catching pandoc dependency and giving informative error

R/BinomialGLMsClass.R

@@ -148,7 +148,9 @@ BinomialGLM <- R6Class(classname = "BinomialGLM",
       self$ParentModel <- ParentModel
       self$classify <- ParentModel$classify
       self$model_contrl <- ParentModel$model_contrl
-      if (!("glm" %in% fit.algorithm)) warning("over-riding fit.algorithm option with 'glm', since fit.package was set to 'speedglm' or 'glm'")
+
+      # if (!("glm" %in% fit.algorithm)) warning("over-riding fit.algorithm option with 'glm', since fit.package was set to 'speedglm' or 'glm'")
+
       assert_that(any(c("glm", "speedglm") %in% fit.package))
       self$fit.class <- fit.package
       class(self$fit.class) <- c(class(self$fit.class), self$fit.class)

R/BinomialMLclass.R

@@ -218,7 +218,7 @@ BinomialH2O  <- R6Class(classname = "BinomialH2O",
   portable = TRUE,
   class = TRUE,
   public = list(
-    fit.class = c("glm", "randomForest", "gbm", "deeplearning", "SuperLearner"),
+    fit.class = character(),
     model.fit = list(coef = NA, fitfunname = NA, linkfun = NA, nobs = NA, params = NA, H2O.model.object = NA),
     outfactors = NA,
     nfolds = 5,
@@ -261,7 +261,6 @@ Please type this into the R terminal:
 
       self$fit.class <- fit.algorithm
       class(self$fit.class) <- c(class(self$fit.class), "h2o" %+% self$fit.class)
-      # class(self$model.fit) <- c(class(self$model.fit), "h2o" %+% self$fit.class)
       invisible(self)
     },
 

R/BinomialModelClass.R

@@ -56,8 +56,8 @@ BinaryOutcomeModel  <- R6Class(classname = "BinaryOutcomeModel",
     is.fitted = FALSE,
 
     binomialModelObj = NULL, # object of class binomialModelObj that is used in fitting / prediction, never saved (need to be initialized with $new())
-    fit.package = c("speedglm", "glm", "h2o"),
-    fit.algorithm = c("glm", "gbm", "randomForest", "deeplearning", "SuperLearner"),
+    fit.package = character(),
+    fit.algorithm = character(),
     model_contrl = list(),
 
     n = NA_integer_,        # number of rows in the input data
@@ -77,15 +77,15 @@ BinaryOutcomeModel  <- R6Class(classname = "BinaryOutcomeModel",
       } else {
         self$fit.package <- reg$fit.package[1]
       }
-      if (!(self$fit.package %in% c("speedglm", "glm", "h2o"))) stop("fit.package must be one of: 'speedglm', 'glm', 'h2o'")
+      if (!(self$fit.package %in% allowed.fit.package)) stop("fit.package must be one of: " %+% paste0(allowed.fit.package, collapse=", "))
 
       if ("fit.algorithm" %in% names(self$model_contrl)) {
         self$fit.algorithm <- self$model_contrl[['fit.algorithm']]
         assert_that(is.character(self$fit.algorithm))
       } else {
         self$fit.algorithm <- reg$fit.algorithm[1]
       }
-      if (!(self$fit.algorithm %in% c("glm", "gbm", "randomForest", "deeplearning", "SuperLearner"))) stop("fit.algorithm must be one of: 'glm', 'gbm', 'randomForest', 'deeplearning', 'SuperLearner'")
+      if (!(self$fit.algorithm %in% allowed.fit.algorithm)) stop("fit.algorithm must be one of: " %+% paste0(allowed.fit.algorithm, collapse=", "))
 
       assert_that(is.string(reg$outvar))
       self$outvar <- reg$outvar
@@ -140,7 +140,7 @@ BinaryOutcomeModel  <- R6Class(classname = "BinaryOutcomeModel",
 
       if (inherits(model.fit, "try-error")) {
         message("running " %+% self$binomialModelObj$fit.class %+% " with h2o has failed, trying to run speedglm as a backup...")
-        self$binomialModelObj <- BinomialGLM$new(fit.algorithm = "GLM", fit.package = "speedglm", ParentModel = self, ...)
+        self$binomialModelObj <- BinomialGLM$new(fit.algorithm = "glm", fit.package = "speedglm", ParentModel = self, ...)
         self$binomialModelObj$params <- list(outvar = self$outvar, predvars = self$predvars, stratify = self$subset_exprs)
         model.fit <- self$binomialModelObj$fit(data, self$outvar, self$predvars, self$subset_idx, ...)
       }

R/reports.R

@@ -2,6 +2,7 @@
 # @importFrom pander evalsOptions
 NULL
 
+# nocov start
 #' Open file
 #'
 #' Tries to open a file with operating system's default program.
@@ -31,6 +32,7 @@ openFileInOS <- function(f) {
     }
   }
 }
+# nocov end
 
 # ---------------------------------------------------------------------------------------------
 #' Generate report(s) with modeling stats and survival estimates using pandoc.
@@ -67,6 +69,14 @@ make_report_rmd <- function(OData, MSM, NPMSM, TMLE, GCOMP, wts_data, SurvByRegi
                             file.name = getOption('stremr.file.name'), file.path = getOption('stremr.file.path'),
                             openFile = TRUE, keep_md = FALSE, keep_tex = FALSE, ...) {
   optArgReport <- list(...)
+
+  if (!rmarkdown::pandoc_available(version = "1.12.3"))
+    stop(
+"Report functionality requires pandoc (version 1.12.3 or higher).
+Please install it.
+For more information, go to: http://pandoc.org/installing.html",
+call. = FALSE)
+
   if ("author" %in% names(optArgReport)) {
     author <- optArgReport[['author']]
     assert_that(is.character(author))

R/zzz.R

@@ -11,6 +11,10 @@ gvars$tolerr <- 10^-12      # tolerance error: assume for abs(a-b) < gvars$toler
 gvars$sVartypes <- list(bin = "binary", cat = "categor", cont = "contin")
 gvars$noCENScat <- 0L       # the reference category that designates continuation of follow-up
 
+allowed.fit.package <- c("speedglm", "glm", "h2o")
+allowed.fit.algorithm = c("glm", "gbm", "randomForest", "deeplearning", "SuperLearner")
+allowed.bin.method = c("equal.mass", "equal.len", "dhist")
+
 #' Querying/setting a single \code{stremr} option
 #'
 #' To list all \code{stremr} options, just run this function without any parameters provided. To query only one value, pass the first parameter. To set that, use the \code{value} parameter too.
@@ -69,7 +73,7 @@ print_stremr_opts <- function() {
 #' The preferred way to set options for \code{stremr} is to use \code{\link{stremrOptions}}, which allows specifying individual options without having to reset all other options.
 #' To reset all options to their defaults simply run \code{set_all_stremr_options()} without any parameters/arguments.
 #' @param fit.package Specify the default package for performing model fitting: c("speedglm", "glm", "h2o")
-#' @param fit.algorithm Specify the default fitting algorithm: c("glm", "gbm", "randomForest", "SuperLearner")
+#' @param fit.algorithm Specify the default fitting algorithm: c("glm", "gbm", "randomForest", "deeplearning", "SuperLearner")
 #' @param bin.method The method for choosing bins when discretizing and fitting the conditional continuous summary
 #'  exposure variable \code{sA}. The default method is \code{"equal.len"}, which partitions the range of \code{sA}
 #'  into equal length \code{nbins} intervals. Method \code{"equal.mass"} results in a data-adaptive selection of the bins
@@ -113,17 +117,13 @@ set_all_stremr_options <- function( fit.package = c("speedglm", "glm", "h2o"),
                             ) {
 
   old.opts <- gvars$opts
-  bin.method <- bin.method[1L]
 
   fit.package <- fit.package[1L]
-  assert_that(fit.package %in% c("speedglm", "glm", "h2o"))
-
   fit.algorithm <- fit.algorithm[1L]
-  if (!(fit.algorithm %in% c("glm", "gbm", "randomForest", "deeplearning", "SuperLearner"))) stop("fit.algorithm must be one of: 'glm', 'gbm', 'randomForest', 'deeplearning', 'SuperLearner'")
-
-  if (!(bin.method %in% c("equal.len", "equal.mass", "dhist"))) {
-    stop("bin.method argument must be either 'equal.len', 'equal.mass' or 'dhist'")
-  }
+  bin.method <- bin.method[1]
+  if (!(fit.package %in% allowed.fit.package)) stop("fit.package must be one of: " %+% paste0(allowed.fit.package, collapse=", "))
+  if (!(fit.algorithm %in% allowed.fit.algorithm)) stop("fit.algorithm must be one of: " %+% paste0(allowed.fit.algorithm, collapse=", "))
+  if (!(bin.method %in% allowed.bin.method)) stop("bin.method must be one of: " %+% paste0(allowed.bin.method, collapse=", "))
 
   opts <- list(
     fit.package = fit.package,

tests/RUnit/RUnit_tests_04a_sims10K_speedglm.R

@@ -173,25 +173,28 @@ test.speedglm.allestimators10Kdata <- function() {
   # report.path <- "/home/ubuntu/stremr_example"
   # file.path = report.path,
   # test for opening file in local OS
-  make_report_rmd(OData, file.name = "sim.data.example.fup2", title = "Custom", author = "Jane Doe", openFile = FALSE)
+  if (rmarkdown::pandoc_available(version = "1.12.3"))
+    make_report_rmd(OData, file.name = "sim.data.example.fup2", title = "Custom", author = "Insert Author Name", openFile = FALSE)
 
-  make_report_rmd(OData, NPMSM = list(surv1, surv2), MSM = MSM.IPAW, GCOMP = list(gcomp_est1, gcomp_est2), TMLE = list(tmle_est_par1, tmle_est_par2),
+  if (rmarkdown::pandoc_available(version = "1.12.3"))
+    make_report_rmd(OData, NPMSM = list(surv1, surv2), MSM = MSM.IPAW, GCOMP = list(gcomp_est1, gcomp_est2), TMLE = list(tmle_est_par1, tmle_est_par2),
                   format = "html",
                   AddFUPtables = TRUE,
                   openFile = FALSE,
                   MSM.RDtables = get_MSM_RDs(MSM.IPAW, t.periods.RDs = c(12, 15), getSEs = TRUE),
                   TMLE.RDtables = get_TMLE_RDs(list(tmle_est_par1, tmle_est_par2), t.periods.RDs = c(1, 4)),
                   WTtables = get_wtsummary(MSM.IPAW$wts_data, cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                  file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Author Name", y_legend = 0.99, x_legend = 9.5)
+                  file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.99, x_legend = 9.5)
 
-  make_report_rmd(OData, NPMSM = list(surv1, surv2), MSM = MSM.IPAW, TMLE = list(tmle_est_par1, tmle_est_par2),
+  if (rmarkdown::pandoc_available(version = "1.12.3"))
+    make_report_rmd(OData, NPMSM = list(surv1, surv2), MSM = MSM.IPAW, TMLE = list(tmle_est_par1, tmle_est_par2),
                   format = "pdf",
                   AddFUPtables = TRUE,
                   openFile = FALSE,
                   MSM.RDtables = get_MSM_RDs(MSM.IPAW, t.periods.RDs = c(12, 15), getSEs = TRUE),
                   TMLE.RDtables = get_TMLE_RDs(list(tmle_est_par1, tmle_est_par2), t.periods.RDs = c(1, 4)),
                   WTtables = get_wtsummary(MSM.IPAW$wts_data, cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                  file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Author Name", y_legend = 0.99, x_legend = 9.5)
+                  file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.99, x_legend = 9.5)
 
   # ---------------------------------------------------------------------------------------------------------
   # TMLE / GCOMP with a stochastic intervention on MONITOR

tests/RUnit/RUnit_tests_04b_sims10K_h2oGLM.R

@@ -84,11 +84,12 @@ test.h2oglm.IPW.MSM.10Kdata <- function() {
                         est_name = "IPAW", getSEs = TRUE)
     # names(MSM.IPAW)
     # MSM.IPAW$St
-    make_report_rmd(OData, MSM = MSM.IPAW,
+    if (rmarkdown::pandoc_available(version = "1.12.3"))
+        make_report_rmd(OData, MSM = MSM.IPAW,
                     AddFUPtables = TRUE, openFile = FALSE,
                     RDtables = get_MSM_RDs(MSM.IPAW, t.periods.RDs = c(12, 15), getSEs = TRUE),
                     WTtables = get_wtsummary(MSM.IPAW$wts_data, cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Jane Doe", y_legend = 0.95)
+                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.95)
   }
 
 }

tests/RUnit/RUnit_tests_04c_sims10K_h2o_ALL_ML.R

@@ -91,11 +91,12 @@ test.h2o.ALL.ML.allestimators10Kdata <- function() {
     wts.St.dhigh <- getIPWeights(OData, intervened_TRT = "gTI.dhigh")
     surv2 <- survNPMSM(wts.St.dhigh, OData)
 
-    make_report_rmd(OData, NPMSM = list(surv1, surv2), wts_data = list(wts.St.dlow, wts.St.dhigh),
+    if (rmarkdown::pandoc_available(version = "1.12.3"))
+        make_report_rmd(OData, NPMSM = list(surv1, surv2), wts_data = list(wts.St.dlow, wts.St.dhigh),
                     AddFUPtables = TRUE,
                     openFile = FALSE,
                     WTtables = get_wtsummary(list(wts.St.dlow, wts.St.dhigh), cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Jane Doe", y_legend = 0.95)
+                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.95)
 
     # ----------------------------------------------------------------
     # FIT PROPENSITY SCORES WITH gbm
@@ -109,11 +110,12 @@ test.h2o.ALL.ML.allestimators10Kdata <- function() {
     wts.St.dhigh <- getIPWeights(OData, intervened_TRT = "gTI.dhigh")
     surv2 <- survNPMSM(wts.St.dhigh, OData)
 
-    make_report_rmd(OData, NPMSM = list(surv1, surv2), wts_data = list(wts.St.dlow, wts.St.dhigh),
+    if (rmarkdown::pandoc_available(version = "1.12.3"))
+        make_report_rmd(OData, NPMSM = list(surv1, surv2), wts_data = list(wts.St.dlow, wts.St.dhigh),
                     AddFUPtables = TRUE,
                     openFile = FALSE,
                     WTtables = get_wtsummary(list(wts.St.dlow, wts.St.dhigh), cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Jane Doe", y_legend = 0.95)
+                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.95)
 
     # ----------------------------------------------------------------
     # FIT PROPENSITY SCORES WITH deeplearning
@@ -127,11 +129,12 @@ test.h2o.ALL.ML.allestimators10Kdata <- function() {
     wts.St.dhigh <- getIPWeights(OData, intervened_TRT = "gTI.dhigh")
     surv2 <- survNPMSM(wts.St.dhigh, OData)
 
-    make_report_rmd(OData, NPMSM = list(surv1, surv2), wts_data = list(wts.St.dlow, wts.St.dhigh),
+    if (rmarkdown::pandoc_available(version = "1.12.3"))
+        make_report_rmd(OData, NPMSM = list(surv1, surv2), wts_data = list(wts.St.dlow, wts.St.dhigh),
                     AddFUPtables = TRUE,
                     openFile = FALSE,
                     WTtables = get_wtsummary(list(wts.St.dlow, wts.St.dhigh), cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Jane Doe", y_legend = 0.95)
+                    file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.95)
 
     # ---------------------------------------------------------------------------------------------------------
     # TMLE w/ h2o random forest

tests/RUnit/RUnit_tests_05_stremr_example_SL.R

@@ -177,11 +177,12 @@ test.h2oEnsemble <- function() {
 
         # report.path <- "/set/your/report/path/"
         # file.path = report.path
-        make_report_rmd(OData, openFile = FALSE,
+        if (rmarkdown::pandoc_available(version = "1.12.3"))
+            make_report_rmd(OData, openFile = FALSE,
                         # AddFUPtables = TRUE,
                         # MSM.RDtables = get_MSM_RDs(MSM.IPAW, t.periods.RDs = c(12, 15), getSEs = FALSE),
                         # WTtables = get_wtsummary(MSM.IPAW$wts_data, cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                        file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Oleg Sofrygin", y_legend = 0.95)
+                        file.name = "sim.data.example.fup", title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.95)
 
         # ---------------------------------------------------------------------------------------------------------
         # ERROR CHECK: no hyper params for randomForest:

tests/RUnit/RUnit_tests_06_direct_regression_def.R

@@ -198,13 +198,16 @@ notest.savedSL.10Kdata <- function() {
                 IPW_estimates
     St.dhigh3
 
-    report.path <- "/Users/olegsofrygin/Dropbox/KP/monitoring_simstudy/stremr_examples"
-    make_report_rmd(OData, openFile = FALSE,
+    # report.path <- "/Users/olegsofrygin/Dropbox/KP/monitoring_simstudy/stremr_examples"
+    if (rmarkdown::pandoc_available(version = "1.12.3"))
+      make_report_rmd(OData, openFile = FALSE,
                     # MSM = MSM.IPAW,
                     # AddFUPtables = TRUE,
                     # MSM.RDtables = get_MSM_RDs(MSM.IPAW, t.periods.RDs = c(12, 15), getSEs = FALSE),
                     # WTtables = get_wtsummary(MSM.IPAW$wts_data, cutoffs = c(0, 0.5, 1, 10, 20, 30, 40, 50, 100, 150), by.rule = TRUE),
-                    file.name = "sim.data.example.fup", file.path = report.path, title = "Custom Report Title", author = "Oleg Sofrygin", y_legend = 0.95)
+                    file.name = "sim.data.example.fup",
+                    # file.path = report.path,
+                    title = "Custom Report Title", author = "Insert Author Name", y_legend = 0.95)
 
     h2o::h2o.shutdown(prompt = FALSE)
   }

tests/examples/4_survMSM_example.R

@@ -88,6 +88,8 @@ IPW_MSM_res$MSM.fit
 
 # ------------------------------------------------------------------
 # Generate automatic html report with results of the analysis
+# This assumes that pandoc is already installed
+# For more information, go to: http://pandoc.org/installing.html
 # ------------------------------------------------------------------
 \dontrun{
 make_report_rmd(OData, MSM = IPW_MSM_res,