# HG changeset patch # User spficklin # Date 1575656146 18000 # Node ID f2d2ec70b7d85a31855d57aff9b8073d931ff045 # Parent 96ba1a8fff064cab04e62cb7fff797747bd1ad3c Uploaded diff -r 96ba1a8fff06 -r f2d2ec70b7d8 aurora_wgcna_trait.Rmd --- a/aurora_wgcna_trait.Rmd Fri Dec 06 13:15:35 2019 -0500 +++ b/aurora_wgcna_trait.Rmd Fri Dec 06 13:15:46 2019 -0500 @@ -8,6 +8,9 @@ ```{r setup, include=FALSE, warning=FALSE, message=FALSE} knitr::opts_chunk$set(error = FALSE, echo = FALSE) ``` +```{r, include=FALSE} +options(tinytex.verbose = TRUE) +``` ```{r} # Load the data from the previous step. load(file=opt$r_data) @@ -64,7 +67,7 @@ hkeep = c() hignore = c() for (field in one_hot_cols[[i]]) { - + # Make sure the field is categorical. If it came in as integer it must be switched. if (trait_types[[field]] == "integer") { trait_data[[field]] = as.factor(trait_data[[field]]) @@ -74,7 +77,7 @@ print(field) next } - + # Now make sure we have enough factors. if (nlevels(trait_data[[field]]) > 1) { hkeep[length(hkeep)+1] = field @@ -82,17 +85,17 @@ hignore[length(hignore)+1] = field } } - + if (length(hignore) > 0) { print('These fields were ignored due to too few factors:') print(hignore) } - + # Perform the 1-hot encoding for specified and valid fields. if (length(hkeep) > 0) { print('These fields were be 1-hot encoded:') print(hkeep) - + swap_cols = colnames(trait_data)[(colnames(trait_data) %in% hkeep)] temp = as.data.frame(trait_data[, swap_cols]) colnames(temp) = swap_cols @@ -100,12 +103,12 @@ dmy <- dummyVars(" ~ .", data = temp) encoded <- data.frame(predict(dmy, newdata = temp)) encoded = sapply(encoded, as.integer) - + # Make a new trait_data table with these new 1-hot fields. keep_cols = colnames(trait_data)[!(colnames(trait_data) %in% one_hot_cols[[1]])] keep = as.data.frame(trait_data[, keep_cols]) colnames(keep) = keep_cols - + # Make a new trait_data object that has the columns to keep and the new 1-hot columns. trait_data = cbind(keep, encoded) }