function (content, measure, analysis, year, pmid, sortby, lefthand, 
    righthand, type, independent_variable, cofactorlabel, topic, 
    label_location, theme) 
{
    if (length(sortby) == 0 || sortby == "sortby") {
        sortby = "none"
    }
    first.row <- substr(content, 1, regexpr("\n", content))
    num.columns <- str_count(first.row, ",")
    num.cofactors <- num.columns - 8
    temp <- content
    temp <- gsub("\r", " ", fixed = TRUE, temp)
    temp <- gsub("\n", " ", fixed = TRUE, temp)
    temp <- gsub("\t", " ", fixed = TRUE, temp)
    temp <- gsub(",", "\",\"", fixed = TRUE, temp)
    temp <- paste("\"", temp, "\"", sep = "")
    temp <- paste("Mymatrix <- matrix(c(", temp, "), ncol=", 
        num.columns, ", byrow=TRUE)")
    x <- eval(parse(file = "", n = NULL, text = temp))
    first.row.header <- FALSE
    if (is.na(as.numeric(x[1, 2])) == TRUE) {
        first.row.header <- TRUE
    }
    if (first.row.header == TRUE) {
        x <- x[-c(1), ]
    }
    x <- x[!(is.na(as.numeric(x[, 2])) == TRUE), ]
    column.names <- c("Study", "year", "pmid", "registration", 
        "exp_events", "exp_total", "control_events", "control_total")
    for (i in 1:num.cofactors) {
        column.names <- append(column.names, paste("cofactor", 
            i, sep = ""))
    }
    dimnames(x) <- list(NULL, column.names)
    myframe <- data.frame(x)
    myframe <- na.omit(myframe)
    remove(x)
    myframe$Study <- gsub("'", "", fixed = TRUE, myframe$Study)
    myframe$Study <- as.character(str_trim(myframe$Study))
    myframe$year <- as.numeric(as.character(str_trim(myframe$year)))
    myframe$pmid <- as.numeric(as.character(str_trim(myframe$pmid)))
    myframe$registration <- as.character(str_trim(myframe$registration))
    myframe$registration <- ifelse(toupper(myframe$registration) %in% 
        c("NO"), "Registration: no", ifelse(toupper(myframe$registration) %in% 
        c("NA", ""), "Registration: unknown", "Registration: yes"))
    if (type == "subgroup.registration") {
        myframe$cofactor <- myframe$registration
    }
    if (type == "subgroup1" || independent_variable == "cf1") {
        myframe$cofactor <- myframe$cofactor1
    }
    if (type == "subgroup2" || independent_variable == "cf2") {
        myframe$cofactor <- myframe$cofactor2
    }
    if (type == "subgroup3" || independent_variable == "cf3") {
        myframe$cofactor <- myframe$cofactor3
    }
    if (type == "subgroup4" || independent_variable == "cf4") {
        myframe$cofactor <- myframe$cofactor4
    }
    if (type == "subgroup5" || independent_variable == "cf5") {
        myframe$cofactor <- myframe$cofactor5
    }
    if (type == "subgroup6" || independent_variable == "cf6") {
        myframe$cofactor <- myframe$cofactor6
    }
    if (type == "subgroup7" || independent_variable == "cf7") {
        myframe$cofactor <- myframe$cofactor7
    }
    if (type == "subgroup8" || independent_variable == "cf8") {
        myframe$cofactor <- myframe$cofactor8
    }
    if (type == "subgroup9" || independent_variable == "cf9") {
        myframe$cofactor <- myframe$cofactor9
    }
    if (type == "subgroup10" || independent_variable == "cf10") {
        myframe$cofactor <- myframe$cofactor10
    }
    PosParenth1 <- regexpr("(", myframe$exp_events[1], fixed = TRUE)
    if (PosParenth1 > 0) {
        PosParenth1 <- regexpr("(", myframe$exp_events, fixed = TRUE)
        PosParenth2 <- regexpr(")", myframe$exp_events, fixed = TRUE)
        myframe$exp_mean <- as.numeric(substring(myframe$exp_events, 
            1, PosParenth1 - 1))
        myframe$exp_sd <- as.numeric(substring(myframe$exp_events, 
            PosParenth1 + 1, PosParenth2 - 1))
        PosParenth1 <- regexpr("(", myframe$control_events, fixed = TRUE)
        PosParenth2 <- regexpr(")", myframe$control_events, fixed = TRUE)
        myframe$control_mean <- as.numeric(substring(myframe$control_events, 
            1, PosParenth1 - 1))
        myframe$control_sd <- as.numeric(substring(myframe$control_events, 
            PosParenth1 + 1, PosParenth2 - 1))
    }
    else {
        myframe$exp_events <- as.numeric(as.character(str_trim(myframe$exp_events)))
        myframe$control_events <- as.numeric(as.character(str_trim(myframe$control_events)))
    }
    myframe$exp_total <- as.numeric(as.character(str_trim(myframe$exp_total)))
    myframe$control_total <- as.numeric(as.character(str_trim(myframe$control_total)))
    if (sortby == "weight") {
        sortvalue <- NULL
    }
    if (sortby == "cofactor1") {
        sortvalue <- myframe$cofactor1
    }
    if (sortby == "cofactor2") {
        sortvalue <- myframe$cofactor2
    }
    if (sortby == "cofactor3") {
        sortvalue <- myframe$cofactor3
    }
    if (sortby == "cofactor4") {
        sortvalue <- myframe$cofactor4
    }
    if (sortby == "cofactor5") {
        sortvalue <- myframe$cofactor5
    }
    if (sortby == "cofactor6") {
        sortvalue <- myframe$cofactor6
    }
    if (sortby == "cofactor7") {
        sortvalue <- myframe$cofactor7
    }
    if (sortby == "cofactor8") {
        sortvalue <- myframe$cofactor8
    }
    if (sortby == "cofactor9") {
        sortvalue <- myframe$cofactor9
    }
    if (sortby == "cofactor10") {
        sortvalue <- myframe$cofactor10
    }
    if (sortby == "baseline") {
        if (PosParenth1 > 0) {
            sortvalue <- myframe$control_mean
        }
        else {
            sortvalue <- myframe$control_events/myframe$control_total
        }
    }
    if (sortby == "year") {
        sortvalue <- myframe$year
    }
    if (sortby == "study") {
        sortvalue <- myframe$Study
    }
    for (i in 1:length(myframe$Study)) {
        if (myframe$exp_events[i] == 0 & myframe$control_events[i] == 
            0) {
        }
    }
    attach(myframe)
    KUBlue = "#0022B4"
    SkyBlue = "#6DC6E7"
    pubbiastext = "Test for funnel plot asymmetry"
    if (analysis == "RE-Knapp-Hartung") {
        hartung = TRUE
        fixed = FALSE
        Random = TRUE
        analyticmethod = "Random effects model (Hartung-Knapp)"
    }
    else if (analysis == "RE") {
        hartung = FALSE
        fixed = FALSE
        Random = TRUE
        analyticmethod = "Random effects model"
    }
    else {
        hartung = FALSE
        fixed = TRUE
        Random = FALSE
        analyticmethod = "Fixed effects model"
    }
    if (type == "ignore") {
        if (measure %in% c("ROM", "MD", "SMD")) {
            meta1 <- metacont(exp_total, exp_mean, exp_sd, control_total, 
                control_mean, control_sd, data = myframe, sm = measure, 
                hakn = hartung, tau.common = TRUE, studlab = paste(Study, 
                  ", ", year, sep = ""))
            if (measure == "MD") {
                xlimits = "s"
            }
            else {
                xlimits = c(-2, 2)
            }
            if (length(myframe$Study) > 9) {
                pubbias = metabias(meta1, method.bias = "linreg", 
                  plotit = FALSE)
                pubbiastext = paste(pubbiastext, " (Egger): p= ", 
                  round(pubbias$p.value, 3), sep = "")
            }
            else {
                pubbiastext = paste(pubbiastext, ": too few studies to test", 
                  sep = "")
            }
        }
        else {
            meta1 <- metabin(exp_events, exp_total, control_events, 
                control_total, data = myframe, sm = measure, 
                hakn = hartung, tau.common = TRUE, method = "Inverse", 
                level = 0.95, incr = "TA", digits = 2, digits.se = 2, 
                allstudies = TRUE, studlab = paste(Study, ", ", 
                  year, sep = ""))
            xlimits = c(0.1, 10)
            if (length(myframe$Study) > 5) {
                meta1.as <- metabin(exp_events, exp_total, control_events, 
                  control_total, data = myframe, sm = "ASD", 
                  method = "I")
                pubbias = metabias(meta1.as, plotit = FALSE, 
                  k.min = 6)
                pubbiastext = paste(pubbiastext, " (Rucker): p= ", 
                  round(pubbias$p.value, 3), " (may be falsely significant if < 10 studies)", 
                  sep = "")
            }
            else {
                pubbiastext = paste(pubbiastext, ": too few studies to test", 
                  sep = "")
            }
        }
        if (sortby == "weight") {
            sortvalue <- 1/meta1$w.random
        }
        meta::forest(meta1, sortvalue = sortvalue, xlim = xlimits, 
            col.diamond = "blue", col.diamond.lines = "blue", 
            title = topic, fixed = FALSE, common = FALSE, random = TRUE, 
            resid.hetstat = TRUE, addrows = 2, print.I2.ci = TRUE, 
            print.tau2 = FALSE, print.p = FALSE, label.left = lefthand, 
            label.right = righthand, text.random = analyticmethod, 
            fs.random = 12, ff.random = 1, ff.hetstat = 2, fs.hetstat = 12)
        grid.text(topic, 0.5, 0.97, gp = gpar(fontsize = 14, 
            fontface = "bold"))
        grid.text(pubbiastext, 0.1, 0.02, hjust = 0, gp = gpar(fontsize = 12, 
            fontface = "bold"))
    }
    if (grepl("subgroup", type)) {
        myframe$cofactor <- gsub("'", "", fixed = TRUE, myframe$cofactor)
        myframe$cofactor <- as.character(str_trim(myframe$cofactor))
        if (measure %in% c("ROM", "MD", "SMD")) {
            meta1 <- metacont(exp_total, exp_mean, exp_sd, control_total, 
                control_mean, control_sd, tau.common = TRUE, 
                data = myframe, sm = measure, print.tau2 = FALSE, 
                hakn = hartung, studlab = paste(Study, ", ", 
                  year, sep = ""), label.left = lefthand, label.right = righthand, 
                title = topic, subgroup = myframe$cofactor, print.subgroup.name = FALSE)
            if (measure == "MD") {
                xlimits = "s"
            }
            else {
                xlimits = c(-2, 2)
            }
            if (length(myframe$Study) > 9) {
                meta1.egger <- metacont(exp_total, exp_mean, 
                  exp_sd, control_total, control_mean, control_sd, 
                  tau.common = TRUE, data = myframe, print.tau2 = FALSE, 
                  sm = measure)
                pubbias = metabias(meta1.egger, method.bias = "linreg", 
                  plotit = FALSE)
                pubbiastext = paste(pubbiastext, " (Egger): p= ", 
                  round(pubbias$p.value, 3), sep = "")
            }
            else {
                pubbiastext = paste(pubbiastext, ": too few studies to test", 
                  sep = "")
            }
        }
        else {
            meta1 <- metabin(exp_events, exp_total, control_events, 
                control_total, tau.common = TRUE, data = myframe, 
                sm = measure, method = "Inverse", hakn = hartung, 
                level = 0.95, incr = "TA", allstudies = TRUE, 
                studlab = paste(Study, ", ", year, sep = ""), 
                label.left = lefthand, label.right = righthand, 
                title = topic, subgroup = myframe$cofactor, print.subgroup.name = FALSE)
            xlimits = c(0.1, 10)
            if (length(myframe$Study) > 9) {
                meta1.as <- metabin(exp_events, exp_total, control_events, 
                  control_total, tau.common = TRUE, data = myframe, 
                  sm = "ASD", method = "I")
                pubbias = metabias(meta1.as, plotit = FALSE)
                pubbiastext = paste(pubbiastext, " (Rucker): p= ", 
                  round(pubbias$p.value, 3), sep = "")
            }
            else {
                pubbiastext = paste(pubbiastext, ": too few studies to test", 
                  sep = "")
            }
        }
        if (sortby == "weight") {
            sortvalue <- 1/meta1$w.random
        }
        forest(meta1, sortvalue = sortvalue, col.diamond = "blue", 
            col.diamond.lines = "blue", fixed = FALSE, common = FALSE, 
            random = TRUE, subgroup = TRUE, print.Q.subgroup = FALSE, 
            print.pval.Q = TRUE, resid.hetstat = TRUE, addrows = 2, 
            print.I2.ci = TRUE, print.tau2 = FALSE, print.Q = FALSE, 
            print.p = FALSE, label.left = lefthand, label.right = righthand, 
            text.random = analyticmethod, fs.random = 12, ff.random = 1, 
            ff.hetstat = 2, fs.hetstat = 12)
        grid.text(topic, 0.5, 0.97, gp = gpar(fontsize = 14, 
            fontface = "bold"))
        grid.text(pubbiastext, 0.1, 0.04, hjust = 0, gp = gpar(fontsize = 12, 
            fontface = "bold"))
    }
    if (type == "metaregression") {
        if (independent_variable == "year") {
            myframe$x <- as.numeric(myframe$year)
        }
        else {
            if (independent_variable == "size") {
                myframe$x <- as.numeric(myframe$exp_total) + 
                  as.numeric(myframe$control_total)
            }
            else {
                if (independent_variable == "size") {
                  myframe$x <- as.numeric(myframe$exp_total) + 
                    as.numeric(myframe$control_total)
                }
                else {
                  if (independent_variable == "cr" | independent_variable == 
                    "tr") {
                  }
                  else {
                    myframe$x <- as.numeric(as.character(str_trim(myframe$cofactor)))
                  }
                }
            }
        }
        attach(myframe)
        if (measure %in% c("MD", "SMD")) {
            if (independent_variable == "cr") {
                myframe$x <- myframe$control_mean
            }
            if (independent_variable == "tr") {
                myframe$x <- myframe$exp_total/myframe$control_total
            }
            myframe[order(myframe$x, na.last = NA), ]
            myframe <- na.omit(myframe)
            dat <- escalc(measure = "MD", m1i = exp_mean, sd1i = exp_sd, 
                n1i = exp_total, m2i = control_mean, sd2i = control_sd, 
                n2i = control_total, data = myframe)
            res <- rma.uni(yi, vi, mods = ~myframe$x, method = "DL", 
                knha = TRUE, data = dat, intercept = TRUE)
            ylabel = "Mean difference"
        }
        else {
            if (independent_variable == "cr") {
                myframe$x <- myframe$control_events/myframe$control_total
            }
            if (independent_variable == "tr") {
                myframe$x <- myframe$exp_total/myframe$control_total
            }
            myframe[order(myframe$x, na.last = NA), ]
            myframe <- na.omit(myframe)
            dat <- escalc(measure = "OR", ai = exp_events, bi = exp_total - 
                exp_events, ci = control_events, di = control_total - 
                control_events, data = myframe)
            res <- rma(yi, vi, mods = ~myframe$x, data = dat)
            ylabel = "Odds ratio transformed to natural log (ln)"
        }
        cofactor.range = seq(min(myframe$x), max(myframe$x), 
            (max(myframe$x) - min(myframe$x))/100)
        preds <- predict(res, newmods = cofactor.range)
        preds <- data.frame(cofactor.range, preds$pred, preds$ci.lb, 
            preds$ci.ub)
        wi <- 1/sqrt(dat$vi)
        size <- 0.5 + 3 * (wi - min(wi))/(max(wi) - min(wi))
        plot(dat$x, dat$yi, pch = 19, cex = size, cex.lab = 1.5, 
            font.axis = 2, xlab = "", ylab = ylabel, main = paste("Meta-regression of ", 
                topic), xlim = c(min(dat$x) - 0.1 * (max(dat$x) - 
                min(dat$x)), max(dat$x) + 0.1 * (max(dat$x) - 
                min(dat$x))), ylim = c(min(dat$yi), max(dat$yi) + 
                0.15 * (max(dat$yi) - min(dat$yi))), las = 1, 
            bty = "l")
        if (cofactorlabel != "") {
            mtext(side = 1, line = 2.25, paste("Cofactor: ", 
                cofactorlabel), font = 2, cex = 1.5)
        }
        lines(preds$cofactor.range, preds$preds.pred)
        lines(preds$cofactor.range, preds$preds.ci.lb, lty = "dashed", 
            col = "blue")
        lines(preds$cofactor.range, preds$preds.ci.ub, lty = "dashed", 
            col = "blue")
        text(par("usr")[2], par("usr")[4] - 1.2 * strheight("A"), 
            cex = 1, adj = c(1, 0), paste("R2 = ", round(res$R2), 
                "% (QM = ", sprintf(res$QM, fmt = "%#.1f"), ", p = ", 
                sprintf(res$pval[2], fmt = "%#.3f"), ")", sep = ""), 
            font = 1)
        text(par("usr")[2], par("usr")[4] - 1.2 * strheight("A") - 
            1.4 * strheight("A"), cex = 1, adj = c(1, 0), paste("Regression coefficient = ", 
            sprintf(res$b[2], fmt = "%#.1f"), sep = ""), font = 1)
        text(par("usr")[2], par("usr")[4] - 1.2 * strheight("A") - 
            2.8 * strheight("A"), cex = 1, adj = c(1, 0), paste("Residual I2 = ", 
            sprintf(res$I2, fmt = "%#.1f"), "%", sep = ""), font = 1)
        abline(h = 0, lty = "dotted")
        if (label_location > 0) {
            text(dat$x, dat$yi, paste(dat$Study, ", ", dat$year, 
                sep = ""), cex = 0.9, pos = label_location, offset = 1, 
                col = dat$color)
        }
        legend("topleft", adj = 0, xjust = 1, inset = c(0, 0), 
            c("Regression line", "95% Confidence\ninterval"), 
            pch = NULL, pt.bg = "white", bty = "n", border = "white", 
            lty = c("solid", "dashed"), col = c("black", "blue"))
    }
    if (type == "funnel") {
        if (measure %in% c("ROM", "MD", "SMD")) {
            meta1 <- metacont(exp_total, exp_mean, exp_sd, control_total, 
                control_mean, control_sd, data = myframe, sm = measure, 
                hakn = hartung, studlab = paste(Study, ", ", 
                  year, sep = ""), label.left = lefthand, label.right = righthand, 
                title = topic, subgroup = myframe$cofactor, print.subgroup.name = FALSE)
        }
        else {
            meta1 <- metabin(exp_events, exp_total, control_events, 
                control_total, data = myframe, sm = measure, 
                method = "Inverse", tau.common = TRUE, hakn = hartung, 
                level = 0.95, incr = "TA", allstudies = TRUE, 
                studlab = paste(Study, ", ", year, sep = ""), 
                label.left = lefthand, label.right = righthand, 
                title = topic, subgroup = myframe$cofactor, print.subgroup.name = FALSE)
        }
        funnel(meta1, common = TRUE)
        if (length(meta1$studlab) > 5) {
            meta1.as <- metabin(exp_events, exp_total, control_events, 
                control_total, tau.common = TRUE, data = myframe, 
                sm = "ASD", method = "I")
            pubbias = metabias(meta1.as, plotit = FALSE, k.min = 6)
            pubbiastext = paste(pubbiastext, ":\np (Rucker) = ", 
                round(pubbias$p.value, 3), " (may be falsely significant if < 10 studies)", 
                sep = "")
        }
        else {
            pubbiastext = paste(pubbiastext, ":\ntoo few studies to test", 
                sep = "")
        }
        grid.text(topic, 0.5, 0.97, gp = gpar(fontsize = 14, 
            fontface = "bold"))
        grid.text(pubbiastext, 0.125, 0.8, hjust = 0, gp = gpar(fontsize = 12, 
            fontface = "bold"))
    }
}
<bytecode: 0x559a4f3f5468>
<environment: namespace:home>