R语言：R2OpenBUGS

OStack程序员社区-中国程序员成长平台 › 门户 › 编程› R语言›R语言教程

原作者: [db:作者] 来自: [db:来源] 收藏邀请

用这个包调用BUGS model,分别用表格和图形概述inference和convergence，保存估计的结果

as.bugs.array 转换成bugs object

函数把马尔科夫链估计结果（不是来自于BUGS），转成BUGS object，主要用来plot.bugs 展示结果。

as.bugs.array(sims.array, model.file=NULL, program=NULL, DIC=FALSE, DICOutput=NULL, n.iter=NULL, n.burnin=0, n.thin=1)

sims.array ：3维数组 n.keep, n.chains和combined parameter vector的长度

model.file ： OpenBUGS编写的.odc 模型文件

DIC ：是否计算DIC曲线

DICOutput ： DIC值

n.iter ：生成sims.array 每条chain 迭代数

n.burnin ：丢弃的迭代次数

n.thin ：thinning rate

attach.all 添加数据到搜索路径

The database is attached/detached to the search path，While attach.all attaches all elements of an object x to a database called name, attach.bugs attaches all elements of x$sims.list to the database bugs.sims itself making use of attach.all.

attach.all(x, overwrite = NA, name = “attach.all”) attach.bugs(x, overwrite = NA) detach.all(name = “attach.all”) detach.bugs()

x : bugs 对象

overwrite ：TRUE 删除全局环境中被掩盖的数据， NA 询问，FALSE

name : 环境name

bugs 最重要，用R运行bugs

自动输入值，启动bugs，保存结果

bugs(data, inits, parameters.to.save, n.iter, model.file=“model.txt”, n.chains=3, n.burnin=floor(n.iter / 2), n.thin=1, saveExec=FALSE,restart=FALSE, debug=FALSE, DIC=TRUE, digits=5, codaPkg=FALSE, OpenBUGS.pgm=NULL, working.directory=NULL, clearWD=FALSE, useWINE=FALSE, WINE=NULL, newWINE=TRUE, WINEPATH=NULL, bugs.seed=1, summary.only=FALSE, save.history=(.Platform$OS.type == “windows” | useWINE==TRUE), over.relax = FALSE)

data ：模型中使用的数据

inits ：n chain 的元素列表，每一个要素是一个模型初始值列表，或者一个生成初始值得function

parameters.to.save ：需要被记录的参数名向量

model.file ： model 文件.txt

n.chains : 默认3条

n.iter ：每条链的迭代次数，默认2000

n.thin ： Thinning rate. 正整数，默认是1，

saveExec ：使用basename(模型.file)保存OpenBUGS执行的重新启动映像。

restart ：执行从上次执行的最后状态恢复，存储在工作目录中的.bug文件中。

debug ：默认FALSE,正在运行行时Openbugs 页面关闭

DIC ：计算deviance,，pD，和DIC。

digits ：有效小数位数

codaPkg ：FALSE 返回 bugs对象，否则输出，用coda 包 read.bugs 读取，

OpenBUGS.pgm ：通向OpenBUGS可执行程序的完整路径。

working.directory：OpenBUGS的输入和输出将存储在此目录中;

clearWD ：是否这些文件的“data.txt”、“init(1:n.chains). txt”,“log.odc”、“codaIndex.txt”和“coda[1:nchain].txt”结束时删除。

bugs.seed ：OpenBUGS随机种子，1-14整数

summary.only ： TURE ,仅对非常快速的分析给出了一个参数概要

save.history : TURE,最后画出trace

# An example model file is given in:
model.file <- system.file(package="R2OpenBUGS", "model", "schools.txt")
# Let's take a look:
print(model.file)

[1] "C:/Users/Date/Documents/R/win-library/3.5/R2OpenBUGS/model/schools.txt"

file.show(model.file)

model { for (j in 1:J){ y[j] ~ dnorm (theta[j], tau.y[j]) theta[j] ~ dnorm (mu.theta, tau.theta) tau.y[j] <- pow(sigma.y[j], -2) } mu.theta ~ dnorm (0.0, 1.0E-6) tau.theta <- pow(sigma.theta, -2) sigma.theta ~ dunif (0, 1000) }

data(schools)
schools

  school estimate   sd
1      A    28.39 14.9
2      B     7.94 10.2
3      C    -2.75 16.3
4      D     6.82 11.0
5      E    -0.64  9.4
6      F     0.63 11.4
7      G    18.01 10.4
8      H    12.16 17.6

J <- nrow(schools)
y <- schools$estimate
sigma.y <- schools$sd
data <- list ("J", "y", "sigma.y")
data

[[1]]
[1] "J"

[[2]]
[1] "y"

[[3]]
[1] "sigma.y"

inits <- function(){
      list(theta=rnorm(J, 0, 100), mu.theta=rnorm(1, 0, 100),sigma.theta=runif(1, 0, 100))
}
parameters <- c("theta", "mu.theta", "sigma.theta")

schools.sim <- bugs(data, inits, parameters, model.file,
n.chains=3, n.iter=5000)
print(schools.sim)

Inference for Bugs model at "C:/Users/Date/Documents/R/win-library/3.5/R2OpenBUGS/model/schools.txt", 
Current: 3 chains, each with 5000 iterations (first 2500 discarded)
Cumulative: n.sims = 7500 iterations saved
            mean  sd  2.5%  25%  50%  75% 97.5%
theta[1]    11.2 8.9  -2.8  5.5  9.8 15.7  32.9
theta[2]     7.5 6.5  -4.9  3.4  7.4 11.5  20.8
theta[3]     5.8 8.0 -12.3  1.3  6.4 10.5  21.0
theta[4]     7.1 6.7  -6.3  3.0  7.1 11.2  20.9
theta[5]     4.9 6.3  -8.8  1.0  5.5  9.1  16.4
theta[6]     5.8 6.8  -9.2  1.7  6.1 10.1  18.3
theta[7]    10.4 7.2  -2.1  5.6  9.7 14.6  26.3
theta[8]     8.1 8.0  -7.1  3.4  7.8 12.4  25.7
mu.theta     7.6 5.4  -2.7  4.1  7.6 11.0  18.6
sigma.theta  6.7 5.9   0.2  2.3  5.3  9.4  21.8
deviance    60.5 2.3  57.0 59.2 60.1 61.6  66.2
            Rhat n.eff
theta[1]       1   410
theta[2]       1   420
theta[3]       1  1200
theta[4]       1   610
theta[5]       1   680
theta[6]       1   490
theta[7]       1   300
theta[8]       1   420
mu.theta       1   310
sigma.theta    1   540
deviance       1  1600

For each parameter, n.eff is a crude measure of effective sample size,
and Rhat is the potential scale reduction factor (at convergence, Rhat=1).

DIC info (using the rule, pD = Dbar-Dhat)
pD = 2.9 and DIC = 63.4
DIC is an estimate of expected predictive error (lower deviance is better).

plot(schools.sim)

bugs.data 生成输入文件

bugs.data(data, dir = getwd(), digits = 5, data.file = “data.txt”)

bugs.inits 生成初始值文件

bugs.inits(inits, n.chains, digits, inits.files = paste(“inits”, 1:n.chains, “.txt”, sep = “”))

bugs.log 读取log文件（summary statistics and DIC information）

bugs.log(file)

plot.bugs 画bugs对象

plot(x, display.parallel = FALSE, …)

display.parallel ：在摘要图的两部分中显示平行的间隔

print.bugs 输出bugs对象

print(x, digits.summary = 1, …)

digits.summary：四舍五入的位数

read.bugs

读Markov链蒙特卡罗输出的CODA格式。并返回一个类mcmc.list对象。使用coda包进行进一步的输出分析列表。

read.bugs(codafiles, …)

validateInstallOpenBUGS 比较R和openbugs软件运行结果

write.model 转化R function创建模型文件

schoolsmodel <- function(){
     for (j in 1:J){
     y[j] ~ dnorm (theta[j], tau.y[j])
     theta[j] ~ dnorm (mu.theta, tau.theta)
     tau.y[j] <- pow(sigma.y[j], -2)
}
     mu.theta ~ dnorm (0.0, 1.0E-6)
     tau.theta <- pow(sigma.theta, -2)
     sigma.theta ~ dunif (0, 1000)
}
## some temporary filename:
filename <- file.path(tempdir(), "schoolsmodel.txt")
## write model file:
write.model(schoolsmodel, filename)
## and let's take a look:
file.show(filename)

LS0tDQp0aXRsZTogIlLor63oqIDvvJpSMk9wZW5CVUdTIg0Kb3V0cHV0OiBodG1sX25vdGVib29rDQotLS0NCiMgUuivreiogO+8mlIyT3BlbkJVR1MNCg0K55So6L+Z5Liq5YyF6LCD55SoQlVHUyBtb2RlbCzliIbliKvnlKjooajmoLzlkozlm77lvaLmpoLov7BpbmZlcmVuY2Xlkoxjb252ZXJnZW5jZe+8jOS/neWtmOS8sOiuoeeahOe7k+aenA0KDQojIyBhcy5idWdzLmFycmF5ICAg6L2s5o2i5oiQYnVncyBvYmplY3QNCg0K5Ye95pWw5oqK6ams5bCU56eR5aSr6ZO+5Lyw6K6h57uT5p6c77yI5LiN5piv5p2l6Ieq5LqOQlVHU++8ie+8jOi9rOaIkEJVR1Mgb2JqZWN077yM5Li76KaB55So5p2lcGxvdC5idWdzIOWxleekuue7k+aenOOAgg0KDQphcy5idWdzLmFycmF5KHNpbXMuYXJyYXksIG1vZGVsLmZpbGU9TlVMTCwgcHJvZ3JhbT1OVUxMLA0KRElDPUZBTFNFLCBESUNPdXRwdXQ9TlVMTCwgbi5pdGVyPU5VTEwsIG4uYnVybmluPTAsIG4udGhpbj0xKQ0KDQpzaW1zLmFycmF5IO+8mjPnu7TmlbDnu4Qgbi5rZWVwLCBuLmNoYWluc+WSjGNvbWJpbmVkIHBhcmFtZXRlciB2ZWN0b3LnmoTplb/luqYNCg0KbW9kZWwuZmlsZSDvvJogT3BlbkJVR1PnvJblhpnnmoQub2RjIOaooeWei+aWh+S7tg0KDQpESUMg77yaIOaYr+WQpuiuoeeul0RJQ+absue6vw0KDQpESUNPdXRwdXQg77yaIERJQ+WAvA0KDQpuLml0ZXIg77ya55Sf5oiQc2ltcy5hcnJheSDmr4/mnaFjaGFpbiDov63ku6PmlbANCg0Kbi5idXJuaW4g77ya5Lii5byD55qE6L+t5Luj5qyh5pWwDQoNCm4udGhpbiDvvJp0aGlubmluZyByYXRlDQoNCiMjIGF0dGFjaC5hbGwg5re75Yqg5pWw5o2u5Yiw5pCc57Si6Lev5b6EDQoNClRoZSBkYXRhYmFzZSBpcyBhdHRhY2hlZC9kZXRhY2hlZCB0byB0aGUgc2VhcmNoIHBhdGjvvIxXaGlsZSBhdHRhY2guYWxsIGF0dGFjaGVzIGFsbCBlbGVtZW50cyBvZiBhbiBvYmplY3QgeCB0byBhIGRhdGFiYXNlIGNhbGxlZCBuYW1lLCBhdHRhY2guYnVncw0KYXR0YWNoZXMgYWxsIGVsZW1lbnRzIG9mIHgkc2ltcy5saXN0IHRvIHRoZSBkYXRhYmFzZSBidWdzLnNpbXMgaXRzZWxmIG1ha2luZyB1c2Ugb2YgYXR0YWNoLmFsbC4NCg0KYXR0YWNoLmFsbCh4LCBvdmVyd3JpdGUgPSBOQSwgbmFtZSA9ICJhdHRhY2guYWxsIikNCmF0dGFjaC5idWdzKHgsIG92ZXJ3cml0ZSA9IE5BKQ0KZGV0YWNoLmFsbChuYW1lID0gImF0dGFjaC5hbGwiKQ0KZGV0YWNoLmJ1Z3MoKQ0KDQp4IDogYnVncyDlr7nosaENCg0Kb3ZlcndyaXRlIO+8mlRSVUUg5Yig6Zmk5YWo5bGA546v5aKD5Lit6KKr5o6p55uW55qE5pWw5o2u77yMIE5BIOivoumXru+8jEZBTFNFDQoNCm5hbWUgOiDnjq/looNuYW1lDQoNCiMjIGJ1Z3Mg5pyA6YeN6KaB77yM55SoUui/kOihjGJ1Z3MNCg0K6Ieq5Yqo6L6T5YWl5YC877yM5ZCv5YqoYnVnc++8jOS/neWtmOe7k+aenA0KDQpidWdzKGRhdGEsIGluaXRzLCBwYXJhbWV0ZXJzLnRvLnNhdmUsIG4uaXRlciwgICAgIG1vZGVsLmZpbGU9Im1vZGVsLnR4dCIsDQpuLmNoYWlucz0zLCBuLmJ1cm5pbj1mbG9vcihuLml0ZXIgLyAyKSwgbi50aGluPTEsDQpzYXZlRXhlYz1GQUxTRSxyZXN0YXJ0PUZBTFNFLA0KZGVidWc9RkFMU0UsIERJQz1UUlVFLCBkaWdpdHM9NSwgY29kYVBrZz1GQUxTRSwNCk9wZW5CVUdTLnBnbT1OVUxMLA0Kd29ya2luZy5kaXJlY3Rvcnk9TlVMTCwNCmNsZWFyV0Q9RkFMU0UsIHVzZVdJTkU9RkFMU0UsIFdJTkU9TlVMTCwNCm5ld1dJTkU9VFJVRSwgV0lORVBBVEg9TlVMTCwgYnVncy5zZWVkPTEsIHN1bW1hcnkub25seT1GQUxTRSwNCnNhdmUuaGlzdG9yeT0oLlBsYXRmb3JtJE9TLnR5cGUgPT0gIndpbmRvd3MiIHwgdXNlV0lORT09VFJVRSksDQpvdmVyLnJlbGF4ID0gRkFMU0UpDQoNCg0KZGF0YSDvvJrmqKHlnovkuK3kvb/nlKjnmoTmlbDmja4NCg0KaW5pdHMg77yabiBjaGFpbiDnmoTlhYPntKDliJfooajvvIzmr4/kuIDkuKropoHntKDmmK/kuIDkuKrmqKHlnovliJ3lp4vlgLzliJfooajvvIzmiJbogIXkuIDkuKrnlJ/miJDliJ3lp4vlgLzlvpdmdW5jdGlvbg0KDQpwYXJhbWV0ZXJzLnRvLnNhdmUg77yaIOmcgOimgeiiq+iusOW9leeahOWPguaVsOWQjeWQkemHjw0KDQptb2RlbC5maWxlIO+8miBtb2RlbCDmlofku7YudHh0DQoNCm4uY2hhaW5zIDog6buY6K6kM+adoQ0KDQpuLml0ZXIg77ya5q+P5p2h6ZO+55qE6L+t5Luj5qyh5pWw77yM6buY6K6kMjAwMA0KDQpuLnRoaW4g77yaIFRoaW5uaW5nIHJhdGUuIOato+aVtOaVsO+8jOm7mOiupOaYrzHvvIwNCg0Kc2F2ZUV4ZWMg77ya5L2/55SoYmFzZW5hbWUo5qih5Z6LLmZpbGUp5L+d5a2YT3BlbkJVR1PmiafooYznmoTph43mlrDlkK/liqjmmKDlg4/jgIINCg0KcmVzdGFydCDvvJrmiafooYzku47kuIrmrKHmiafooYznmoTmnIDlkI7nirbmgIHmgaLlpI3vvIzlrZjlgqjlnKjlt6XkvZznm67lvZXkuK3nmoQuYnVn5paH5Lu25Lit44CCDQoNCmRlYnVnIO+8miDpu5jorqRGQUxTRSzmraPlnKjov5DooYzooYzml7ZPcGVuYnVncyDpobXpnaLlhbPpl60NCg0KRElDIO+8muiuoeeul2RldmlhbmNlLO+8jHBE77yM5ZKMRElD44CCDQoNCmRpZ2l0cyDvvJrmnInmlYjlsI/mlbDkvY3mlbANCg0KY29kYVBrZyDvvJpGQUxTRSDov5Tlm54gYnVnc+Wvueixoe+8jOWQpuWImei+k+WHuu+8jOeUqGNvZGEg5YyFIHJlYWQuYnVncyDor7vlj5bvvIwNCg0KT3BlbkJVR1MucGdtIO+8mumAmuWQkU9wZW5CVUdT5Y+v5omn6KGM56iL5bqP55qE5a6M5pW06Lev5b6E44CCDQoNCndvcmtpbmcuZGlyZWN0b3J577yaT3BlbkJVR1PnmoTovpPlhaXlkozovpPlh7rlsIblrZjlgqjlnKjmraTnm67lvZXkuK07DQoNCmNsZWFyV0Qg77ya5piv5ZCm6L+Z5Lqb5paH5Lu255qE4oCcZGF0YS50eHTigJ3jgIHigJxpbml0KDE6bi5jaGFpbnMpLiB0eHTigJ0sImxvZy5vZGPigJ3jgIHigJxjb2RhSW5kZXgudHh04oCd5ZKM4oCcY29kYVsxOm5jaGFpbl0udHh04oCd57uT5p2f5pe25Yig6Zmk44CCDQoNCmJ1Z3Muc2VlZCDvvJpPcGVuQlVHU+maj+acuuenjeWtkO+8jDEtMTTmlbTmlbANCg0Kc3VtbWFyeS5vbmx5IO+8miBUVVJFICzku4Xlr7npnZ7luLjlv6vpgJ/nmoTliIbmnpDnu5nlh7rkuobkuIDkuKrlj4LmlbDmpoLopoENCg0Kc2F2ZS5oaXN0b3J5IDogVFVSRSzmnIDlkI7nlLvlh7p0cmFjZQ0KDQpgYGB7cn0NCiMgQW4gZXhhbXBsZSBtb2RlbCBmaWxlIGlzIGdpdmVuIGluOg0KbW9kZWwuZmlsZSA8LSBzeXN0ZW0uZmlsZShwYWNrYWdlPSJSMk9wZW5CVUdTIiwgIm1vZGVsIiwgInNjaG9vbHMudHh0IikNCiMgTGV0J3MgdGFrZSBhIGxvb2s6DQpwcmludChtb2RlbC5maWxlKQ0KZmlsZS5zaG93KG1vZGVsLmZpbGUpDQpgYGANCiAgIG1vZGVsIHsNCiAgICAgICBmb3IgKGogaW4gMTpKKXsNCiAgICAgICAgIHlbal0gfiBkbm9ybSAodGhldGFbal0sIHRhdS55W2pdKQ0KICAgICAgICAgdGhldGFbal0gfiBkbm9ybSAobXUudGhldGEsIHRhdS50aGV0YSkNCiAgICAgICAgIHRhdS55W2pdIDwtIHBvdyhzaWdtYS55W2pdLCAtMikNCiAgICAgICB9DQogICAgICAgbXUudGhldGEgfiBkbm9ybSAoMC4wLCAxLjBFLTYpDQogICAgICAgdGF1LnRoZXRhIDwtIHBvdyhzaWdtYS50aGV0YSwgLTIpDQogICAgICAgc2lnbWEudGhldGEgfiBkdW5pZiAoMCwgMTAwMCkNCiAgICAgfSANCiAgICAgDQpgYGB7cn0NCmRhdGEoc2Nob29scykNCnNjaG9vbHMNCmBgYA0KYGBge3J9DQpKIDwtIG5yb3coc2Nob29scykNCnkgPC0gc2Nob29scyRlc3RpbWF0ZQ0Kc2lnbWEueSA8LSBzY2hvb2xzJHNkDQpkYXRhIDwtIGxpc3QgKCJKIiwgInkiLCAic2lnbWEueSIpDQpkYXRhDQpgYGANCmBgYHtyfQ0KaW5pdHMgPC0gZnVuY3Rpb24oKXsNCiAgICAgIGxpc3QodGhldGE9cm5vcm0oSiwgMCwgMTAwKSwgbXUudGhldGE9cm5vcm0oMSwgMCwgMTAwKSxzaWdtYS50aGV0YT1ydW5pZigxLCAwLCAxMDApKQ0KfQ0KcGFyYW1ldGVycyA8LSBjKCJ0aGV0YSIsICJtdS50aGV0YSIsICJzaWdtYS50aGV0YSIpDQpgYGANCmBgYHtyfQ0Kc2Nob29scy5zaW0gPC0gYnVncyhkYXRhLCBpbml0cywgcGFyYW1ldGVycywgbW9kZWwuZmlsZSwNCm4uY2hhaW5zPTMsIG4uaXRlcj01MDAwKQ0KcHJpbnQoc2Nob29scy5zaW0pDQpgYGANCmBgYHtyfQ0KcGxvdChzY2hvb2xzLnNpbSkNCmBgYA0KDQojIyBidWdzLmRhdGEgICAg55Sf5oiQ6L6T5YWl5paH5Lu2DQoNCmJ1Z3MuZGF0YShkYXRhLCBkaXIgPSBnZXR3ZCgpLCBkaWdpdHMgPSA1LCBkYXRhLmZpbGUgPSAiZGF0YS50eHQiKQ0KDQojIyBidWdzLmluaXRzICAg55Sf5oiQ5Yid5aeL5YC85paH5Lu2DQoNCmJ1Z3MuaW5pdHMoaW5pdHMsIG4uY2hhaW5zLCBkaWdpdHMsDQogICAgaW5pdHMuZmlsZXMgPSBwYXN0ZSgiaW5pdHMiLCAxOm4uY2hhaW5zLCAiLnR4dCIsIHNlcCA9ICIiKSkNCiAgICANCiMjIGJ1Z3MubG9nICAg6K+75Y+WbG9n5paH5Lu277yIc3VtbWFyeSBzdGF0aXN0aWNzIGFuZCBESUMgaW5mb3JtYXRpb27vvIkNCg0KYnVncy5sb2coZmlsZSkNCg0KIyMgcGxvdC5idWdzICDnlLtidWdz5a+56LGhDQoNCg0KcGxvdCh4LCBkaXNwbGF5LnBhcmFsbGVsID0gRkFMU0UsIC4uLikNCg0KZGlzcGxheS5wYXJhbGxlbCDvvJrlnKjmkZjopoHlm77nmoTkuKTpg6jliIbkuK3mmL7npLrlubPooYznmoTpl7TpmpQNCg0KIyMgcHJpbnQuYnVncyDovpPlh7pidWdz5a+56LGhDQoNCnByaW50KHgsIGRpZ2l0cy5zdW1tYXJ5ID0gMSwgLi4uKQ0KDQpkaWdpdHMuc3VtbWFyee+8muWbm+iIjeS6lOWFpeeahOS9jeaVsA0KDQojIyByZWFkLmJ1Z3MgIA0KDQror7tNYXJrb3bpk77okpnnibnljaHnvZfovpPlh7rnmoRDT0RB5qC85byP44CC5bm26L+U5Zue5LiA5Liq57G7bWNtYy5saXN05a+56LGh44CC5L2/55SoY29kYeWMhei/m+ihjOi/m+S4gOatpeeahOi+k+WHuuWIhuaekOWIl+ihqOOAgg0KDQpyZWFkLmJ1Z3MoY29kYWZpbGVzLCAuLi4pDQoNCiMjIHZhbGlkYXRlSW5zdGFsbE9wZW5CVUdTIOavlOi+g1LlkoxvcGVuYnVnc+i9r+S7tui/kOihjOe7k+aenA0KDQojIyB3cml0ZS5tb2RlbCAg6L2s5YyWUiBmdW5jdGlvbuWIm+W7uuaooeWei+aWh+S7tg0KDQpgYGB7cn0NCnNjaG9vbHNtb2RlbCA8LSBmdW5jdGlvbigpew0KICAgICBmb3IgKGogaW4gMTpKKXsNCiAgICAgeVtqXSB+IGRub3JtICh0aGV0YVtqXSwgdGF1Lnlbal0pDQogICAgIHRoZXRhW2pdIH4gZG5vcm0gKG11LnRoZXRhLCB0YXUudGhldGEpDQogICAgIHRhdS55W2pdIDwtIHBvdyhzaWdtYS55W2pdLCAtMikNCn0NCiAgICAgbXUudGhldGEgfiBkbm9ybSAoMC4wLCAxLjBFLTYpDQogICAgIHRhdS50aGV0YSA8LSBwb3coc2lnbWEudGhldGEsIC0yKQ0KICAgICBzaWdtYS50aGV0YSB+IGR1bmlmICgwLCAxMDAwKQ0KfQ0KIyMgc29tZSB0ZW1wb3JhcnkgZmlsZW5hbWU6DQpmaWxlbmFtZSA8LSBmaWxlLnBhdGgodGVtcGRpcigpLCAic2Nob29sc21vZGVsLnR4dCIpDQojIyB3cml0ZSBtb2RlbCBmaWxlOg0Kd3JpdGUubW9kZWwoc2Nob29sc21vZGVsLCBmaWxlbmFtZSkNCiMjIGFuZCBsZXQncyB0YWtlIGEgbG9vazoNCmZpbGUuc2hvdyhmaWxlbmFtZSkNCmBgYA0KDQo=