zjr)N3DMe1q^Al{h(hOx_nK$zw>DK01npa0{Tx}(g;TjsQ_~Qyr-sW;C4`T{uE?mmN zg&W)~!_|dcUd3&j+_ptwas3eY_^H&G(YW})8^DM1ieQg4N z3?-oZ$yxH``c5w2<$4EhN9L**?y2B*PJ9HnE^*Z1Dvt2v z;ruF~dGo$mxoh(*^Q)HI(w(0rz^H-63+~6`QZX(A<042d9%U_qJCpDc{J7kNtEY7D zCl`bAle%06#!oO;4$jX!l%6(M?QktA*AJG?S~5qQXDPYJ+87q_65uT?m&npTbH5Dt zRC3V~HwSXpCpUg_S2-)ATvy0ggX4MJ=SE)eAkGo@U>fGiPQyPAfG-VrJT3Jb5vu4WI-lTdoJ6H$~CrS>{bSFYKvc(7gEIEi5aV zykzB)W3J%mMJx-V>=WfaZ9ePAcSX2)pIfORV7S5N3x*ylI95PB+~B@@uH9bP3$$7J z=X*GeS-8%*S=^;DwRx7(%Bzh!wFFq&V2>7)6uu?GmngUqnvZqyl>%PD@(lnk9>%k9 zb0Q8xn>!G~yA^(t#yb~&f|=u_Ceem$c?qyDCiOf~%Oxizw0V|fnY3XVl>l2dnUyen z;Nw{AE9W~qtaR|v55BCz8y1!}*gwi!5{4c)5nie2^FYjKEQMVjhP?sIm^NUZl>nP- z_?A;kKCn}~s%@U7__F#DQu12=tET$hzmKZ#8$Y|c=rp>d=n{u8%srcv`a}B3x1@RP=zVX^9blhAM*&{hJ4OQ6lO6fU0yHx}=} zw|@C%7~lP6gE>nWYz*ZP3g6ab9hA2Ktd;V`JWj#t>t%dzkg*9bU)*c)u^Vs7WQW+0 z*`HfpaP!=zS<6a*6?AMy@8DV{scG&u&(d@mEjSd68W?);9YVHjFmvI(EL%C)xWQ*? z`EDtC>JYsYuHQJ?uyd4?EDKI9+BBJ8C4kNNb=%z5q0O_*?GkQH`Hw&TxW)0^Xv4gS zG28~Ul|a4{;9SDUkt6vqcKGqne8erpw|SNp7GN8DTM7JNO5ncx?we=2wt1F$R@7|? zwUxj^R03_DWg)8Uwyd_5z&tB~HqSE8in=YKwh~x~N}$cNEJT&vmesZrm}e!>=2_-h zQP1au*vr5M79A0u&#L4}p8J+Kd^ERjY|N8*X=q*VQW{TdQ?t>rh1cS?wnmNQwRx6C z#j)TAcyY*m0{iT~XMw z-4^)1*Is*#JQmO1JW9UH!{)$dY%J!62R55=@UnU7wh9;Lm7jX*sS7W>uvMhZ4Iyow zWp0$}+>(hw(+IgIt)v9-BV5JE(LpXQ;Qq9^>Q?B?xdoCdV@uMIyJkhthjlm%&3@vN z*qcF2htGyXZWU+-?}EU>UA%YQb=QKRNj@FgJWIZkSrGADv&hXRXvWrlKG0jFid<#z zzylBP-8@bYa(U9R#~#ZGfFgNIC8IyreF&v)Aty()7WNI(8#8|4AVR50mqd@AjIB_T z1D5k(+gM4uHA&3%6pRWvR{r6KA8PM)lPvR|pv|+)`!a8Sirk9C>13{qXLeJ(6wV)W zwuml{kqf;v{-t={QYpE^84bCxx}@_29H-WILOET{eC=QV`d6vxmP*gC2w%pPF$|GP z;$_UpiWMvX9%_{&q2LzGB&;gR+Y-Sk3;h%Xsc;*IJvCF&HQQt5vc$zHI2k)FNR$u7F_f~DuQ z?{XzM>EleztW3K_sO7RnH_UYCvtSkD{GScg)z&z^>- z|F-{AS^v4XK+x=p`Q(#NxZH&Gg=R}=9%J43_p6`WCo^--_i-$zZq$>t!>-H)xsIiM z@A6Hr%J$jNvGWd?GNSs)z3QfXRWQXYez1Jd)!D8C3zE&v$SGgf>zP|UiGLqoKN}-G zugP{@dNyX>{jZz&QO*DMt()GV_8|?CF?x>c+KR`sva7WsjY50 zr@Z`_F}nKkJs7-H4EZ<1gr-?qg<{&Mss}d7b*fGO>KMpWKe$DvqC>?t4`jQq&@@LY zQcN9L^}wJ^ZSi_;o~h_qvCaMSOivm}^~2j{>g+uVWxj6oE7kXJlxf$!V!Ma4o#y(R z`$DSBw=ql2jH)|U%+%GljoRw-Pal18=WZXI+PQzG_VF9W{cXS^w{2OG=JV=m-afDP z>xU|a+)_AKEs5pxlzgg!4`T6|8}4@G9V$0SBk;Gs{p~Hc+`KXY)|iiei3F*H9}Q89me`^4_; zdT+b*&w6)UJe8sL%X@2HI3hJFFOrs1HMRP`+m^3=X4wjdx#<@6&0?ZI?Q* zd;9d^WOMbL(pq}{{F8MFBE!|wYKQ$L+iRV&{@WDEkxEumTlM;P zkMl=tdA4G$MF*o{ua2`v;59R9UN|z9v1uf^d;Rom&o#2W*KL}(DTX$}7$RC8| zWfki$wp}AHvjFHhvu66UBktx84cnJ@+@Rm~TMgK^YrCEeIh!`BWVhlnsjI=c=Snhd zj!9!2KjZs*hF{4a99=5AZ_@v#IGPR>oh+c`z+CYx%XD8I&oZNWs!dRe_U*jXnte83 zzt<*Rr+pKuUcPweqt}^w$Rjf^yR7$Sb}LY{uh?SaF8qbWnaXappWQn5Jd4ue#UA^S z*gwKY#o7CCQT?xx8kx-7Z@%3x;%-{p!3@&zCK?AjOWEnWmwwKJ=0tCGZw zG=`0Ri9e$7GOP67XzRL;{p#wHltp5&vonTYHsg~^illFntfsEoRjk&#)wIc7GtJj9 z(`DdU@1I^%-l55)ntHTeZw}Ix1oz1y`>#0%d_5*+3 zzsqtp6HaSMNwTmuoz;e-OR`J5MHsehzQk@7W$k8{R*U|($g{ll{ZsdSdiBJaqeB&# z1TEEV)sM!#V&qcYR^Df&6Z&=qEqUGwm6cbzcl zkcZpq<;{K>zmisk-shVE*?im`S z&<1u|=FmCwpnB@EfqT`wdRnG-W-4zJ$PT^hCJZ}o?OU>oZQcZB!RSA|anx6nK2#fE zwPne!D}?0bEBw6n?F%!RW;Kcqy)qL&x@6s3vwb%WWi}7~%bUl1KJk5cDM|eDfNm>= za$2Nkd+B*w3s=N1dgYMO(?+Ny8_jv^6)ju!sM=5d8R}d-%nPgzlfSrZ)5o(kw8W43 z1nU)`8KDg7ze}sngOp%%Jj=IJzqt0@)0nDB1g>QBC3fepW5t|l%h@mN)3am0eO5d^ zbtSTL+xx5TUOQ7YdFNmC-g$P;%ldSF@`l0#z-+eXT1*Ve*Ewr8Q!D@H6v^qQpU(ML ze+(4=^3g{h?Xkxm?5QtxCwMcI+k3@h=ghQ|{Mm|%9UiTIaB!y78AqqZGgXty*BshH zw}M88ta?^AnpVDFMdymGZm)h~&wTOu@Y#-i>S|_{uW@3lxD`@4VC7SKbm*HexlB&a zJF1`9BQHUoh@uKk+PAz#O?4|M z*S&qO^EbVRW?QVH&7v#KvrL*f_786y$&<7gHSN9PaqIToDmC3RH@Php-+(OdSh?qi zHP8LBcEqD|mU0U*WU`BHQa9$6?BYYpS3kZWVKXD~DSR!E1zNr_%GWk5g-b(@tN(%F zq+HLQvlrM*(W*c`w=%0{#`qcEji34bgqfo!YLvZE>XZ88;y3+I<{Yu4<+ge4> zq+5wfa%yGyl=aV8Y}i#ROdpe}En*&A-kufuy7BK;Z2Mq#kw#;*u#9NAtu^Hd#YS)m zp5^JU|8w7`SHaKt?Z8z|TWirRB3Y#d%QwBIysoSbzr>6a;-BA|jv&DA;(Q4|piTLj*)nL@A*Np(G@vLmH5tkhZ%s z-+%UIC$l@Vo2`N8=ltfE+_~l4GqW@2o_o%@|J!FN?bF#5HQVwt%P)RPw7j(r_nvhe71guB{X?a~MRgdrs3bg5}unQa>V$PVTl?_HG?`8|6A94hWjEI5@DcJB?Ik*~RUtD~oFK31gTH z!)4&Yxa~?`hgiwiO<=O_kst-XM?0--CI4hvBQ_%eG14ug{mugArr zW4#0KeX8pl){=F2{q)kr-5Klit1<=3@F9$ecpfo_h|0|70qwZ58m3utQTWnV8u8xT@8aKj(aNH21%PVggP^gieeP9wo3hsoNs(ARD3%0|(@e z5VTfRUq(3jG2;uu;k_N64)F}PnLW}F@8_>q|MPq9P9fkB!NGU6c@T10?5}(&b`^pW zW#ieKfzQhNl06w;oW1^w@Fcxm{18~sJLq0PdxB1iYO-T0IR7F2Ggj};S9A5`-uGt0 zG~tHyz3rd!b!%<<}~&An`<(xi*Ux11()gG`wZZ zxQ?@65g=U{{go9fsDsh?frx2K6DHQ_RCl4(mD00FKh<@fJz@n!en{U?a4Sn73#-JK z$mhHG#R)k^30bj%I?m%0rtY0->k33V6y~$zI?lv=$A~v_DJxb`1kMMwdt~{kDMr?4 zh$@s$iCJ<-i50Nagk-@WkW+D)+%smUB;mZjWHk!t_h@~O^Ln`xD8mW*{gpM42pCi! z2%p+7WVCHuVaBu|{_FQnK3iT_B&L8!qP{`*iF2Av)7K+JQ|JOJ?bwJHp-Blj_9yJ*)6|pf6Exg@-UtP$aCy#zSUH-v z`%uo7YE7jef{l29&9s~UT}GJ>ylg_U2$B1~u1=rAZAG z+lPj=f65*m%KY8gz$dK2@RW%sWsxmoVS+U^-!acXlci$e0z>62w$#a1!UWMvPnW`5}N znU$&CDOavKpT>d%xdY}Ws3*mG2N|khY0}up7sGtI$yIlJo+W%bX})9O&G;+5Ukjg! zN7R+^pz165@Hxo=4-}3TN>7-##jZjap{8|TDyi=H3}Z>b7ALa5PXZ7b=?XvQ`y^G6 ztWMpbOVYGQ8y*zT5g-~MSg)DgY#-NvQDKuEsHHb>VDt|G)<<7tbN30^@L@zz zlt#D)#~jT>K6NX4_JQ2*P8IAI)Dt+u;CzMWs2M~ahD`0T7l9}y@MVV2A9MaKo;dt~olL%{wTpA;KCHoQ+QstRv{O+w987r}kuS#2_trpgdK+YPOclLlJWf zo0M>yPE9mQ7B+jQ4y;Tp;AFez>Qge4O2@>yP(Nmd_^Bt}J3rEsKe zx%~dcn@33ngoFcM%T(n%FdsYkV5w19SlGsm8(X(-jeSv@iAh$i`5_J~9xUmOD?H}D ze`H=jD5X%r&i7E{OQt+cmQb^_j<%lXP(MDIY{kE89Gkv{$@(@aAd;nNtVkXk@j^2M zH#D$hC7*QGjecJ`B9UvCbE^=vhK_L>7Dv4VqSvL$t^c2<7!TYYs0_c@PvzY$CA-Cj2P zkT!N;i|G$|j`XKg?ObBhp&DM!TIjpyNpb~Qu98% z7>{>;mAP|(=Jij<%zvz4Ja+8ZTW`J9vSrI5Lxw~|L<}B0_`wJ59T5JLvI4GR*k9q7 zk?*}SLk-keeXyo{9m8CPQ_s1PhRKm4_;E%6e>1FO`4!AzVcyYZIE31({n zo$5Ne)y=-U*?>^RjnkzcR(K?Q=@tmxRd;YT>W%yAyWa2d*$hl(Za|Y%4A_7Y_}KPj zett54pP-2$Go$=sOBeskB`1@MPhP7?6&D42hCkYAw#_jYgc&FO#WsC_WdT0R)TvWR z7eG;%H*X$tpq?NHD;SBAQdn3xdGh2VM~*mTlXHagoG$$NyDO{F1tMV3nk5-UPO;7DuFE~b6D&_M$mGp$b{=IBh}LBb)%z@hD*KqPMCr~sVAgJZ0x zs%ybf+CQe!qkp5lJK9J=)U2eY@a*+p37=5YhPE4TAY~lD=MjFrf*nV}t;`EazY_v* zSzzjCAZ0YSv_~e|NS;w>4vkYie3p{KyP#=Cz$yfN;MV|m86ZkF_GO|OFIm{DyQQU3 zV}0K`b93gaMWu^d{T8nmD!O{KyK2fK`ggsyr25+4pDw?D^7^l0sUZ;jF6F~2bs4#Y zY=+q!xW0ITW@br3+uN7`EK5d4#=&Abz}+*)mrxlwLBqx-79f2|)=^VQ``Zrgwb3J&Kbo1e%wl&rzQ3Kff@a z@`{ivwJs052@N0(A!79MItXyHy-U;V=5wB%e2PPmupHs*q$Px5B4whpVfSg$G$Gug zVS?4vxd(rBW~S`&F?h$twc*hxfs4ot^NGGc^x2NSJ;fDlOY@S2h7noP`s%UDUjp5F z=GOj{%)Zv1$9$7Nlx}Y7sz=*B*}v^0XRaT_@_!+Mo7DGl3s~6y-=<<88zOpi{$%@uiz*5ydfy288+`|+liS5VM(xH_t(lr&VE<)GZwwBa~fCfYv46bxZ z+MUTKIE)J#0beqm`GfSUwB#O19ZuzzrY@pNi!5w{Q6EdI0KVs_fDfHy2_DK_`%&P4 ztEpeOs;d1|zk>s*?EOVC+e+8=@}JGm0Avn+*=1cR?b)F%MIm?z20~H^ah#Wmj^q1D zEjeRq8Yx5WPRhdO>QGhE)6++d8U+y}hJ&(`P54l8$`bkF!h#eX&?mmtLG2zx_=bT{ z#y$_HUeQ6Z7-*5Qq>M8e7p@}X?M};>@W+jdthruu_8%lHn;@qebJ%erdvT?WgRRPw zD^Fh@vw>+l(i)EB>k4RX%Uxdg2eb%{cXkppo_LQ}5Gbif`9 zm)0d*Te($at?c_f-MXK&A>7r7t0eidps3O&HstV6^c1DuTKRMJH%avU zBRr=EQe0L2p)y4$Ma}$oi!(=hzB$5GFtw1L=+^Pv%|De7{4VVnF{T(9obHX2cI=3p z&6_vFBD-M00uaMA2fE4Fv19k_+2aV>P0+xIIx%{oNmnbQR1V#tIrLByRHVf~8*33v zx9Vf3SOMYtLZ69j*;8!#G3d?JPtkvkWllzec9g;$08w=4IERQ8kcQbJvSknPw`$NI z&*xeIK3OdU!KbSa9Rfa0ij|VuLSK9_GqUV_!aGeInZC^fb$rx9~?&a1EZ{k zc3Go_PP}w$tsHlOtIE;-F&#VpQo#3-yF93(B_V@V)hUX~%UGe0`Ki>z zgGFH&-mwN3e+Bqhd-#V3O7(aqck{IN*z+*#z@tTJ)y4#A1?|@ZxZm8Z+}yx8uJSxv zlLv4~oMihkp+20Zm}u+R>mKG4V^X0L8d#jLp3iu$@*G=}59}+PWD_!>H|R&9K2cb6 z2(f}p%o+cA{IM1)ugTGiZC>955HHn}*Vx>ju!9{Yz#O(Hk_u~@;gRR-%Msap+&7}2 zteCC8Ay}+oi|Fjv2ly{t{oO!=8rD{u5O%$wI~3HU|8lg z2){l}fKC~;)LlZof-cM_)s<}S0qi=wsp>*Gt%Z{Oq)Yoi|9*74k!1VfP>PzbYa;xC z5X>yU1fNA#`6b92>_dX{@YU@?1tj3=4)Kftho*D~ab+`U3-}yG)*^&T_=ML99A;G{ z!Py6d9pFwGb%f}xPZD3j2RuCV8Q{ejm0Ea-nHTeDDk^ipgub>H-{$(d%zA|2(u(Jl z?Yslx=yszq*rUQdx)TX3M7&V+2UcC-S57H8jnk+vI9m{^3#bhznJf$?OJnCWzps5j z9Ld0R2q#y1?w8B&!)3MU(jw$Tot_Z2z-Fbov8%sRc0K_N{y_s3K?5t!&MK)$fu~Z_ zX)FrRzkuIbR>whjNT_#oIOYO=eJz3oQF$?8 zk5XZe79>2Q8_zke$D520c*Z4WOi*DCWN+>hcpvIZD_EEV z$X>eB6spsd{6EGvEdGjH&d8xv!h8d*1Sg@4STkUEVuh!;BZP?{VgL|)vwD6kyL*qg zvbAYbl0}9NyxagQ1~)($+mrD*;u=<*o`!mWDa3jN2kNLYegB=+EKdN_a$ZC6?Pb&( z;%@L^;A6&DeYGMLpEOSKLL^cj-%Ll}UWm}`=o<@9xe*cFWrHIXioIgJHvj#Hr4yNP)z`C!JTBu9G+o zo3x962qJMnffY8*5xxv@Jv-9a1xZ%ofF&|pLG%OEB_eTP!W4HU=s1u#P{d5Slf-RI zq#x@v$3%(4vjyR7QPc~};mZa#hjUZ7Ppc)1&=LyX#OSwhiC+sFrWDoWk$_P+=V2Mb zt`S1bzZ1@T0JvYaQ{-&hG$N_rp)R?74MUK6MwUS7j(S~%f2|09}Y~8d)R^qHY zJsm9|uzCg#1%-|+d&1IbNW(E#5NbGu*@{3^kT}FXA}euF`NK0W;xEj`FZ5V$($Ovu zb#ekgzrdF;>g3&oXOM%Hcd(ck72V!T+a-&@Hx$45@tbgonw@ z&J%B`hp}Qnn+NRy>=Qcu3nXB!YUka#U+`!k@z{Jj@E_0Wxmqe*;2UmFbRY1E(H=vKz_4rrZw1VASjH`s_6i(|eV2txK>x!< zC3u2N_?~U|1Yw@oxMou5RgwT#$P)_}tXFterA(OuJvh&y%g*WI(mLR~eg5|Zg zNgO;M@$W8Stsw(e2&BZq1y;^GO0|A(!4aKikSTypYg?k>Eh*c${VVH{&V7UsqWjaqYV7T|&Zb?Ki7 zGnvzCovE-Zv2a0zn4>_$FlQkkjKFY2s0-j-y7>Zxe6-lP5g4Opo;t== zj$^l{u0kEGjfOHSt_!3eGZQGO@v z@g)8!g|y=(i$EIw$SE=Jet6~u*y@E$B;WvBlJL}m_-{;oC*sO~?KEByGHW2WiJ|*= z{vLQPVW+^%QEX_0^$}rLX7pYnTn^@Jnb!S1hdeYmH3zkOG^zN+m6AU(N*lSKF~Ej+ zg_Cv9&uFLsnvB2L3K^a81=Q%6Et-DAXqg4p&f4D?PXH=)>@5Pl z=OXM1tk1FHKg?keWH`lbAWDFV;|cd8TDHS~x`oT!;h=c$x#S2L1%d;dC7vS8P% znLR&}-X0`+acu$Q&iVX3{m!kVYD=Vb8gn?7%{x$gV8`S?=|17%iIo#x-tbCakH&I` zrvMfQkSBL{cTm=7G@hQGAt511W4W_Z$v>{z+FGnvbaizFEDJ8RTJ84j+X4P$r&s|g z;n<$CGNbf7^mZuHn7!jfnolrpuF4kJq_z6Si5`g`Hnp4tkM1B3o(ia+t}_)qEY%bQ zHlzvi!9X{`NBhyAFG(vP2W#A6Q6A*hA^QBsO>n5KWiM}FQ$FLDuJvETbl{B$_8d~X zER4i&9UW)189Taw_Vjc1H;6;M*DIz+BHmKgT)*0SHuX!K80QpmE3lQaxyq)?Fl~fQ z3MVm_kcyb`Ej~V8+6c2?Ov(_9Kh~#b=ku#&g8D^mICZ7={4(wN7Gh#B)Y*01sJ;d@czl=FaGD~d z*&WYS=b9>&$%lpG@~3Rd=gN@>Y3~S`3f5#14)GT;2M{y??rj>>6A=;h&TbY#w~{72 zIuNH#c$Gc3L^=A1wfc*wTZ1HY7R<*F*p3EYUudnZICRu3GLapjl~M9S-SA%au4szZ+f($*h8C3*0*oD zeW*@J5bm00?(5D>_#HT*0>Vjw>I-5`nhZ#$Z^7rwrf=brdyw2U#ap2p{~WE|g~9VCC2| zw0od!ypdMlrmlij9%+QJ-WNJi3fipR6>Ei95m;gCFz~AP^jAHUh=EtgwmAVlG1nk` z#s`nUf<_}!n%#XZL@76DsI8~yTRfi+pEvdC8q_y^vGLMdWM_Ls^k51sb+M^;cBk|^ zVVfo+Dbdz)^~IVW2P(&$gpwsGrpoJzSDkqt`Ir*o1;ggw(jQUU%Gw+U{DjtaiTdmn z!4{xz%_!&1@Z2-6;<;_CM<)03tN!2f(~*WF`}c}f#Z07%%gj^GdZq()?j366A$_5< z{y+Wn(@9B5H4D8Z6yr^7*OyDSOZcHJ^`BsDZ{Ujtj?D-vMfVWOS`wOy0}u& zZH{FP9~qG|8a|@tED~*5m_qD_8mt6p0m7486x&Z}g)m51NM#;2}b3 z2@Qv3l$VzSQ*OzUCE3~8{rmUN%ggij_I4C?Or8Zdyd^YYlfr(6a5oI0JXDr6Fj0qg z@%mR?dWi}9h>1oPW#jAJcK6uoi(sWNdsMUD{y?=8bh10Twbd6v% zY|dlGab_7YzSixKEUvz^%7ati5_4I1t?KbD+@d)FD?Yd}|I=He-}VSIAvb6@e;p#Z ziDEHPo6`05_3ym%&a1D!`oRYufWX$RTf2Af4&id?(j{1kY$9pgh{XFC=jyJ%xbUX1 z;2kDg?4dcN-T21XpSn^t)iC_ZHg6(RQ#;RRn4;%Do2X{Cc(|5mBup)9E)A3m;1u}=^gmP{`f`7cQL-Z=J7b*Jzn(Q`C7%~{jR@` zQcCBDv`=TQ_?&Y19((Rtw(a%n*XPWcla-YP0}9~de0_aMHCS8VCxQbARu!0O*REab zR7^RQ!WFJc@f&8qt*bA5wA$p4ZP*xu2^_$dB`bT(Hy~*J1*8dF8`yJBc4J^b5HW4< z$6UjD!W13QpQ_1cx=gpBt+WY4E8+hZOA~P&yhy;5X`fSz$Cz~k^{d4-w@8pPf>rX2AT?Q5?KM&DcG|G>b&ZQHgr zUh7jZ;^NR>IsN~n=q3~jru$sVJqV!~vwT9SUg$8z)4 z6azO>r{}X7TPy_AnQiuY|58b@TqJ;dxhbnRgO6U zN4fdu&}8ZCUX4?H*~*GxJE)`2bGiFDy@^D&nn9+@lLx6{>y@K^F~}EP0K}`wu0_>t zOx_;J0%qyJ0jei+wPzQxSH9vmqBpaR4qiLI?to1L7?o~sWH~*8AGIrY}i;fjq(Q7eIs@{r}$_Muj1lA9eBQu6)I1?*y7G|1=ctI$UcF#i;1u zL5hx(=-?r@=P9}RDevEpuDP;g_VRilxl+Jz<$qi+4pF4_Si85D@|OMN|-!S3y*;(FH+NkP<}%`-e(VL`4Ne zn!Xf4AcWpR?>!`h^mgBWZ}0Z@_V!wG7ZTrhe*3$ddFH9}Y?+;TW>%-y>%p`)2rW{D zjKh@n3$A~Lh~ISrW|GGI+uFd$lK$HxGoEZZ){J8y(cw^2W*)x}OaBn6Uf+28$JXMb%4RnY)LYX2NTI&<6XeLkSNGnGV2Q>l_q6=vCH8jXRi3C@3OD-1Ga7!+vWtrY8}(ar4oyA07%rdVN3tjbF% zQzACH%1};-Q*o6#rCcjpd3%(aHh>oTo$9nY;q1G)6xTmNlTGR<3h4eYG`~Wa&TJ3Q zL6c(@$mXK#JT+=N;V`B(Q*oM^21p1w%_2?E?-IZ|RP2Dvze&bq>+BKg#FV5F$*7$U zk=smQL8dL@W2w78+c#G39*eeZU-Lm*#0D~p6U&0KU*lJJSQA@AsHO#_sJ7zlQzcWU zBy?(rC6c$~wu!98Ulp0GsCckCf5Y410#BfQp@_XMr?E$gzz1;a_lrda95pgl7|y=D zb2v~}$<^Wcq$IE^=U>GBa3xeY5l-}^UP46?;(x^M*|;fQ8%dh&HC8DJY?psHHb<;% zT|Hj7!yfN}^tFV<$6>=p^C`=A*P+pBWAkfqcrVh$HZf51gc7+p^GlKA+I?nm!<(cLSICkc|31sL$87I-f2s`;yRg?_}u8WdI32}lW({r z5%MhRenEij{IPVNc;<=%0(`4Z?yywM+yAAaBEr*AlZ}{`YzdT!v|=VnP0vlSs#F2w z7G@FZrSS^&uN@QGSp~xYvL})VQgeywOGRsT4iO|02O=drQ5gB^v_3N`u9o1a^2qDw z(^<5mjTzaHb5^8CmQZ>6mTk#R7>aW_Fm#xr?V*EL#= zml>wY_COuO^BsoU5^oA^P_PR@FCg_}17UL4?Ul!vSvbm#-qb@xnbmczRzow4Y`yMB z!S|SxhF9wu>W_CbpFWf)by0-M+><|Co@2mr?PvC`p<~*9DbcLG2ikaJ&qn~}95;)s z(iZwe#LUHkkp}$c7aa7&C9Z~ka`lRJ<&vbz_;H8|luC5x8J*xWgP+q9g$koQ{}wVV z$61wIkBKdb4-PkRs1VJZ!Ae%2e-+o}7;6mI+0?(Ee6z#1CR~^Qc|r-DpR+@8LwH~b z$-tKXriQKk0eHyLt=#J*W6#Z{2Q?Xxioa7jelIYq`?+vfCX+}@C`*H)CM`VlHlOu8 zW_Mj@{hnh*foVGir(|Ce;gx-Ytr0|sO_8yocS|(R$Rjd4BBM9m&rIi9eI#1zj7^b5 zt>#}lBR?%G|Mmw-FTC1Xyjra}m?5G(mLHCI$qcp1*%LXo-2&ntN!j`3fj2f^`04KO zN3F=s0MUgir>+^rn6oERL}^;JUGvbpe?9SObV%o5KcP=;ou+K-Pay9*)>YOJ$0A8H zIBL?Qt&hCBecB7ddaj&RqmwP0hfcn)?h|sL{k#hzIkpj?736K0 zj&45*Y{mD@!8vd7(ih6kC09fqZYlDN9}oR_RC&%h307FHh^#gO*FF?EXbg`XD{f8! zToGA&jwsTvjlkOcaIN~3Q=&{ru81U!;gOWH%S6rps)JGQYgyTGR@TEK^5GW9-ltCF z*BDXB>Q#3NdZaC~!7D(AoMaq|l+!%a4Zr20XtMw-I1BGb@$oMh9&SR}7lNj{;NRL*cIUtN zFcLX(#NqMCQ5HR8R^%KNM|DM=CjPY_v^={C2g|-7G zFTu)ZaIlWrIOow0I@FjX^d>D+Py?a7;wborp|p?{HQwwOX1v#0(v`F%4IGJ#*cBfV z{Z>eYT4ugiGH`LMxyYOkigpQOJPijUXM{zKu~1{84%J>k{xw{gF4fO^7Gw6KaF^ip z%j*P}?%ayDzNY$WFQLZ7_NJ11{~HCh45?7l9?!5b;1J!lxz;#36>o6(YB=BQ4K=l< zfd}opIDy2Yzr<)0j~d;f?=BVIT$JY1EA+V_6Tt2N4lFdt%L;pE`R(ch12b`Dnxuen z=u9dz6E+sL-V*^b^LYvLu$!K$33eRWks}uS@qBx*i|)?KUrv0y<-*V8$Vr*EI{qNA zuJx{y0%S%yN5NMq@gBJE`erf19_@toi){6#o#U#Gm6@MDWm~@@eUZG{XrUbBDDSe7 zWXOK3{9dnzp8a>i-|hUXpU%+}%#lK3eRe^{#J?W4#&zaZ$C|1*Mv{x%PQHM!u`7B9 zy30cHub`_m;jVA^W-QmZcovdFF^weUv@XBS-Tm_DwqNiKo!VsP-=ce4R=a0lGvV%4 zlEhPkFIgf7w%ycDKh91ZIqfQx+(~;>(TZ=Rf?a&wCwV078g?n;)Mxv~v37cwl|Ej6 z3N$;#xw}W6O+O+cywmw-p0Re2U8-`Mk)qVs+kD0hhenI^h2yt8#JuDUWbVL?FOjuJ zz^vE?A;aKr>^5z^xMKP7j|4dIO!SoKkWOOWUB7NNQdEmW&b$?gLrtf52G#dUQ>x2R zdxl?_(%Ky1Mj92;VMz2;6U?3e9x;*%@|K%W`44G$+rG+E{eI^C6w&MLqZPZ>Q$`}W zi^9SiC$RmpLC*s89*u5eu=X*T+WG>V`WG7fCoFs<_bYS^X@p*lFsL^crwZ>4 zDk)1zqJmQ_kjISVE4h5O2}n&u@NleLhlRm5G>txHqoSl>j~PkDI0m(^Its3dItFpP zP1S*{_Nj-M(x?ZPx7%iRER>egTyiT@Nl#P^F(I3Cvl*@M{f-Fl$Pp_w02;gh?#vz z7-5Y%szb&$u;66MJA?E!2)tB731hdLguTq3={I?#~;V5J8^C~Hf7@24tc$hKYu&< z;%)a|kcE7%m=a&OZJ27I?fK=<^~Dc2(JvT-9s>L=%`_h&R?BKbNyS3K zPi=tqUIxeE&&O40N*~m{-c$G28>A|C$|Ahm4!%#K-ov=dR|Q4g!XAC$6^f$kcy%{E zXr&~J3F#EZiBi-2ymD>vaGhq)!c;^b!q$R{g%hVeGeo4I9Dp8^nZ0^)UO__M{^ zKt7r#pwC+f)1c23)L4y>0mxn=V;QkdR&ZywNBzO}gGdBn5x-k#>rj5gdc#@DrO@8%iAv<=@V&c(7GI_EZqwT5KiGCAQ1~L%z-ZQohBs$^mD>8S%*( zFM0B=j)*^mZXR$m=PJjApNJgOCajwrrF8P*vWHrg8a>$HS~h>t`lDdo}Z{Sj~{O4%;94Sm5txjKbSTVugcn=9|I@J#w_RE#C+S-IU-F`p`Qlhe){TQ&@!l=U z;Nzl7-op(OI(i!FW1|vsnTJge^)&78G!5}8#8+GEj05Gg1Fa6K^*M#%)I+|d9CyL6RCSwo%}+?=N&sTxSjK~ zwlLjeA-VJXqyw2LEGl?wTmQ-8S_fNV)!-@Fmp|DzTCCypFz?uEST4idW+D4-Ja6Cl zt1neZ&$~G%>7`d%&vRa#yu#dRJB;HLVHN9f4%~TK`3sq7+`k+48BKeiDK1`gOB~RExv7-Rcfle|8D;KOdi8(_7Qv zd|wN786%v&4(CS*7DgW%yJMtO>*TT%3E!EF`z@r@9`&ETl6i)m5`V2;X`Y2rdM-G8 zp=!6e*9E92%p+TmSV+E4M^E`?d)#2}O@i)rQ{9~P6RUOhYy7*Di+JoZQ>pZng`Dtl z#ynA@pzObpe;T?N6$MKS@|cCmofNuxwc(xv=9Ih>jkV??j4MF&HZ%Hvl|ip_;l6b`jJ)yBbR`Vb4|u=9g0PjD_;W)iOgk z5mNt6Ok98#zA~tXv!1a~RcSpc^hDH?c)Pi3wWoj{v(S&B4~OMi$8>@|90y*L!(@3= z=!6X9o~*|#j5;KrLP=ePgL&+&3QJnV^x6U>$er1Eqk9ap;^Hw2DXH0h95xye)C+oJ z*V;-20-{f;LEQuMq=gk?(BdZc20PHSWd^hr9pIFp+LIPiEw3OUa3l36obIoD;&EcA z;jaknSpwyqBDIQZGgB1_yB}SfhuiCw%9(LYIX)gL#R)G&fxU%Qa{o&XT0jrR*j<~C zTdVEOXo^fUKEe)J+>|LGUV7hDu(vu-8Si)QMWifL;5jI=8^Z`oDaevhe%!FhE6@Se ze_J&d2%mr1yfcWC=Iis2_N#D+DP#z0e*~??$h6Fhxu=9l;5_4Oos%Ne(Ncc^5AT|2 z{3raHsN?)sLAi|_+m_*vB~LdNfQN5=p=}vuZAd0Bo*C@&_2*Eb{7XUjcoaNF2wU7H5CRc|%em0J+$M&DDmGDY z%CDE-vFE}fd_4;P-i#s*x{bw}f0@^4ock3m-bHE=f;u4SL)6#WZn%uCxHLtF8(-N* zG1i3l#O3xF^b{8FMze01zX(IU(03qKrU~mD<}SpDI*xdxe6>2SQK`k?I?~1s^|t)V z$gzkjEsgi#8~QWq}Tb zY4#+U>NLSvvos*(YTTY;oUpZODm@4k_65@o_ z%?9CWEZPmj;bjOviD!Wil7v|8mp{gw*3Ma8kg5|_)9{VenWRDENCaw-BF4{a5Y`Aa z>|E?}o|WR{rGHZ1!7~H!O%mRx2diP(VmuP;kj{iq z`EDJ;=LnJCYU9=>EW>{x=8eG1hNs|wnV92)Hm~8>_XthIhOg1Z;_bzP&QQ>x&fK29 z5!}bkx~^Mm~hz z&byGavbE>vgoTIA4uC7f^!c`AaW*XeX>87fHW-VSU_hvVczO+<8i*yQ&@~?8KF73Z zThv-K-2vYSTh(P5F8Yl7XB0EYjc;|_$m}S$%P#bcnOa2t&9*hXY4|l|?7FbwI49k$ z(*Cc}aXvEjsM;Ax@1w5i7Ro2(;#GKH90vV~>Ca%@bNFj0{1oIB+oQ=wXi9N=iq0=S`?<^RDZ={mF7<)<7pf6 zwcy=#fLdiyrMRnM@7D-ysaE-42IB&cCY!I_J{WRwxHj`ITgG!WN&ZkR z%ik5LW}yo(uC+|Oyms~Juh@NMJan~4!{x}eycb@VZ~p2mM~;SvCxG9Lu}YR+l&ZO; zyP?!^)gx;kGPff8AUM`qa1$Ic=S&h;u5cWWn@jW55Z14lZnY*R^#h~$2?u^EXWf}a zJP2lhSf7T^*9(j?q5L!d87#iy6e3%3C!>I6QO98DA1@S-GN$Y{WuI}g_r-kjbs5^n4NfW*sg+r^c%l$*IJ%;DPDwqe~BKx0vu(d#4}B}>kDu|yKM zU^FG?vJ~1a@?t#gpXa{k!6KK*^l3D_exS`hZi&Qun3(GrR-E|s=*>i1a_B#{fUR10 z!9DIpD#{g&ydqEEI;g7j`P;|l?Rk||d~kLN=LPztyOwpRiYqx#yzJJT`6j0z(>^!e zMR_l^kK>lw((ob`eWAr{2D7K;T(hHeev$bC?O3zdqm5s-gM3G-xFH?4rE^j=w-y~9 zJKIU*5~7uo7de~5&HSeR{X|}2j(v@rhjtmzbeyZpN$vA=SM4JaBYHP{{=1_y_Fi)^ z*~I7lYi&ODP;9idIi=)93g%{!ue6%S?{RSzhP57C;8nr95?3N~F=E+tp{O$#EI-w3 zVqERV=Iwr^z{Y47ZF!~jd^c{6Y2923Na=Zzf;pfyBxdUW8Fu55^q|kQ8V&n9T)a-; zN4sCHTA>CP(A)XAi$q)cnT-5f)bK|3MVV-g;;njU?DS%e*vXi>lr2>+QnO}N+xKns z0+*~a&&>^nHlObA9iV6zc7?cEccY*d&&5pRx$lxao%c(>t}uF^e+AHn6yxKH+umQN#aZ6^08zNb$A zUfeU|PABfPvq=`fA*#c--eIvb>jgApaZ-nGx#?H(&BRvK?<;*}?|>I+r1e%R+~kM@ z&|Coc-H{pc2j_pzeZTSCvL~93;Q&JEw+}<@wgt@8hA^B7`l-VLbNoUucnFy^n{S+`W0OvHy7QDDfkra?TxVKJ;kg zk#fi#&zCJPip{|X=R*yKesN&J@zgz11LYKCai=ZOI2>qaPXQVh1at# zC^$E)63Yj9*?Ck;>Xk(=QaNsc^nC10esg#3o|jW|u1f0O$iBo)n7DTczjfnUM4`mO zZiZc$1qa@=9#P|#8r-m`if^c0>N{3hM!hJu2S3(4vHfC>p?|V(6kkQ8MmmzRgFAS1 zt2>aJ-$}rdXI7s4eB=2a6nr=?_$oV>ZY^nl;k6+g49@i|y|_|nVA6}`|- zwR&^BO-@vn^X0~iY()0&`!{+0k%*T*-8c5ktwTXxg+u*zR?}vdsD1T$(di$~sE8{* zZNJE9kd~XmiHKp%-z&?bfY$bwCohWH^FcKtI0Ho+tc@uNBrGp+l20P)n79OS_^jL%}?zr6-e@#F&PD0pE{E*lGylB6PieJn5 zynA!>pGuZnv>pfF0*l?2>klZmJiU>)j7)WS&k%>t@`rj z#Y%Sz9mtap+iJzaOL+7d{F)*JIp04N&o@T$RwT=@HE7fl(E)OhWxmN~a6_qbmCBnJ z*|_{jOKgg;5Q#e!Sf%vcSVI>w6=Ch?SZjqq@4oh><(5|xj$b4qcYw zh`*BjP*zd|AR#9$$fGeMri%W+fdk#UcQ-LHA@IhH8~4HN&jBNQd`8l;$Z~oRIjG_r@wpm z?&HUg@fH$Ds2(RLr{HlzO8$)!)gPT%n8!$5PEUM%cmP1KAYM|)2QlJJMHCc6Wp7iARuB6A8hrcPu4>V}dfQwAAiFDZfliU8eymdw!(`GTKc0%AA56$bYs%gxRuYeo&>*tO922s^6IzjxRG^Z0;IU_xO3582 zWdHKZFK5r5rSVsZ&v+D*a^~#*sHpybiDk0!Y@ZmD z?%@jnp_8ffTp`;4jzHt`7oECur*FzJsie4;pY&a&#w*Oip)^v*oH3tGlDSSwFG#gnp00ebYd zpw&lwOja@>b#vby$I}#{XO6Zh>zLxQQ@pL5$P94-=;DcP7d?min{7JzUg^~uY^ zTmZ}3H9PU-20?l(BbwB;$4H6kZ3#U5NM=o%= zcyg$-ZdCsXTlSNbDJdzKbCHiE(7@LH!OLglBY=ljZ|d>#tF))`B4JGb7{$Aq+v4YR z;$15fvz(%WKzBdL*4n$1Y)PP8JnTCdPZF_X$Byj~x^6@x26z!a^MfvbS&FK6vn8mVq)sr4og!h{Ly)~!?LqY{r<)$`}i)iDeb1KDU2zW(}asX(s~e;`0zgco}{HlUA|-(X=+^XRUbNg8>X=CeTN*4Rt~oj zl;ZuR7N_i8m6r4}OV6mxl<1tSvL}Q6!4kr#WtVwQ!6PD->cOnj-zN!^QPIWaDu+g zU`I#C)YMeaUcVDlE?b@;R}}Ds^@xm^o12TemB|LuQA^<(v1;)Em873kDiy&bu;-xK zK~t6S0?+MVb{G+Pp1^4n7$QoSm)e(Es^}Bm*E+nz^1;|dTP91NaA{q8ya3WO)Tu@LT z+8Vvz-^T8gr9CMn&)s$5?(z?VjC@64>z8fuwo{n|+C%3y$`9{mW(EdTLG9YrbN6V` zwabzvOQ?8;f?>mqh={=25E_8!=xE3W=U1BJ>7BDX=?SU%*)($04OYOst(|G{#(Pw? zxyF-0<*Yt8BvTbU-jkeE{OD&LCi!p&s#JLv&8aztZTei&{f=%8sgcHjjio%sj5~Jh zfHfS`2cq3Qm!xdj)WeaFJN?L?_WP#()T9!5X%*XU@a|yBx1IJ;*{l+bp|x>vyI#9OzF+H}o@u@{b-p!sLKO5V695 z`XpwAn8?5i;^pOqB@Y&=4sH$3KBHECX}4OfpsEYxJXIJ>4QV5OAv_eQ@fZpW8Z@YD*RIq^7Ffs^bZ~2UK8bvrgvX!WX2iEQ zwS;lXQewuKYQRXk@-)raCFSSmU%!66dGqG%Xk;qt6V{KcBH8V1m8c+~jd9WsXG8osgv$LrtDTv2+Xl-rXs8J&W zW`YZ=l`B^QZRygb=!0RfGswh`L3uJO_3PJPuwVhr20^A)3qZI$ojP@D*REZQ7ca(l z_bS<7Qy$irFTVKVRk$087G7frX=!O_4n~X^(Y}3q{RFWokDi>P_l|zDUy%xHkYtP4 zuwla~Q>N(3P^0p3adEoxsup}|iwJo6^5xvK!K^%9CxXS5D_2^#Zq1!tHDVGi;))e3 zFeO2YICbh&tp&lXJlc|3vu3%wyT2})lZB3#A45aJ-vw33-dZy&k5&{5X<=buB-7Uj zz_^geEnWfkMYKxF@1Ht-{Hm4Fj_y7p@m$y-_tgcLb-9^83FJ@G(_bU~ef>L19ekSi z7|-(=MH&yYQxguZ7?JYgIoUy$D<5ECdBoPGBae62)r+R&1Phx9Wj6gVS=O8fcVUyJ zA=IcxvBYfHf>Z53TB@EaD@V^^3rmYJrwlxrIywJvPlcJIn10|yRNXI&FwTzvzeIhm9HE$xufYJ0EFUU%1Y z{?~2kfcIJvNj@!m&)=8Y$NPq8Z+HY%b0lAO_g$Cz*sVz6Ztj-iQfX3c!sl;{n5VA% z%a<=nku<;OX=UqHX&ww~+zon$ zY~H+?f?N$C*`8BnZ(O~^qhY%)A6K6C2|TIGG$g)Q#QP$2lJc8BoS*A%CcW+8PNp#e zGIzIComy{>jATr$36Y8SFPcithlHPC&x~uyouS=Z`TGwqZ|$Gu%P^9+wy-#1<%oWc zMjRH<#Ho2Rbo(_|?}{83$npTTZr#STRy(iOgDawy|Ex}T8E)jRzTtj_!V3Fkc0y>NLNU{E%ZtOBC{JDX}e`7={%bIdDO%I?W5p_8P15RB~fz zN99bj=w))IxbKPE^HoVaUa=88?y39%8%uN83}8Zg@4feEu5K#vqGmfU9p#lKb@A%C zGrp4Az$ucek4jaf27%Z#hxhNfbLX(|Mo5B5yl&mPuv=>tOt=lhW`IB-!%du5$4;Fv zv~iD+&1akh^psM$9PbFrW>h+X?#~)EKX~vU&F>1256c3UCF;~s&2u3XN!X_R*I$3Z zx0IF#c-0Xb!_<8*?kpPkz-`3JZVU!?H1DD^S1=kk13ofn4mF@Qiwukb6B85R@DEH( z7~y6CWO$)tO}cH{w!~`5H{X0Sefo5k37jJ1Z0NegBMH+P1zAbVszGyhwTX;kvp5W( zSboN5{M~oo!LNd*09JBnF+AmA2C8ve%%=NkpOy~ zU%GUOqa@gE2?Hoc{8r9B*a}V}*PXHNW5RX2X&u8C>#7L!6?4UuYTX~|^_ z&xI8W4xA@2aF{e{(yKd?(4^2c$9_NrI1|Oxxsj?2CcH#H_vxpfavwWUIjF-7nga=*I!Q@MH*em=aU4_{ z$_>XGToHRliFvkXD1y?DYw5MrbEf+x=HHXy!JvJ8>)rEsI z0+%gYhH(yF0X!e@_H02}so%7#YG@7%d#3`caGGRFlO zR$UD8@Y&;h0dl7&7ebAiW8Zs~jncSelm(}A=g!zXfP)<{v85d)%0k7UIcDjE$pM`? z3uTQ;BD~jO@EtyUIL9pt44PvD5X`Ai%y=_IW0V@kmpJSWd(WOddxjPYkL)i$gXZv? zBXWkf$QW5P9#6cg#2^nFc+lFT&1aFtpgD;>bm&kLeq8`^qJdQqbWhEiHPe)HX3epE z563BKqN&O8*rfyO9R{{}d3mUR7Efqq&9Mf`er`}rRvz{RjA~KmYuB#D@yXbJ%9vXQ z%?~bIyzlqDnfdt*{rxAeSk8EC6C~3yLLqKxvS5bwQiUGeC~$zAB`GSK~mHIU%2*fQ2Q4Br|)Xi_Z$2W z*}W#?HE1``zqWMu&Yy{xa?PKfI61U;pYx9HG_~A?qN`4|GQNrM{_5+v=Sj5G8u&PL zLfG?M)1mWDsOPQqdo!@yZDwYkpRaTnRBE0VpYVC#{&4hH$0lOz1j1_qb01I7*n|X4 zI5i=DRn-1;xl`??0U9&I=MRoJVQFQZlbxm1+|k=B(cMQQc16PBQqga}{r3L*?^i_B zxc~FRXZM`=qt%u{+&-11naL90bBC zMHt=@fpVF@H(d&LicQ5Hhl4WsvL;@18NHo z9(!2p?xzkR+eozugm(td6IrQn%b|^~o(_pm4&Y!(52g1!HPjc1Adhw~1B~;S*iq3) z7_JqXwn4+Hi7uDKqyj8S4W?7cRt_e-AT@Gt1! zvb9THQNxOZZz*KhhOa2%38j=M62wdxfqblX0Q&#>>#rEAS=g$%S@3OJ)Dk|fzlP70 zOHsO7b!|?Kb1`@8UidsXyYQ5i35^PNZ&(5l$zq(;l!Uib)A|IAs>B8jnm9bFCt z@mR^lC=KUT5+va(cBs>`#E58z8Y|>^GYWYx9Hk;2-ud`6D_;i+wOk@O_{YJPEn8CJ zssY5BFBV&9f}h;{tHaQ_v`9^lR`M63unC{>4jdyf_yY5Ek2+aGA=ujF#;H>*2-Tjn;rp<)hwnOgGFG!H zcU!;ihQ3*;?PrrGci;EN%75dWH7h^F%jI<0g+(d`H4p0DtJl>lb1prSxz=U&RTB#M zyTgJpbw}RI>-pHJlV~T;MlM@hDx_(YQrD-&<+Isw_o>4M8cxl5#9}ktjN5nY91{AA z#MFYCF@Ef?FP=Ui)JDAO4Tp6dA)@RszMv*e?JO*AB|pa+a`dWYGNTfE(uC{?-hfj8RYtQTqd*+op}J}LR|c~YT?o4Ku< zg&q8@xpRKKFjepS>y>Yva*G{-*c8RZ=@#zTvXVj^(T^XDdL=?}V|!0Ho$Dm7<|F^d zYyDwn(zDzQ{i*?z$->fN?4;uJ_`WtS9v1fc@u`7S|Jhp4FDbf`{7BueT8}yalbFos zxFoZ+b6uluEz%=z`jPxMYaQ2k-b#ztuS6KB@ZH0bboE7SBYO{=cFBnI8b$sGYrUwn zL`#(yUs=(1Z1f8`y`D_E&$;(qpoAVxK9?#=aoBZ$ldrj$ZnPqCUJKwgt96_zilbJr zVUA@}$*pv4U%9yTl>gnFQc*vO7f0szo_Q+Ovelf>4w|c+K$nQbxRSM*hRakdSl0Ca zhN*QJ;9)W^Dk_4l6<%s^C`15kaP4u^0_{RW2!gW<=BZfzC-$b~>9E_eG;!;MiD8%W zoO;f_x;V&L`&6xQn3Z+Ta$AtEo6<0)_%?XV%r{4E-oX z;2=jFJ)T?^$ORXwT-@4q?bj2Paj@u*qx!i?+vdJ4kd>NDvV0B zyVvlp;LLTb^I*Zl+1VK!;Qa`hXyI`{GtP9zv8J>}1>55^8ez(FtH?dSc;fJ#iJ}G{ zMsNGLnGP%N#J0M z8oWAjGBsKSc0VPpe%@YIPMD@sMOTL9@jrYpCh~p>Ut(dGeq>DVBV;mP+F|MMlLKk1 z&}6a_kO)Nf7Tnu~O0+fSL3^#^L&UKXs566P8td<%^J%WvkP= zTiCtG&&twx0llF1=X;KQu1=>t#?jO&ucWYE*TT^oledh*90R?p%8WRfDXaV)-ERulTMOe6YN>TxIju!@$*pyt&OV$WYBgo+ zx4{;iheSMl^frVh0;3UY6!#pBu_Eh=bBF3il6n44N!nKiOkIaz5V zs?q6ewT_Nl*RvV8>xwP|o@W@MlbZ~<({jYLGqnixX!Jss_26Y9NaQ+x_%OdaNG)F$ z-49I^ zcF$qEEG86i59fZPnehr!C(dlOjt>zPk1M-5a4i0N#jyS>D2RNku5-3bZlRGs>Klg) z8ss)^iqt9sVRFD>*swt3YH9c+BeZMR4xZU8s{>V(Gj3uu)(SxmGbfx1aq5UhZlb2R z;%tJ(Wab2Pjo^vnWSEm*lUTUk;~*f6@^IE7!o`ahF*3z)PgmP>)vK$e){!CHB7mbY zG1!KB2FV}Jb-`*44i2rdPTe5c*2DyjJT8|&z{N*fw{FFSI#>k4a1ZulO(fMGuf5ih zMHta=8ZPGUn8<0`!NNro)c1=QF9@SL7yQ=?2RA+;V7P}f(9r&4xQDI&D7qS`1*cj^ zzR{Xt%!2V3PG-T34&OcZ>c_hen+$qE9Mn|pyGD#%2N>?*To`Qd!f+1(EdfqMhB3<6 za&_X*wbluyaEK?4G((l+`~$3{VS5scEf~c>JpuUAU=8WibgOl;yzs5#x@|Q0qehLwsn=xuTYy>uc>B@FqpXPG z9_&2c-rg7hR{M5XJ!>6j;G@d%M!&6cYBe*$Gzy9(Ds`e<&k<6-4S; z>uBDQIjoabK^FfrBHk$rIMxZ)y*YE{V7P~kxW>D@o2A_Jsdb#$iJ>@{!efbr_1h%I z5`e2GhI^R1VyuViM@xV+e~1e(81an}sZXur94BN2775(TZj3y=2|RH?(zoA!OAPmZ zFAJFUjT||$YKD9I*E-7qUT;FR8AL~7xCd5nE<)AA&Vz{`&igm^a8Ku2$HD{Ayu%yS zAez5rf)YRrj{sW;W_lRzVVI2Jo`;8rkqSrWS|@J*!2BB%aU# z7?v>%_nI_mV(TnM)0SX+h!?5!c z#e%k0)0sZp%Z+QT_wU~yc6joLa7@;*MWIKJ9^56p#nh<;$l+d8R1{(4!ABn(8;j*> z*hpX;(}^1mnQN`%D~CN9hi{YgIQIfg5c&8PFh&VrxHo3Z7#!Y<<(%8MZxbT0Hv!(| z=>GMV$hFoX0GALFws*idkQ5D(-csKZaxVd_b?x83A1>-t65NUjJ3mZ4crB>$9AVvP zaPBVLhH&4>TdN$$@_ltQI31B^tQ8j*V^iOeBS*0Ag&r2?-p`*uUk|H^ZnX~QK-lAk z4IB3A%Dm;EM+so>0O7Pr(Bc+Y+&k2*TQ~S7>Eyh{R_j|COy2ImHarkd+()-!r=hNSQHGko^}7h z#2&q(6XU6WPO^$@y+E?njexV&keBXgFySA=&?)7Rn32xD|+>3GE?zk%-j4w!kvj300aOR{ERjT#p zXV16q+9g12vdY$Jm0KMu9y=`L@disR@uP@M46pCElUI8F%9Sd(yO2Fs|AEx8Ga|RH zp4a7*G*Hs-f1Hm6J2h}$lPG9QM^+5D~^aTVHRjr3MX?FJEg9}czOe(x$bQJ@9 z_|wd_v)+QK)~|o}{(YSW8v056k5S!=dDg6P zy)K`8@8qT4G0U^hO+0w;MyYkb;Hh7^S8UWNeiV6N)4p7RU-QWunz*PwX0l}bdrf(I z^Qf1XcTCJ9W%6e&0-Os9!)%>MjLLxg_{)S3^GZxj98Nphpxu|3WDQP9JtY^uYv;L9 zD%7-(btGS$B9OqggK=hMhHv5%ylCJo3O94D-vB!Dz|KMMygQAl%29LztBz|2ScOIJ zh(5WrAaL4TPbshPPI&asM+?k4wx7D##|j@EbExGX!jmO`@WBV7xCaG6;ZW%gs=nz>O zaH8tftJk1GgJ9LB&3z(QZ!5v2c`bX9$Z7!J)PLRpD~E#{cAtqXXwg}I=kWO=+xNOo z`O>xG(^n%YkML+Sd~ZZ|(SfL_AXBMS+ex{#%-!ya(>kPNew2p$watHUaUF^w9EVW5 zs2D~)4v23S3QVeHcUb&Ps*NiY`t<38_u$Ow)2_a)$aqbF$UC(EG{>VWR$n{B6MBs8 zGjX&nhZ^DY1^zADiE;~zv{pF;7Z<<@7%#I z=7lV|Q$}uvztG=-RhHZ_BWqxC@e(JuPZ-7xm6(||_S-OS9C<%SrL;Hc0nTw4 z-(V37#x>M2?34xPIzR{iai3gGBQ4q`LJR)(iZ& z?mGtXRze#sh~0hW%4wWNfRC`eTgGB}c0vB|Aw%XJJ`!!^X!#0Jcw9}j06TxdxQM=PhlebTuErca-SA2`lV`aIaL z-GBzcpNcdbo1@O!&ZO}1>9yxhAK$TK$D-U^DuSxkiCBD{+js5+jD6Ts=g!@=b7vPX z?{`aio%4(A1@y^5j*aytU$1>RSPo$VZz&UKLHNd<1(TuGwJwGpFVYTbVSj{jT?`r z-ateHGD_+A1hTSf4tP#3%*rV(Z0p(}B_+kz*O$nr{#N53IB;OVfC1Dp#OvF-;tvrZ z&!0bk-n@AUrK+chJ zIeaTitCG?Z4QT!P^*ea*;K=B|MOL=kW`~Skc2ba?lfzRyBz&NWEJPZRQ4|^)ieD~I z0X%BiN&E_2TwHSTzcS?K=Rb=w|37QiEEN3wRT!AZbLlF%DA4@=-Oo#F=-|a%nKD;X z{}a!DYEiC(g>@TOf5uvv@zFhQZf;={MpWL76*L}~S60l3jLaJNWUmqrDd+{{5Sg;Am+hNzUY`8G)+tX)gv*OMCY~MWmG)Y?6Of;v^TEBUSVniEG9r zN~`&21FZ_9CqDVUEAa8=;WzW|{Y{!QQKg`oBeW2|v+&rq|ITln)lu=<%B4lgd>g;s z13B8>-~Idd4;(m9fjqnW=wrluWz zbX~@gBU9a1Rvk~;$xod+wXCcxDk{p;)ARS=e;+br2(>Qfch~Hk(wLf~Q(P_iNs}gh z_uY4>Yt5Q9WQ`nIRE^w$=f)26?=M6DLkI zgenb}7>`lsh7B8NMMSa&dBEvCfHT%kWD>N0`st_m_;_0V)#P*UCuE+C-xnJ@Xq@6x zY4q@}GB|tq6&K{Br6xN$Ie~7q`(niY<{6myLbPx3R0FL z=lMe6LI|I>OE7o$F5;VqG)Wur_)>`i#7N=836m=VM=2_MMjO$8V?R(dKGDuCTC`|v zZL{~^i!zx^0cX^(X=BwG01Y)KC+FeAhxO~%=ghjP{}&JJN#(NJxsRGT`SQ3#bLz}; z^^>nFD*4en`S>Pq8OLC)b?Ql;q)igg<7*3CeaI)|rxR4S4D9=NwmjnG;f{*emXlPF z>2K#o6tVo(f&A&yr?+q4j(Uk%I3n-wKlGDdI6)i9WTTV6d-v`)-+a@(dv^?3U6lc3 zu4FxYT2_R}iq5T%KD>em5$r_bXHmBQtCc=8Gjrj>h0rx`bv1GHzVq}kHS^xZt9m^A4CFBjP_;-|$g|ftdeqdc(M`H?Kqns&5wUvpYE?E_=70kY%7tnrbn;eK zR;oZcS2_lB)CG zd+*)3b4Mr?1_T7Cr(W~XJAeMXvB-lYX3S9gckS9WV#Ek_9Mq6r9gyDnQ2Rj}6Ep{iDS(O)M2^f^%vTk0Z@ePBZQ~cu+z0i@UnIW@Tjo z_wtQ|D%!FXnY@(4Z$vO+ZEY>+t)va0!?yf&!s^8Xbdr8lsaQlMA+HasA39x;&-2~! zS=Z4qrwN#jA>o1?nbcWnsgB}#3@T3Mh#L@PdRx2*bo}P#<`VPB2M?0w@5vzZ$;g>w z2L%h;wOo1m7P1c5ia54av{!U)x0r;|H$u<`c7q1~9C4$S*A|6>N>$7v$h{ z+}4?-Qta)ycvsbnL1r2l+4@;0Jnhsk7lb2twRv+$>%b(pM&eWX@}*C08SW7wuorQPo4)s&iPinytShtBx(&p-F?-@i|v zKE$|tDoru6sfI)EcmJL{!a#EKrozZV2H5^#5tu!DwvCOA z439NENWy(PuDEsWccwKQ$WI41c3OYoF$c!tHTc0q220p}etuZ;(7;|-uO{)2shuzF zU9SV=i)h*sSUopg`Z8QZq6>NaRxW;I(nJxRh2OSq+nFkkCPR*my7j!TQAL(WCX_1pNu?M}E2FdUi^M ztWvcSCkq=`ORk^$-@4t4@xga+a8QLYVh$_eVO+-C2Nos^3k$<(+l$0_1vJz{JX2_B z3>M?Zj}HzG)&Oisk#^w=3kzq?oJpIK8A6gJE@s&vigw48G$<(O8EYU!IVV0E3-TCVrZPhnE_d z)@%(kTq>PvAY7h~A3y%fFTcR-NcFGNsU8AYlkl*)248<(6vIf!Y3sq=~`Evz*9#=#8<9d$({{n z;qf>TRb04mp+kob?5nFoNMb~UpUdsrw=p74pFUk{Lof@Ev1HDiIo{sh&x_<_qvPR+ zYDoCIpbLr5Bw0#}Fbj`S6v`ad4QbHNMG|@;!6zO8&5LL?gx@!P=EOCt<6ON11i~oT zAon&zE#DR8{lJqw%*lB!Bqx{*4C*F!324)IA}3%hX*?{*N;$AvwGRk zp+i-xHisO~z+gyDa`M0BT{GM7=-0#V=DHqty_OGovprE1(60Z&y;%eNuL$Rc6 z?c@1UmuX0Np@8#g#1zFl|5r(ox0U$1i#J)u7?F5;t?AxjOH2%7sX1h(-acz7wjMG6 zsAgt7D{qYK(>^F@RMlwzxI_X?-rmOMsGTdmb2Q}0;FfM}TH~{CzH&d~xKvV%Xon6R zE$emm>oELTw&H(#p=4Xs5ZIT!y%lR!2z9euMbOTadIT9y29QMNj1IqYVv~CJI@{FNYi@yy}mQKhS-29m#J!{)u={^nM87ILj!ZL z86Xf`!cCmhz)6+kh!%arx14n2(MzN<8J-ahn^EZqx~GjJIaEMwQ~u|lf8bk6;{jfE z;o;%PRQ0{Mv1I5SuhFY}JJL{)#PBXqx`L?|87s#i$ssT%jBv96H`sVQY}l|)ojMV# zC0~B|C04RDOyK0X@e$vp9!iNHb0eAmi2GeapIaE8IKiVM}aLA+dhZq)Z1^rZJ^i$Q=L2-SIDH7 zvy!ATsfaJ&a(I<;DdNqXe5{nGNnr5^4op~G#_Bp-Idq3(B{>Faczj@qSNHmvF))U` zMjTwhT+aFS+i$V>#d4#!nmV*(YnL3KF8CX0UcS;oM6Y64nqy*Ow62`#Rks33yX0Se z_0?N%y`?t+LlEKE0`ENtzjp0fYQ#Y{OY{Cyztt?-rk@=^}3jWY-o`jI~dHtmkJYJ;yd@@haa|Y z-`+r(33N@86Dwp!@z*k(R}T*lI7z^_37w&n)~t@SCdqg1+-a-^R!3f^eC%(;-aa_6 z!hZ@MJe_b=fi+2f_3BlOE2=nkm_zss!AU}~yE7`v5Kc5nj(RmrunnPVATE?V_>aO# z$PoWgwUXm|W*~niMTe^t3(;H+an3AxuU@@Q zoH(Hl$DH$Va15+>P;85fi_!lYOi9d=V-2)fGn3nWP1J&Rc6QKe(dX;eug4BlY(Hfr z#31>B#Y^}8z9+Awq-jvl)Kx37V8)1P=9!moCB^PKbnW6XOliI9H<~?b#?;9lQF*3F zE|WF0wizt4nrY|MnE*L;@p0=8opYW?6bT_n1+d(sbX+H2v{1L2P zl5m=I9vW0%yldxnBBx67M@Ns1>_6bNt2a#~Z@%E7TfN*bqy0bsBI!vQt<(&YBPT^Z zDY6{7;HWBZ9p0RU6TGdgtV>FYQL2!z8slb{Z_$hIpK1LM<2Ddvo9>jS4i&a@0aSW-5uvIso=NYeyb}Bv|3>(1qcjmwSkgz~Jx>jd`2x_Roy$;AP=JXY6^x8y zWlG(Ns(E>N_~eq1)QE;6a>E3U#3W9|L=$YC{!cl!qduj- z%k~}(1x=qFd`m&YLXNzg!xvMs$S7vQ2xPL(Bar{sUw?(JX5*;hX2ErATaW*s@j5OO zmx3IX?%I--%XxFYC zC9e*m&d$zQY@rc;c=eyIBco`U>XLS{3_;|ik2(8~kPuv+b&e0Qc2%{lg)BlOvRT0T zx)vL3sEUouh#7NosJtrU&8M7du9At2BU@$*qSmh)w#{y%qdOA-J@0l=AwA1%(%-}^ z^l8$u=Lf5_a%5v6pEPNr;;Rt4cwzd|l^%RWTtf)6x43frxCVeaFWUG`dGF;>fjw2bJr%#Df2VMib~MX_*jf=@-nY02sskj_&po52xjkA9Bm?72++v_Ej}35~gb@qxJuFEVGZLFS`{1-oMwsU)^9Ahna%rWO zF0Z(-s?$et88Ui5S@fBA>+28^J)2BRttc<$ zlvm1R0`-p_32e3z^X+wfM{xxg4syiNX3 zse&f&#%+7QwGJ+{_@b9qRy^;TeBJaqc88rhbqXiyn*Eft z`g!}=xnY`4eROKZWy#;ZJwE1kC0AtQlyhi&|3hRkSKM{^?^8o)tI%YzA|m1moNd^T z3pH@fCsBL7juHFZxpRB=>>;kyf=$13=gxRpVJOzAIcRUTDz3MUQ$|UCzS;}udG$Zp zedH5WJncEImUhLJWv%T!w1b*AL~Zms989t3Pk5^0v_&{8s#`)gR$H}o^~dBbx2%xu zH=YGyH!DS#zoz?5;VkN|Z4eXoIy`HzyBL1-kg09kwqwVRHOkEqEL*6ROa0u^qU+ho zxHgC-iaM~eyke2HrLCP+bLVO&D|NunMZQUT9U2)tk#QRti2Nq;?-9V)`|E5DUey?xdy!qnAi*Reg zmfBXWTA7Y@mNmPJ?eT}(#ghxN9u?(4!$(}*y-_&O$yTUXVaqQscC)e%u=fx-2{>L& z`*rAnBcyOV0{rDLkHD{Hm6IiJbp>Y3UdO~5XJ)`(5_`KD-J9x4NYKS%c>{ZQ{1H6h zh_-@?lH!W83Tb79Tq@x5ggk**Xkja~v=%Z?+$jNhYlUWM-KNHzEoXKJ{%b2aikyu zEVE;Z^D1 z>loYNUGN@}#;4JCDD=KNP?^pXTV<>sTkn5j}XKanO0~x90OjkyQ6!vV^)fW5x`q zd)N(({h@Uwh@QO;n+!O{I%!?;{XEirO&PEp1?%3tdGny|VIwYlLuy|ZeR>^dc4lN` zfLC}d)jr41J4Lv9LfylHBy>G=KSlza`9oZQQEQ^<)9X0L3H-q#0b^XPG5UfL!U0J* zXq!;?epd#}`o@eIQ#tW}?#1H5Fn_At|x!1Aq zKn(BjMm1e?ylgdNgh#+BwV3Ha-GiD8bK?gfS*GWjZEiWY#nH-7wh9NvrNoa@)G z6C~hZj(r69{Oe7TZLfns_wL3sd@Rb|i39Q&a!=;}6^{2De?E*1$CNy0p0X*V&){ z_}g#CGt)i!f*_%Vzd&fug*ZM6z1KT-!H?el~*n8`oTrMk@6c*;D=A_zul z$39=i24U#+{M&aX_w5&#n*6z)(^zZ!K0<4b=9835WYVQ8H%CTC73BqY>a=;+FXrq# z#MlbvoQa(~pD+1gW%P8PX5HGq5z%$@H(eRob@X)dZdAWzSlibp5#hg_zH~H0x9XaEnBo|)M=<5{Y|UN zw?z;9&04%lJTL0>=~K`El)WyK^>XuQQC@k}tsV_Qza$R&*kNJ2_20`)8#lgr`?gMs zhJF-ZFsxe%k6X8B)boRH-27BO%(8-0lMfuYBDH@bZ2ITkmQ|ux+>hD6d2bP~S(~XF zTY6N#%w)?1^_%wS>R~@W|Ad75isFyk2D_J*MLN2X5H%yRzn`YOTU=>r;c~*&0pq@` zGJkkx)^Qnsu#@j5F<)V!WeQ`s!c3kBwjJo1Oyw_E9=vF9g>08DU9`Rf_>hP62z_nv z45lteqw(yztsi0+8TWJCvE`*9GopOOoU$A9Xi*arc=eRM+EGRA&{HHa~S5 z?pCk-FL2Ay!Gj0M-hdNTzkdCO4I2ilHf`<`vvyl0F3oG#j|5hUa4mxt46$=LuyNPP zn9{aAWH7%wR>a0hMhk%R!U}ONYcCK7UlA+*SmWT=X0_o>D_Ykho)C7bMyLuL2n@O z6cMWy%{ES$K;F+$qwHl$z&Q^34HmIrTtgqjz6nz+V*?E68#?EUd5lGdXY#o6tC!9S z!iNpBnLOoJN#QWDEmNhkbNQ)n7M62`j(@p12P?nhu40d?g4{k?xwF%g6I|;pu~4f% zTncJ&7xaO=d12=6LpBPguNjlA&OV^ z&kN_91~zOXvIxYMZhGK#viD85r{`LToKAW>(Q|92^SwRaI(6*NKmVj@D>b2Br5IY^ zf9K=t`}y0s)^qkTrkk3hCcFtjzVg@dB&$=JnI_$_n ztjxmoQ+Mo_9RnLA(x0R?Y}~X_$omanGlI7g#%NyRu9Fu|q@<+aC9L`^W3jxTv}DwX z5exo06ldpZYrr9?l}?VM9mM7NIXD3HmD`2h&o5N^?9B8PE54q;pnvFxYaBfS^sdT_ zoA_O{#>Bn-opI2GpFI$8$EZ< zEi*WDiYnfV5&u9L;Gv3Hz@kNqZrr%pGPoP=?-Dr$@~m7$Vmo-RiZ~LBiY!Ui{RcPB oo4Hr_5c6? literal 0 HcmV?d00001 diff --git a/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb b/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb index b3b93c22..734aa7ca 100644 --- a/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb +++ b/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb @@ -22,25 +22,7 @@ "metadata": { "scrolled": true }, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/plotting.py:26: UserWarning: [Errno 2] No such file or directory: '/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/../versions.py'\n", - " warnings.warn(str(e))\n", - "OMP: Info #276: omp_set_nested routine deprecated, please use omp_set_max_active_levels instead.\n", - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/gpdc.py:27: UserWarning: [Errno 2] No such file or directory: '/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/../../versions.py'\n", - " warnings.warn(str(e))\n", - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n", - " from .autonotebook import tqdm as notebook_tqdm\n", - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/gpdc_torch.py:33: UserWarning: [Errno 2] No such file or directory: '/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/../../versions.py'\n", - " warnings.warn(str(e))\n", - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/models.py:29: UserWarning: [Errno 2] No such file or directory: '/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/../versions.py'\n", - " warnings.warn(str(e))\n" - ] - } - ], + "outputs": [], "source": [ "# Imports\n", "\n", @@ -70,27 +52,19 @@ ] }, { - "attachments": { - "image-2.png": { - "image/png": "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" - }, - "image-3.png": { - "image/png": "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" - }, - "image-4.png": { - "image/png": "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" - }, - "image.png": { - "image/png": "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" - } - }, "cell_type": "markdown", "metadata": {}, "source": [ "# Background: Pearl's causal effect framework and optimal adjustment\n", "\n", - "A standard problem setting in causal inference is to estimate the causal effect of a variable $X$ on $Y$ given a causal graphical model that specifies qualitative causal relations among observed variables, including a possible presence of hidden confounding variables. This is different from causal discovery where the task is to estimate a causal graph from data. The PC algorithm is a typical causal discovery method and the PCMCI method and its variants PCMCIplus and LPCMCI a modification for time series. Once a causal graph has been estimated with a causal discovery method, one may use it to estimate quantitative causal effects.\n", - "\n", + "A standard problem setting in causal inference is to estimate the causal effect of a variable $X$ on $Y$ given a causal graphical model that specifies qualitative causal relations among observed variables, including a possible presence of hidden confounding variables. This is different from causal discovery where the task is to estimate a causal graph from data. The PC algorithm is a typical causal discovery method and the PCMCI method and its variants PCMCIplus and LPCMCI a modification for time series. Once a causal graph has been estimated with a causal discovery method, one may use it to estimate quantitative causal effects." + ] + }, + { + "attachments": {}, + "cell_type": "markdown", + "metadata": {}, + "source": [ "## Causal effects\n", "\n", "In Pearl's framework the __causal effect__ of setting $X=x$ on $Y$ is a function of the interventional distribution $p(Y|do(X=x))$. This distribution is fundamentally different from the conditional $p(Y|X=x)$! The basis of Pearl's framework is the assumption of an underlying (but unknown) structural causal model (SCM) among random variables, for example,\n", @@ -103,7 +77,7 @@ "\n", "where $f()$ are to be understood as *assignment functions* by which the value of the random variable on the left is determined by the direct causes (also called parents) and noise terms $\\eta_{\\cdot}$ on the right-hand-side. These noise terms represent further causal drivers that are assumed to be independent of each other. In the associated graph an edge is drawn from, e.g., $Z$ to $X$ if $X$ occurs as an (non-trivial) argument in the assignment function. This causal graph is assumed acyclic and represents the qualitative causal relations:\n", "\n", - "![image.png](attachment:image.png)\n", + "\n", "\n", "In this SCM $p(Y|do(X=x))$ is to be interpreted as the (interventional) probability distribution of the intervened SCM where the assignment equation of $X$ is replaced:\n", "\n", @@ -117,8 +91,13 @@ "\n", "\\begin{align*} \n", " \\Delta_{yxx'} = \\mathbb{E}[Y|do(x)] - \\mathbb{E}[Y|do(x')]\\,.\n", - "\\end{align*}\n", - "\n", + "\\end{align*}" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "### Conditional causal effects\n", "\n", "Sometimes, one may be interested in the causal effect *conditional* on values $S=s$ of another variable $S$ in the graph. The __conditional causal effect distribution__ is denoted as\n", @@ -131,12 +110,22 @@ "\n", "\\begin{align*} \n", " \\Delta_{yxx'|s} = \\mathbb{E}[Y|do(x),s] - \\mathbb{E}[Y|do(x'),s]\\,.\n", - "\\end{align*}\n", - "\n", + "\\end{align*}" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "### Multivariate causal effects\n", "\n", - "Multivariate causal effects of $\\mathbf{X}$ on $\\mathbf{Y}$ are defined accordingly. Note that one can always decompose causal effects on a multivariate $\\mathbf{Y}$ into the individual effects on $Y\\in\\mathbf{Y}$. On the other hand, an intervention in a singleton $X\\in\\mathbf{X}$ refers to a fundamentally different experiment than an intervention in the whole multivariate $\\mathbf{X}$.\n", - "\n", + "Multivariate causal effects of $\\mathbf{X}$ on $\\mathbf{Y}$ are defined accordingly. Note that one can always decompose causal effects on a multivariate $\\mathbf{Y}$ into the individual effects on $Y\\in\\mathbf{Y}$. On the other hand, an intervention in a singleton $X\\in\\mathbf{X}$ refers to a fundamentally different experiment than an intervention in the whole multivariate $\\mathbf{X}$." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "## Adjustment\n", "\n", "Pearl's theory allows to utilize purely graphical knowledge to employ criteria to characterize whether a causal effect of $X$ on $Y$ among observed variables $\\mathbf{V}$ is *identifiable*, i.e., whether the interventional target query can be written as\n", @@ -159,9 +148,13 @@ "1. $\\mathbf{Z}\\cap \\text{forb}=\\emptyset$, and \n", "2. all non-causal paths from $X$ to $Y$ are blocked by $\\mathbf{Z}$. \n", "\n", - "A causal path from $X$ to $Y$ consists of only directed edges towards $Y$, all other paths are called non-causal. An adjustment set is called *minimal* if no strict subset of $\\mathbf{Z}$ is still valid. \n", - "\n", - "\n", + "A causal path from $X$ to $Y$ consists of only directed edges towards $Y$, all other paths are called non-causal. An adjustment set is called *minimal* if no strict subset of $\\mathbf{Z}$ is still valid. " + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "### Optimal adjustment sets\n", "\n", "We denote an estimator given a valid adjustment set $\\mathbf{Z}$ as $\\widehat{\\Delta}_{yxx'|\\mathbf{s}.\\mathbf{z}}$. Estimators of causal effects based on such a valid adjustment set as a covariate are unbiased, but for different adjustment sets the *estimation variance* may strongly vary. An __optimal adjustment set__ may be characterized as one that has minimal asymptotic estimation variance. More formally, the task is, given a graph and $(X,Y,S)$ ($S$ may be empty), to choose a valid optimal set $\\mathbf{Z}$ such that the causal effect estimator's asymptotic variance $\\text{Var}(\\widehat{\\Delta}_{yxx'|\\mathbf{s}.\\mathbf{z}})=E[(\\Delta_{yxx'|\\mathbf{s}} - \\widehat{\\Delta}_{yxx'|\\mathbf{s}.\\mathbf{z}})^2]$ is minimal:\n", @@ -191,8 +184,13 @@ "\n", "These results were proven to hold for linear estimators. However, in extensive numerical experiments it was shown that the $\\mathbf{O}$-set or variants thereof typically outperform other adjustment sets also for estimators based on neural nets, k-nearest neighbors, random forests, or Gaussian processes.\n", "\n", - "One such variant of the $\\mathbf{O}$-set is the *collider-minimized* $\\mathbf{O}$-set where collider nodes that do not block non-causal paths are removed. This set has smaller cardinality than the $\\mathbf{O}$-set which can help for more complex estimators that suffer from the curse of dimensionality.\n", - "\n", + "One such variant of the $\\mathbf{O}$-set is the *collider-minimized* $\\mathbf{O}$-set where collider nodes that do not block non-causal paths are removed. This set has smaller cardinality than the $\\mathbf{O}$-set which can help for more complex estimators that suffer from the curse of dimensionality." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "## Estimating causal effects through adjustment\n", "\n", "We focus on the average treatment effect defined as\n", @@ -213,9 +211,13 @@ "\\mathbb{E}[Y|do(x)] &= \\frac{1}{n} \\sum_t \\widehat{f}(X=x,\\mathbf{Z}=\\mathbf{z}_t)\n", "\\end{align*}\n", "\n", - "where $n$ is the sample size of $\\mathbf{Z}$.\n", - "\n", - "\n", + "where $n$ is the sample size of $\\mathbf{Z}$." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "### Linear case\n", "\n", "For the linear case, but here assuming a multivariate intervention in $\\mathbf{X}$, the causal effect wrt to $X_i$ can be written as\n", @@ -228,8 +230,13 @@ "\n", "\\begin{align*}\n", " Y &= \\sum_i \\beta_{YX_i\\cdot \\mathbf{Z}\\mathbf{X}\\setminus X_i} X_i + \\sum_j \\beta_{Y Z_i\\cdot \\mathbf{X}\\mathbf{Z}\\setminus Z_i} Z_i \n", - "\\end{align*}\n", - "\n", + "\\end{align*}" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "### Estimating conditional causal effects\n", "\n", "Conditional effects cannot occur in linear models, so this applies only to nonlinear models.\n", @@ -251,9 +258,13 @@ "1. Fit inner expectation $Y = f(\\mathbf{X}=\\mathbf{x}_t,\\mathbf{Z}=\\mathbf{z}_t,\\mathbf{S}=\\mathbf{s}_t)$ using the ``estimator`` model on the observed data (indexed by $t$)\n", "2. Predict $\\widehat{Y}_{\\mathbf{X}\\mathbf{Z}\\mathbf{S}} = \\widehat{f}(\\mathbf{X}=\\mathbf{x},\\mathbf{Z}=\\mathbf{z}_t,\\mathbf{S}=\\mathbf{s})$, where $\\mathbf{z}_t$ are the *observed* values, $\\mathbf{x}$ the *interventional* values, and $\\mathbf{s}$ are the conditional values\n", "3. Fit outer expectation $\\widehat{Y}_{\\mathbf{X}\\mathbf{Z}\\mathbf{S}} = g(\\mathbf{S}=\\mathbf{s}_t)$ using the ``conditional_estimator`` model on the *observational* values of $\\mathbf{S}$\n", - "4. Predict $\\widehat{Y}_{do(\\mathbf{X}),\\mathbf{S}} = \\widehat{g}(\\mathbf{S}=\\mathbf{s})$ where $\\mathbf{s}$ are the *conditional* values of $\\mathbf{S}$\n", - "\n", - "\n", + "4. Predict $\\widehat{Y}_{do(\\mathbf{X}),\\mathbf{S}} = \\widehat{g}(\\mathbf{S}=\\mathbf{s})$ where $\\mathbf{s}$ are the *conditional* values of $\\mathbf{S}$\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "## Estimating linear effects with the Wright estimator\n", "\n", "Sewall Wright (Wright 1921) suggested already hundred years ago to estimate causal effects in linear models in a particular way that first estimates the so-called *path coefficients* for all links belonging to causal paths and then takes the sum over all causal paths of the products of these path coefficients. \n", @@ -273,39 +284,58 @@ "MCE = \\sum_{\\text{causal paths through at least one $M\\in M^*$}} \\prod_{\\text{link $i\\to j$ in path}} \\beta_{i\\to j}\n", "$$\n", "\n", - "This is also implemented in the class through the parameter ``mediation``. For ``mediation='direct'`` only the __direct effect__ in the coefficient $\\beta_{X\\to Y}$, if non-zero, is returned.\n", - "\n", - "\n", - "\n", + "This is also implemented in the class through the parameter ``mediation``. For ``mediation='direct'`` only the __direct effect__ in the coefficient $\\beta_{X\\to Y}$, if non-zero, is returned." + ] + }, + { + "attachments": {}, + "cell_type": "markdown", + "metadata": {}, + "source": [ "## Types of graphs describing qualitative causal knowledge\n", "\n", "### Non-time series data\n", "\n", "Qualitative knowledge may come in different forms. For example, one may further assume that some variables in the __directed acyclic graph (DAG)__ are unobserved, here L:\n", "\n", - "![image-3.png](attachment:image-3.png)\n", + "\n", "\n", "Another way to represent the presence of hidden variables is through an __acyclic directed mixed graph (ADMG)__, which would here be\n", "\n", - "![image-2.png](attachment:image-2.png)\n", - "\n", - "An ADMG has directed and bidirected edges representing one or more latent confounder variables.\n", + "\n", "\n", + "An ADMG has directed and bidirected edges representing one or more latent confounder variables." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "### Time series data\n", "\n", "In the context of time series, we consider time-dependent SCMs. An important assumption is that of stationarity, i.e., the SCM does not depend on a time point $t$. Then one may represent each variable at different instances of time as a node resulting in a __stationary DAG (statDAG)__: \n", "\n", - "![image-4.png](attachment:image-4.png)\n", - "\n", - "The stationarity assumption implies that this graph repeats into the past and future. Knowledge of all causal edges at time $t$ then suffices to represent all causal relations. The causal link with the maximal time lag (here 2) then defines the *order* of the process.\n", - "\n", + "\n", "\n", + "The stationarity assumption implies that this graph repeats into the past and future. Knowledge of all causal edges at time $t$ then suffices to represent all causal relations. The causal link with the maximal time lag (here 2) then defines the *order* of the process. In the above example you can see that a feesback cycle in the summary graph (left) actually is still a non-cyclic time series graph." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ "## Limitations of currently implemented methods\n", "\n", "The theory of stationary time series DAGs *with hidden variables* is much more complex and currently not treated in Tigramite. However, ADMGs are treated for the non-time series case.\n", "\n", - "As mentioned above, the theoretical results on optimal adjustment sets have so far been only established for linear least-squares estimators and singleton $X$, but the numerical results show that the $\\mathbf{O}$-set or minimized variants thereof also hold for multivariate $X$ and often yield smaller variance also in non-optimal settings and for non-parametric estimators such as kNN, neural networks, gaussian processes, or random forests.\n", - "\n", + "As mentioned above, the theoretical results on optimal adjustment sets have so far been only established for linear least-squares estimators and singleton $X$, but the numerical results show that the $\\mathbf{O}$-set or minimized variants thereof also hold for multivariate $X$ and often yield smaller variance also in non-optimal settings and for non-parametric estimators such as kNN, neural networks, gaussian processes, or random forests." + ] + }, + { + "attachments": {}, + "cell_type": "markdown", + "metadata": {}, + "source": [ "## References\n", "\n", "* Pearl, J. (2009). Causality: Models, reasoning, and inference. Cambridge University Press. \n", @@ -911,7 +941,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Hence, for this graph the $\\mathbf{O}$-set yields the smallest asymptotic estimation variance for *any* distribution consistent with the graph." + "Hence, for this graph the $\\mathbf{O}$-set yields the smallest asymptotic estimation variance for *any* distribution consistent with the graph, at least for linear estimators." ] }, { @@ -1099,7 +1129,7 @@ { "data": { "text/plain": [ - "" + "" ] }, "execution_count": 20, @@ -1136,7 +1166,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "y1 = [[0.73288064]]\n", + "y1 = [[0.73288078]]\n", "y2 = [[-0.00799763]]\n" ] } @@ -1398,7 +1428,7 @@ "$S$ may be a continuous variable, but here we consider the case where it is binary $\\{-1, 1\\}$ and defines two causal regimes:\n", "\n", "$$\n", - "Y=0.5*S*X+Z+\\eta^Y\n", + "Y=0.7*S*X+Z+\\eta^Y\n", "$$\n", "\n", "I.e., for $S=1$ we have $Y=0.7*X+Z+\\eta^Y$ and for $S=-1$ we have $Y=-0.7*X+Z+\\eta^Y$. $S$ also causes $Z$ here, and $Z$ causes $X$." @@ -1492,7 +1522,7 @@ { "data": { "text/plain": [ - "" + "" ] }, "execution_count": 30, @@ -1521,7 +1551,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "Causal effect for S = -1.00 is -0.70\n", + "Causal effect for S = -1.00 is -0.72\n", "Causal effect for S = 1.00 is 0.70\n" ] } @@ -1566,7 +1596,7 @@ }, { "cell_type": "code", - "execution_count": 32, + "execution_count": null, "metadata": {}, "outputs": [ { @@ -1590,16 +1620,6 @@ "\n", "Oset = [('$X^2$', -3), ('$X^1$', -4), ('$X^1$', -3)]\n" ] - }, - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" } ], "source": [ @@ -1652,7 +1672,7 @@ }, { "cell_type": "code", - "execution_count": 33, + "execution_count": null, "metadata": {}, "outputs": [], "source": [ @@ -1677,20 +1697,9 @@ }, { "cell_type": "code", - "execution_count": 34, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "" - ] - }, - "execution_count": 34, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "# Fit causal effect model from observational data\n", "causal_effects.fit_total_effect(\n", @@ -1702,17 +1711,9 @@ }, { "cell_type": "code", - "execution_count": 35, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Causal effect is 0.85\n" - ] - } - ], + "outputs": [], "source": [ "intervention_data = 1.*np.ones((1, 1))\n", "y1 = causal_effects.predict_total_effect( \n", @@ -1742,42 +1743,9 @@ }, { "cell_type": "code", - "execution_count": 36, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "##\n", - "## Initializing CausalEffects class\n", - "##\n", - "\n", - "Input:\n", - "\n", - "graph_type = dag\n", - "X = [(0, 0), (1, 0)]\n", - "Y = [(3, 0)]\n", - "S = []\n", - "M = [(2, 0)]\n", - "\n", - "\n", - "\n", - "Oset = [('$Z_1$', 0), ('$Z_2$', 0), ('$Z_3$', 0)]\n" - ] - }, - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "graph = np.array([['', '-->', '', '', '', '', ''],\n", " ['<--', '', '-->', '-->', '', '<--', ''],\n", @@ -1824,7 +1792,7 @@ }, { "cell_type": "code", - "execution_count": 37, + "execution_count": null, "metadata": {}, "outputs": [], "source": [ @@ -1851,7 +1819,7 @@ "We fit and predict with Wright's method. Next to the data handling arguments ``fit_wrights_effect`` takes:\n", "\n", "* ``mediation : None or 'direct' or list of tuples``: If None, the total effect is estimated, if 'direct', only the direct effect is estimated, else only those causal paths are considerd that pass at least through one of these mediator nodes.\n", - "* ``method : {'parents', 'links_coeffs', 'optimal'}``: Method to use for estimating Wright's path coefficients. If 'optimal', the Oset is used, if 'links_coeffs', the coefficients in links_coeffs are used, if 'parents', the parents are used (only valid for DAGs). 'links_coeffs' can be used for testing purposes if you play around with toy models.\n", + "* ``method : {'parents', 'links_coeffs', 'optimal'}``: Method to use for estimating Wright's path coefficients. If 'optimal', the Oset is used, if 'links_coeffs', the coefficients in links_coeffs are used, if 'parents', the parents are used (only valid if there are no bi-directed links adjacent to mediators or Y). 'links_coeffs' can be used for testing purposes if you play around with toy models.\n", "* ``links_coeffs : dict``: Only used if method = 'links_coeffs'. Dictionary of format: {0:[((i, -tau), coeff),...], 1:[...], ...} for all variables where i must be in [0..N-1] and tau >= 0 with number of variables N. coeff must be a float.\n", "\n", "The default is ``method='parents'`` which is suitable here since we deal with a DAG." @@ -1859,17 +1827,9 @@ }, { "cell_type": "code", - "execution_count": 38, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Causal effect is 0.75\n" - ] - } - ], + "outputs": [], "source": [ "causal_effects.fit_wright_effect(dataframe=dataframe)\n", "\n", @@ -1898,17 +1858,9 @@ }, { "cell_type": "code", - "execution_count": 39, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Mediated causal effect through M = [(2, 0)] is 0.23\n" - ] - } - ], + "outputs": [], "source": [ "considered_mediators = [(2, 0)]\n", "causal_effects.fit_wright_effect(dataframe=dataframe, mediation=considered_mediators)\n", @@ -1936,17 +1888,9 @@ }, { "cell_type": "code", - "execution_count": 40, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Direct causal effect is 0.51\n" - ] - } - ], + "outputs": [], "source": [ "causal_effects.fit_wright_effect(dataframe=dataframe, mediation='direct')\n", "\n", @@ -1982,26 +1926,9 @@ }, { "cell_type": "code", - "execution_count": 41, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "##\n", - "## Running Bootstrap of fit_wright_effect \n", - "##\n", - "\n", - "boot_samples = 1000 \n", - "\n", - "boot_blocklength = 1 \n", - "\n", - "(2, 1)\n" - ] - } - ], + "outputs": [], "source": [ "# Let's generate shorter length data\n", "T = 100\n", @@ -2023,37 +1950,9 @@ }, { "cell_type": "code", - "execution_count": 42, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "\n", - "##\n", - "## Running Bootstrap of fit_total_effect \n", - "##\n", - "\n", - "boot_samples = 1000 \n", - "\n", - "boot_blocklength = 1 \n", - "\n" - ] - }, - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": { - "needs_background": "light" - }, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "# Now the total effect estimate\n", "# First fit \n", @@ -2116,9 +2015,9 @@ ], "metadata": { "kernelspec": { - "display_name": "Python (py39)", + "display_name": "Python 3 (ipykernel)", "language": "python", - "name": "py39" + "name": "python3" }, "language_info": { "codemirror_mode": { @@ -2130,7 +2029,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.9.12" + "version": "3.9.7" } }, "nbformat": 4, diff --git a/tutorials/tigramite_tutorial_pcmciplus.ipynb b/tutorials/tigramite_tutorial_pcmciplus.ipynb index 7f935656..d50d6028 100644 --- a/tutorials/tigramite_tutorial_pcmciplus.ipynb +++ b/tutorials/tigramite_tutorial_pcmciplus.ipynb @@ -14,9 +14,6 @@ "J. Runge (2020), Discovering contemporaneous and lagged causal relations in autocorrelated nonlinear time series datasets\n", "http://www.auai.org/uai2020/proceedings/579_main_paper.pdf\n", "\n", - "See the following paper for theoretical background:\n", - "Runge, Jakob. 2018. “Causal Network Reconstruction from Time Series: From Theoretical Assumptions to Practical Estimation.” Chaos: An Interdisciplinary Journal of Nonlinear Science 28 (7): 075310.\n", - "\n", "Last, the following Nature Communications Perspective paper provides an overview of causal inference methods in general, identifies promising applications, and discusses methodological challenges (exemplified in Earth system sciences): \n", "https://www.nature.com/articles/s41467-019-10105-3" ] @@ -25,23 +22,7 @@ "cell_type": "code", "execution_count": 1, "metadata": {}, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/plotting.py:26: UserWarning: [Errno 2] No such file or directory: '/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/../versions.py'\n", - " warnings.warn(str(e))\n", - "OMP: Info #276: omp_set_nested routine deprecated, please use omp_set_max_active_levels instead.\n", - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/gpdc.py:27: UserWarning: [Errno 2] No such file or directory: '/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/../../versions.py'\n", - " warnings.warn(str(e))\n", - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n", - " from .autonotebook import tqdm as notebook_tqdm\n", - "/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/gpdc_torch.py:33: UserWarning: [Errno 2] No such file or directory: '/home/jakobrunge/anaconda3/envs/py39/lib/python3.9/site-packages/tigramite-5.0.1.18-py3.9.egg/tigramite/independence_tests/../../versions.py'\n", - " warnings.warn(str(e))\n" - ] - } - ], + "outputs": [], "source": [ "# Imports\n", "import numpy as np\n", @@ -163,7 +144,7 @@ "\n", "The general idea behind PCMCIplus follows that of the PC algorithm: \n", "\n", - "* **Skeleton discovery phase**: Starting from a completely connected graph first a skeleton of adjacencies $X^i_{t-\\tau} - X^j_t$ is estimated by identifying which pairs of nodes are conditionally independent for certain subset of the other nodes. See the paper for the particular way that conditions are chosen, which is different from the original PC algorithm. The adjacency between conditionally independent pairs is removed. The lagged adjacencies in that skeleton are then automatically oriented by time-order. For example, an undirected link $X^i_{t-2} - X^j_t$ can only be oriented as $X^i_{t-2} \\to X^j_t$ since causal effects cannot go back in time. \n", + "* **Skeleton discovery phase**: Starting from a completely connected graph first a skeleton of adjacencies $X^i_{t-\\tau} - X^j_t$ is estimated by identifying which pairs of nodes are conditionally independent for certain subsets of the other nodes. See the paper for the particular way that conditions are chosen, which is different from the original PC algorithm. The adjacency between conditionally independent pairs is removed. The lagged adjacencies in that skeleton are then automatically oriented by time-order. For example, an undirected link $X^i_{t-2} - X^j_t$ can only be oriented as $X^i_{t-2} \\to X^j_t$ since causal effects cannot go back in time. \n", "\n", "* **Collider orientation phase**: The contemporaneous adjacencies $X^i_{t} - X^j_t$ are then oriented based on the following collider rule. For an unshielded triple $X^k_{t-\\tau} - X^i_t - X^j_t$ with $\\tau\\geq 0$ (for $\\tau>0$ we always have $X^k_{t-\\tau} \\rightarrow X^i_t$) with no adjacency between $X^k_{t-\\tau}$ and $X^j_t$: If $X^i_t$ is *not* part of the conditioning set that makes $X^k_{t-\\tau}$ and $X^j_t$ independent, then orient $X^k_{t-\\tau} - X^i_t - X^j_t$ as $X^k_{t-\\tau} \\rightarrow X^i_t \\leftarrow X^j_t$. This rule is applied to all unshielded triples. There are three options (``contemp_collider_rule={'none', 'majority', 'conservative'}``) to decide whether a middle node $X^i_t$ is *not* part of the separating conditioning set: ``'none'``: In the original PC algorithm the conditions that lead to conditional independence in the skeleton discovery phase are stored (``sepset`` in Tigramite) and then used in the collider phase. Alternatively, all separating conditioning sets are *re-computed* based on the neighbors of $X^k_{t-\\tau}$ and $X^j_t$ and collider motifs are oriented based on the ``'majority'`` or ``'conservative'`` rule as discussed in the paper.\n", "\n", @@ -1007,13 +988,7 @@ " Subset 0: () gives pval = 0.00000 / val = 0.400\n", " No conditions of dimension 0 left.\n", "\n", - " Link ($X^{7}$ -2) --> $X^{2}$ (23/27):\n" - ] - }, - { - "name": "stdout", - "output_type": "stream", - "text": [ + " Link ($X^{7}$ -2) --> $X^{2}$ (23/27):\n", " Subset 0: () gives pval = 0.00000 / val = 0.395\n", " No conditions of dimension 0 left.\n", "\n", @@ -1215,7 +1190,7 @@ " Still subsets of dimension 2 left, but q_max = 1 reached.\n", "\n", " Link ($X^{7}$ -3) --> $X^{2}$ (15/17):\n", - " Subset 0: ($X^{2}$ -3) ($X^{2}$ -2) gives pval = 0.84770 / val = 0.009\n", + " Subset 0: ($X^{2}$ -3) ($X^{2}$ -2) gives pval = 0.84769 / val = 0.009\n", " Non-significance detected.\n", "\n", " Link ($X^{7}$ -2) --> $X^{2}$ (16/17):\n", @@ -1265,7 +1240,7 @@ " Non-significance detected.\n", "\n", " Link ($X^{0}$ -1) --> $X^{2}$ (5/13):\n", - " Subset 0: ($X^{3}$ -1) ($X^{2}$ -1) ($X^{3}$ -2) gives pval = 0.94750 / val = -0.003\n", + " Subset 0: ($X^{3}$ -1) ($X^{2}$ -1) ($X^{3}$ -2) gives pval = 0.94749 / val = -0.003\n", " Non-significance detected.\n", "\n", " Link ($X^{3}$ -3) --> $X^{2}$ (6/13):\n", @@ -1292,12 +1267,18 @@ " Subset 0: ($X^{3}$ -1) ($X^{2}$ -1) ($X^{3}$ -2) gives pval = 0.10329 / val = 0.074\n", " Non-significance detected.\n", "\n", - " Link ($X^{2}$ -3) --> $X^{2}$ (12/13):\n", - " Subset 0: ($X^{3}$ -1) ($X^{2}$ -1) ($X^{3}$ -2) gives pval = 0.46521 / val = -0.033\n", + " Link ($X^{2}$ -3) --> $X^{2}$ (12/13):\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + " Subset 0: ($X^{3}$ -1) ($X^{2}$ -1) ($X^{3}$ -2) gives pval = 0.46520 / val = -0.033\n", " Non-significance detected.\n", "\n", " Link ($X^{4}$ -2) --> $X^{2}$ (13/13):\n", - " Subset 0: ($X^{3}$ -1) ($X^{2}$ -1) ($X^{3}$ -2) gives pval = 0.93596 / val = -0.004\n", + " Subset 0: ($X^{3}$ -1) ($X^{2}$ -1) ($X^{3}$ -2) gives pval = 0.93595 / val = -0.004\n", " Non-significance detected.\n", "\n", " Sorting parents in decreasing order with \n", @@ -1324,7 +1305,7 @@ " Still subsets of dimension 4 left, but q_max = 1 reached.\n", "\n", " Link ($X^{1}$ -1) --> $X^{2}$ (3/6):\n", - " Subset 0: ($X^{2}$ -1) ($X^{3}$ -1) ($X^{1}$ -2) ($X^{3}$ -2) gives pval = 0.21702 / val = -0.056\n", + " Subset 0: ($X^{2}$ -1) ($X^{3}$ -1) ($X^{1}$ -2) ($X^{3}$ -2) gives pval = 0.21703 / val = -0.056\n", " Non-significance detected.\n", "\n", " Link ($X^{1}$ -2) --> $X^{2}$ (4/6):\n", @@ -1678,13 +1659,7 @@ " Subset 0: () gives pval = 0.00000 / val = 0.428\n", " No conditions of dimension 0 left.\n", "\n", - " Link ($X^{1}$ -3) --> $X^{4}$ (6/27):\n" - ] - }, - { - "name": "stdout", - "output_type": "stream", - "text": [ + " Link ($X^{1}$ -3) --> $X^{4}$ (6/27):\n", " Subset 0: () gives pval = 0.00000 / val = 0.450\n", " No conditions of dimension 0 left.\n", "\n", @@ -1930,7 +1905,7 @@ " Still subsets of dimension 2 left, but q_max = 1 reached.\n", "\n", " Link ($X^{2}$ -3) --> $X^{4}$ (9/17):\n", - " Subset 0: ($X^{4}$ -3) ($X^{4}$ -2) gives pval = 0.55942 / val = 0.026\n", + " Subset 0: ($X^{4}$ -3) ($X^{4}$ -2) gives pval = 0.55941 / val = 0.026\n", " Non-significance detected.\n", "\n", " Link ($X^{4}$ -1) --> $X^{4}$ (10/17):\n", @@ -1958,7 +1933,7 @@ " Non-significance detected.\n", "\n", " Link ($X^{7}$ -3) --> $X^{4}$ (16/17):\n", - " Subset 0: ($X^{4}$ -3) ($X^{4}$ -2) gives pval = 0.75666 / val = -0.014\n", + " Subset 0: ($X^{4}$ -3) ($X^{4}$ -2) gives pval = 0.75665 / val = -0.014\n", " Non-significance detected.\n", "\n", " Link ($X^{7}$ -1) --> $X^{4}$ (17/17):\n", @@ -2036,7 +2011,7 @@ " Non-significance detected.\n", "\n", " Link ($X^{2}$ -2) --> $X^{4}$ (13/13):\n", - " Subset 0: ($X^{3}$ -1) ($X^{4}$ -1) ($X^{3}$ -2) gives pval = 0.67180 / val = 0.019\n", + " Subset 0: ($X^{3}$ -1) ($X^{4}$ -1) ($X^{3}$ -2) gives pval = 0.67179 / val = 0.019\n", " Non-significance detected.\n", "\n", " Sorting parents in decreasing order with \n", @@ -2063,7 +2038,7 @@ " Still subsets of dimension 4 left, but q_max = 1 reached.\n", "\n", " Link ($X^{1}$ -2) --> $X^{4}$ (3/6):\n", - " Subset 0: ($X^{4}$ -1) ($X^{3}$ -1) ($X^{1}$ -1) ($X^{1}$ -3) gives pval = 0.12058 / val = 0.070\n", + " Subset 0: ($X^{4}$ -1) ($X^{3}$ -1) ($X^{1}$ -1) ($X^{1}$ -3) gives pval = 0.12059 / val = 0.070\n", " Non-significance detected.\n", "\n", " Link ($X^{1}$ -1) --> $X^{4}$ (4/6):\n", @@ -2071,7 +2046,7 @@ " Non-significance detected.\n", "\n", " Link ($X^{1}$ -3) --> $X^{4}$ (5/6):\n", - " Subset 0: ($X^{4}$ -1) ($X^{3}$ -1) ($X^{1}$ -2) ($X^{1}$ -1) gives pval = 0.55066 / val = 0.027\n", + " Subset 0: ($X^{4}$ -1) ($X^{3}$ -1) ($X^{1}$ -2) ($X^{1}$ -1) gives pval = 0.55067 / val = 0.027\n", " Non-significance detected.\n", "\n", " Link ($X^{3}$ -2) --> $X^{4}$ (6/6):\n", @@ -2161,13 +2136,7 @@ " Subset 0: () gives pval = 0.00000 / val = 0.416\n", " No conditions of dimension 0 left.\n", "\n", - " Link ($X^{5}$ -2) --> $X^{5}$ (17/27):\n" - ] - }, - { - "name": "stdout", - "output_type": "stream", - "text": [ + " Link ($X^{5}$ -2) --> $X^{5}$ (17/27):\n", " Subset 0: () gives pval = 0.05232 / val = 0.087\n", " Non-significance detected.\n", "\n", @@ -2413,7 +2382,13 @@ "\n", "Testing condition sets of dimension 2:\n", "\n", - " Link ($X^{6}$ -1) --> $X^{6}$ (1/3):\n", + " Link ($X^{6}$ -1) --> $X^{6}$ (1/3):\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ " Subset 0: ($X^{5}$ -1) ($X^{5}$ -2) gives pval = 0.00000 / val = 0.513\n", " Still subsets of dimension 2 left, but q_max = 1 reached.\n", "\n", @@ -2889,7 +2864,7 @@ " Iterate through 1 subset(s) of conditions: \n", " with conds_y = [ ($X^{5}$ -1) ($X^{6}$ -1) ]\n", " with conds_x = [ ($X^{0}$ -1) ($X^{1}$ -1) ]\n", - " Subset 0: () gives pval = 0.42604 / val = 0.036\n", + " Subset 0: () gives pval = 0.42603 / val = 0.036\n", " Non-significance detected.\n", "\n", " Link ($X^{0}$ 0) o-o $X^{6}$ (7/86):\n", @@ -3030,13 +3005,7 @@ " Link ($X^{2}$ 0) o-o $X^{7}$ (28/86):\n", " Iterate through 1 subset(s) of conditions: \n", " with conds_y = [ ($X^{7}$ -1) ]\n", - " with conds_x = [ ($X^{2}$ -1) ($X^{3}$ -1) ]\n" - ] - }, - { - "name": "stdout", - "output_type": "stream", - "text": [ + " with conds_x = [ ($X^{2}$ -1) ($X^{3}$ -1) ]\n", " Subset 0: () gives pval = 0.54563 / val = -0.027\n", " Non-significance detected.\n", "\n", @@ -3393,7 +3362,7 @@ " Iterate through 1 subset(s) of conditions: \n", " with conds_y = [ ($X^{4}$ -1) ($X^{3}$ -1) ]\n", " with conds_x = [ ($X^{2}$ -1) ($X^{3}$ -1) ]\n", - " Subset 0: ($X^{3}$ 0) gives pval = 0.60740 / val = -0.023\n", + " Subset 0: ($X^{3}$ 0) gives pval = 0.60741 / val = -0.023\n", " Non-significance detected.\n", "\n", " Link ($X^{3}$ 0) o-o $X^{2}$ (5/17):\n", @@ -3407,7 +3376,7 @@ " Iterate through 2 subset(s) of conditions: \n", " with conds_y = [ ($X^{2}$ -1) ]\n", " with conds_x = [ ($X^{3}$ -2) ($X^{1}$ -3) ]\n", - " Subset 0: ($X^{3}$ 0) gives pval = 0.57260 / val = -0.026\n", + " Subset 0: ($X^{3}$ 0) gives pval = 0.57261 / val = -0.026\n", " Non-significance detected.\n", "\n", " Link ($X^{3}$ -1) --> $X^{3}$ (7/17):\n", @@ -3429,7 +3398,7 @@ " Iterate through 2 subset(s) of conditions: \n", " with conds_y = [ ($X^{4}$ -1) ]\n", " with conds_x = [ ($X^{3}$ -2) ($X^{1}$ -3) ]\n", - " Subset 0: ($X^{3}$ 0) gives pval = 0.90730 / val = 0.005\n", + " Subset 0: ($X^{3}$ 0) gives pval = 0.90731 / val = 0.005\n", " Non-significance detected.\n", "\n", " Link ($X^{4}$ 0) o-o $X^{2}$ (10/17):\n", @@ -3580,7 +3549,7 @@ " with conds_y = [ ($X^{2}$ -1) ]\n", " with conds_x = [ ($X^{3}$ -2) ($X^{1}$ -3) ]\n", " Subset 0: () gives pval = 0.00000 / val = -0.402\n", - " Subset 1: ($X^{3}$ 0) gives pval = 0.57260 / val = -0.026\n", + " Subset 1: ($X^{3}$ 0) gives pval = 0.57261 / val = -0.026\n", " Fraction of separating subsets containing ($X^{3}$ 0) is > 0.5 --> non-collider found\n", "\n", " Triple ($X^{4}$ 0) o-o $X^{3}$ o-o $X^{2}$ (3/10)\n", @@ -3588,7 +3557,7 @@ " with conds_y = [ ($X^{2}$ -1) ($X^{3}$ -1) ]\n", " with conds_x = [ ($X^{4}$ -1) ($X^{3}$ -1) ]\n", " Subset 0: () gives pval = 0.00000 / val = -0.243\n", - " Subset 1: ($X^{3}$ 0) gives pval = 0.60740 / val = -0.023\n", + " Subset 1: ($X^{3}$ 0) gives pval = 0.60741 / val = -0.023\n", " Fraction of separating subsets containing ($X^{3}$ 0) is > 0.5 --> non-collider found\n", "\n", " Triple ($X^{2}$ -1) --> $X^{2}$ o-o $X^{3}$ (4/10)\n", @@ -3596,8 +3565,8 @@ " with conds_y = [ ($X^{3}$ -1) ($X^{1}$ -2) ]\n", " with conds_x = [ ($X^{2}$ -2) ($X^{3}$ -2) ]\n", " Subset 0: ($X^{2}$ 0) ($X^{4}$ 0) gives pval = 0.00000 / val = 0.247\n", - " Subset 1: ($X^{4}$ 0) gives pval = 0.62678 / val = -0.022\n", - " Subset 2: () gives pval = 0.62897 / val = -0.022\n", + " Subset 1: ($X^{4}$ 0) gives pval = 0.62679 / val = -0.022\n", + " Subset 2: () gives pval = 0.62896 / val = -0.022\n", " Subset 3: ($X^{2}$ 0) gives pval = 0.00000 / val = 0.248\n", " Fraction of separating subsets containing ($X^{2}$ 0) is < 0.5 --> collider found\n", "\n", @@ -3616,7 +3585,7 @@ " with conds_y = [ ($X^{4}$ -1) ($X^{3}$ -1) ]\n", " with conds_x = [ ($X^{1}$ -3) ]\n", " Subset 0: () gives pval = 0.00266 / val = 0.135\n", - " Subset 1: ($X^{3}$ 0) gives pval = 0.87116 / val = -0.007\n", + " Subset 1: ($X^{3}$ 0) gives pval = 0.87117 / val = -0.007\n", " Fraction of separating subsets containing ($X^{3}$ 0) is > 0.5 --> non-collider found\n", "\n", " Triple ($X^{2}$ 0) o-o $X^{3}$ o-o $X^{4}$ (7/10)\n", @@ -3624,7 +3593,7 @@ " with conds_y = [ ($X^{4}$ -1) ($X^{3}$ -1) ]\n", " with conds_x = [ ($X^{2}$ -1) ($X^{3}$ -1) ]\n", " Subset 0: () gives pval = 0.00000 / val = -0.243\n", - " Subset 1: ($X^{3}$ 0) gives pval = 0.60740 / val = -0.023\n", + " Subset 1: ($X^{3}$ 0) gives pval = 0.60741 / val = -0.023\n", " Fraction of separating subsets containing ($X^{3}$ 0) is > 0.5 --> non-collider found\n", "\n", " Triple ($X^{3}$ -1) --> $X^{3}$ o-o $X^{4}$ (8/10)\n", @@ -3632,7 +3601,7 @@ " with conds_y = [ ($X^{4}$ -1) ]\n", " with conds_x = [ ($X^{3}$ -2) ($X^{1}$ -3) ]\n", " Subset 0: () gives pval = 0.00000 / val = 0.369\n", - " Subset 1: ($X^{3}$ 0) gives pval = 0.90730 / val = 0.005\n", + " Subset 1: ($X^{3}$ 0) gives pval = 0.90731 / val = 0.005\n", " Fraction of separating subsets containing ($X^{3}$ 0) is > 0.5 --> non-collider found\n", "\n", " Triple ($X^{6}$ -1) --> $X^{6}$ o-o $X^{5}$ (9/10)\n", @@ -4383,9 +4352,9 @@ ], "metadata": { "kernelspec": { - "display_name": "Python (py39)", + "display_name": "Python 3 (ipykernel)", "language": "python", - "name": "py39" + "name": "python3" }, "language_info": { "codemirror_mode": { @@ -4397,7 +4366,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.9.12" + "version": "3.9.7" } }, "nbformat": 4, diff --git a/tutorials/tigramite_tutorial_prediction.ipynb b/tutorials/tigramite_tutorial_prediction.ipynb index 31874b7d..fd42d29f 100644 --- a/tutorials/tigramite_tutorial_prediction.ipynb +++ b/tutorials/tigramite_tutorial_prediction.ipynb @@ -689,9 +689,9 @@ ], "metadata": { "kernelspec": { - "display_name": "Python (py39)", + "display_name": "Python 3 (ipykernel)", "language": "python", - "name": "py39" + "name": "python3" }, "language_info": { "codemirror_mode": { @@ -703,7 +703,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.9.12" + "version": "3.9.7" } }, "nbformat": 4, From f7b17347b7959cc32255919ad173fe99a3bd49ce Mon Sep 17 00:00:00 2001 From: jakobrunge Date: Tue, 20 Sep 2022 20:34:27 +0200 Subject: [PATCH 2/2] fixes regarding typos in tutorials --- setup.py | 2 +- ...orial_general_causal_effect_analysis.ipynb | 169 ++++++++++++++++-- 2 files changed, 151 insertions(+), 20 deletions(-) diff --git a/setup.py b/setup.py index a214796c..c6bec529 100644 --- a/setup.py +++ b/setup.py @@ -63,7 +63,7 @@ def run(self): # Run the setup setup( name="tigramite", - version="5.1.0.4", + version="5.1.0.5", packages=["tigramite", "tigramite.independence_tests", "tigramite.toymodels"], license="GNU General Public License v3.0", description="Tigramite causal discovery for time series", diff --git a/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb b/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb index 734aa7ca..b868388e 100644 --- a/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb +++ b/tutorials/tigramite_tutorial_general_causal_effect_analysis.ipynb @@ -1596,7 +1596,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 32, "metadata": {}, "outputs": [ { @@ -1620,6 +1620,16 @@ "\n", "Oset = [('$X^2$', -3), ('$X^1$', -4), ('$X^1$', -3)]\n" ] + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" } ], "source": [ @@ -1672,7 +1682,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 33, "metadata": {}, "outputs": [], "source": [ @@ -1697,9 +1707,20 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 34, "metadata": {}, - "outputs": [], + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 34, + "metadata": {}, + "output_type": "execute_result" + } + ], "source": [ "# Fit causal effect model from observational data\n", "causal_effects.fit_total_effect(\n", @@ -1711,9 +1732,17 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 35, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Causal effect is 0.85\n" + ] + } + ], "source": [ "intervention_data = 1.*np.ones((1, 1))\n", "y1 = causal_effects.predict_total_effect( \n", @@ -1743,9 +1772,42 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 36, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "##\n", + "## Initializing CausalEffects class\n", + "##\n", + "\n", + "Input:\n", + "\n", + "graph_type = dag\n", + "X = [(0, 0), (1, 0)]\n", + "Y = [(3, 0)]\n", + "S = []\n", + "M = [(2, 0)]\n", + "\n", + "\n", + "\n", + "Oset = [('$Z_1$', 0), ('$Z_2$', 0), ('$Z_3$', 0)]\n" + ] + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], "source": [ "graph = np.array([['', '-->', '', '', '', '', ''],\n", " ['<--', '', '-->', '-->', '', '<--', ''],\n", @@ -1792,7 +1854,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 37, "metadata": {}, "outputs": [], "source": [ @@ -1827,9 +1889,17 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 38, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Causal effect is 0.75\n" + ] + } + ], "source": [ "causal_effects.fit_wright_effect(dataframe=dataframe)\n", "\n", @@ -1858,9 +1928,17 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 39, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Mediated causal effect through M = [(2, 0)] is 0.23\n" + ] + } + ], "source": [ "considered_mediators = [(2, 0)]\n", "causal_effects.fit_wright_effect(dataframe=dataframe, mediation=considered_mediators)\n", @@ -1888,9 +1966,17 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 40, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Direct causal effect is 0.51\n" + ] + } + ], "source": [ "causal_effects.fit_wright_effect(dataframe=dataframe, mediation='direct')\n", "\n", @@ -1926,9 +2012,26 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 41, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "##\n", + "## Running Bootstrap of fit_wright_effect \n", + "##\n", + "\n", + "boot_samples = 1000 \n", + "\n", + "boot_blocklength = 1 \n", + "\n", + "(2, 1)\n" + ] + } + ], "source": [ "# Let's generate shorter length data\n", "T = 100\n", @@ -1950,9 +2053,37 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 42, "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "##\n", + "## Running Bootstrap of fit_total_effect \n", + "##\n", + "\n", + "boot_samples = 1000 \n", + "\n", + "boot_blocklength = 1 \n", + "\n" + ] + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], "source": [ "# Now the total effect estimate\n", "# First fit \n",