completely fixed constructConDfAndParseTCR

R/combineContigs.R

@@ -82,20 +82,19 @@ combineTCR <- function(input.data,
       }
     }
     if (!is.null(samples)) {
-        out <- .modifyBarcodes(input.data, samples, ID)
+      out <- .modifyBarcodes(input.data, samples, ID)
     } else {
       out <- input.data
     }
     for (i in seq_along(out)) {
-        data2 <- out[[i]]
-        data2 <- .makeGenes(cellType = "T", data2)
+        data2 <- .makeGenes(cellType = "T", out[[i]])
         Con.df <- .constructConDfAndParseTCR(data2)
         Con.df <- .assignCT(cellType = "T", Con.df)
         Con.df[Con.df == "NA_NA" | Con.df == "NA;NA_NA;NA"] <- NA 
         data3 <- merge(data2[,-which(names(data2) %in% c("TCR1","TCR2"))], 
             Con.df, by = "barcode")
         if (!is.null(samples) & !is.null(ID)) {
-            data3<-data3[, c("barcode","sample","ID",tcr1_lines,tcr2_lines,
+            data3 <- data3[, c("barcode", "sample", "ID", tcr1_lines, tcr2_lines,
                 CT_lines)] }
         else if (!is.null(samples) & is.null(ID)) {
           data3<-data3[,c("barcode","sample",tcr1_lines,tcr2_lines,

R/utils.R

@@ -283,7 +283,7 @@ is_seurat_or_se_object <- function(obj) {
 .constructConDfAndParseTCR <- function(data2) {
   rcppConstructConDfAndParseTCR(
     data2 %>% arrange(., chain, cdr3_nt),
-    unique(data2$barcode)
+    unique(data2[[1]])
   )
 }
 
@@ -400,8 +400,7 @@ is_seurat_or_se_object <- function(obj) {
                   str_c(str_replace_na(v_gene),  str_replace_na(j_gene), str_replace_na(c_gene), sep = "."), NA)) %>%
             mutate(TCR2 = ifelse(chain %in% c("TRB", "TRD"), 
                   str_c(str_replace_na(v_gene), str_replace_na(d_gene),  str_replace_na(j_gene),  str_replace_na(c_gene), sep = "."), NA))
-    }
-    else { # assume BCR (`c("B")`)
+    } else if (cellType %in% c("B")) {
         heavy <- data2[data2$chain == "IGH",] %>% 
           mutate(IGHct = str_c(str_replace_na(v_gene), str_replace_na(d_gene),  str_replace_na(j_gene),  str_replace_na(c_gene), sep = "."))
         kappa <- data2[data2$chain == "IGK",] %>% 

src/constructConDfAndparseTCR.cpp

@@ -26,7 +26,7 @@ std::vector<std::vector<int>> constructBarcodeIndex(
 
 class TcrParser {
 public:
-    // variable for the eventual output conDf
+    // variable for the eventual output Con.df
     std::vector<std::vector<std::string>> conDf;
 
     // variables for *references* to columns on data2
@@ -39,6 +39,8 @@ class TcrParser {
     // variable for helper barcode index
     std::vector<std::vector<int>> barcodeIndex;
 
+    // constructor: in the future if more columns are ever added its probably a much
+    // more general algo would be to just find the indicies of the dataframe by looking at the colnames
     TcrParser(
         Rcpp::DataFrame& data2, std::vector<std::string>& uniqueData2Barcodes
     ) {
@@ -48,19 +50,22 @@ class TcrParser {
         );
         conDf[0] = uniqueData2Barcodes;
 
-        // set references to data2 columns
+        // set references to fixed data2 columns
         data2ChainTypes = data2[5];
-        data2Tcr1 = data2[19];
-        data2Tcr2 = data2[20];
         data2Cdr3 = data2[12];
         data2Cdr3Nt = data2[13];
 
-        // barcodeIndex
+        // setting reference to the TCR columns assuming all extra columns come before
+        data2Tcr1 = data2[data2.size() - 2];
+        data2Tcr2 = data2[data2.size() - 1]; 
+
+        // construct barcodeIndex
         barcodeIndex = constructBarcodeIndex(
             uniqueData2Barcodes, Rcpp::as<std::vector<std::string>>(data2[0])
         );
     }
 
+    // Rcpp implementation of .parseTCR()
     void parseTCR() {
         for (int y = 0; y < (int) conDf[0].size(); y++) {
             for (int index : barcodeIndex[y]) {

tests/testthat/test-utils.R

@@ -93,10 +93,17 @@ test_that(".theCall works", {
 })
 #TODO .theCall Add custom header
 
-# TODO .constructConDfAndParseTCR !!!! Need to use testing data from the old version
-
 test_that(".constructConDfAndParseTCR works", {
-    # TODO create testdata with the original .parseTCR and test here. also do edgecases
+  expect_identical(
+    .constructConDfAndParseTCR(
+      getdata("utils", "constructConDfAndParseTCRInput")
+    ),
+    getdata("utils", "expected_con_df")
+  )
+
+# TODO: add more cases! This is not comprehensive.
+# there was a case in the past where this passed but
+# the function caused a segmentation fault.
 })
 
 # TODO .parseBCR