Add low-level utilities read_auxdata() and write_auxdata().

Also the 'auxiliary_data' argument of igblastn() can now be a data.frame
in addition to the path to a file containing auxiliary data (.aux file).

Author: hpages

DESCRIPTION

@@ -51,6 +51,7 @@ Collate:
 	ndm_data-IO.R
 	compute_V_gene_delineations.R
 	clean_allele_set.R
+	auxdata-IO.R
 	compute_auxdata.R
 	LATIN_NAMES.R
 	IMGT-utils.R

NAMESPACE

@@ -91,6 +91,10 @@ export(
     compute_V_gene_delineations,
     compute_imgt_intdata,
 
+    ## auxdata-IO.R:
+    read_auxdata,
+    write_auxdata,
+
     ## compute_auxata.R:
     compute_auxdata,
 

R/auxdata-IO.R

@@ -0,0 +1,62 @@
+### =========================================================================
+### Read/write ".aux" files
+### -------------------------------------------------------------------------
+###
+
+
+### Not exported!
+### Not the true colnames used in IgBLAST ".aux" files: ours are shorter,
+### all lowercase, and contain underscores instead of spaces.
+AUXDATA_COL2CLASS <- c(
+    allele_name="character",
+    coding_frame_start="integer",
+    chain_type="character",
+    cdr3_end="integer",
+    extra_bps="integer"
+)
+
+
+### - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+### check_auxdata_col2class()
+###
+
+### Not exported!
+check_auxdata_col2class <- function(auxdata, what="'auxdata'")
+{
+    if (!is.data.frame(auxdata))
+        stop(wmsg(what, " must be a data.frame"))
+    expected_colnames <- names(AUXDATA_COL2CLASS)
+    if (!identical(colnames(auxdata), expected_colnames)) {
+        in1string <- paste(expected_colnames, collapse=", ")
+        stop(wmsg(what, " must have the following columns ",
+                  "(in this order): ", in1string))
+    }
+    col2class <- vapply(auxdata, function(x) class(x)[[1L]], character(1))
+    if (!identical(col2class, AUXDATA_COL2CLASS)) {
+        in1string <- paste0("    ", names(AUXDATA_COL2CLASS), " -> ",
+                            AUXDATA_COL2CLASS, collapse="\n")
+        stop(wmsg(what, " must have the following column types:"), "\n",
+             in1string)
+    }
+}
+
+
+### - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+### read_auxdata()
+### write_auxdata()
+###
+
+read_auxdata <- function(filepath)
+{
+    read_broken_table(filepath, AUXDATA_COL2CLASS)
+}
+
+write_auxdata <- function(auxdata, file="")
+{
+    check_auxdata_col2class(auxdata)
+    header <- paste0("#", paste(colnames(auxdata), collapse=", "))
+    cat(header, "\n", sep="", file=file)
+    write.table(auxdata, file, append=TRUE, quote=FALSE,
+                sep="\t", row.names=FALSE, col.names=FALSE)
+}
+

R/auxdata-utils.R

@@ -32,16 +32,6 @@ get_igblast_auxiliary_data <- function(...)
 ### load_auxdata()
 ###
 
-### Not the true colnames used in IgBLAST auxdata files. However, ours are
-### shorter, all lowercase, and contain underscores instead of spaces.
-.IGBLAST_AUXDATA_COL2CLASS <- c(
-    allele_name="character",
-    coding_frame_start="integer",
-    chain_type="character",
-    cdr3_end="integer",
-    extra_bps="integer"
-)
-
 ### IMPORTANT NOTE: Unlike with the data.frame returned by load_intdata(),
 ### all the positions in the data.frame returned by load_auxdata() (that is,
 ### the positions reported in columns 'coding_frame_start' and 'cdr3_end')
@@ -50,7 +40,7 @@ load_auxdata <- function(organism, which=c("live", "original"))
 {
     which <- match.arg(which)
     auxdata_path <- get_auxdata_path(organism, which=which)
-    read_broken_table(auxdata_path, .IGBLAST_AUXDATA_COL2CLASS)
+    read_auxdata(auxdata_path)
 }
 
 load_igblast_auxiliary_data <- function(...)

R/clean_allele_set.R

@@ -119,7 +119,7 @@ checkarg_with.intdata <- function(with.intdata, gapped)
     stopifnot(is(dna, "DNAStringSet"))
     ## We only care about the annotations found in the Metadata Columns
     ## Of Interest listed in 'MCOI' when comparing the V allele annotations.
-    MCOI <- setdiff(names(NDM_COL2CLASS), "chain_type")
+    MCOI <- setdiff(names(NDM_DATA_COL2CLASS), "chain_type")
     ok <- .repeated_vector_elts_have_identical_metadata(dna, MCOI=MCOI)
     if (!ok) {
         msg <- c("V alleles with identical ungapped sequences and names ",

R/compute_V_gene_delineations.R

@@ -84,11 +84,13 @@ EXTENDED_V_GENE_DELINEATION_COLNAMES <- c(
 ### Sanity checks.
 stopifnot(
     identical(
-        setdiff(names(NDM_COL2CLASS), EXTENDED_V_GENE_DELINEATION_COLNAMES),
+        setdiff(names(NDM_DATA_COL2CLASS),
+                EXTENDED_V_GENE_DELINEATION_COLNAMES),
         "chain_type"
     ),
     identical(
-        setdiff(EXTENDED_V_GENE_DELINEATION_COLNAMES, names(NDM_COL2CLASS)),
+        setdiff(EXTENDED_V_GENE_DELINEATION_COLNAMES,
+                names(NDM_DATA_COL2CLASS)),
         c("seq_len", "starting_gap", "all_gaps_in_frame", "all_gaps_contained")
     )
 )

R/create_region_db.R

@@ -149,7 +149,7 @@
     internal_data_path <- file.path(destdir, "internal_data")
     stopifnot(!dir.exists(internal_data_path))
 
-    ndm_data <- as.data.frame(dna_mcols[ , names(NDM_COL2CLASS)])
+    ndm_data <- as.data.frame(dna_mcols[ , names(NDM_DATA_COL2CLASS)])
     check_ndm_data_col2class(ndm_data)
     dir.create(internal_data_path)
     destfile <- file.path(internal_data_path, "V.ndm.imgt")

R/download_OGRDB_germline_json.R

@@ -156,7 +156,7 @@ makeogrannote <- function(ogrdb_json_file)
                      "in JSON file: ", ogrdb_json_file))
         data <- character(0)
     }
-    col2class <- head(NDM_COL2CLASS, n=-1L)
+    col2class <- head(NDM_DATA_COL2CLASS, n=-1L)
     m <- matrix(data, ncol=length(col2class), byrow=TRUE)
     df <- matrix2df(m, col2class)
     cbind(df, coding_frame_start=integer(nrow(df)))
@@ -190,7 +190,7 @@ makeogrannote <- function(ogrdb_json_file)
     intdata <- compute_V_gene_delineations(V_fasta_file)
     locus <- extract_loci_from_OGRDB_set_names(organism, names(germline_set))
     intdata$chain_type <- paste0("V", substr(locus, 3L, 3L))
-    intdata[ , names(NDM_COL2CLASS)]
+    intdata[ , names(NDM_DATA_COL2CLASS)]
 }
 
 ### Returns the intdata in a data.frame.

R/ndm_data-IO.R

@@ -5,7 +5,7 @@
 
 
 ### Not exported!
-### Not the true colnames used in IgBLAST intdata files: ours are all
+### Not the true colnames used in IgBLAST "ndm" files: ours are all
 ### lowercase and we've replaced spaces with underscores.
 ### Note that many columns are redundant:
 ### - columns 'cdr1_start', 'fwr2_start', 'cdr2_start', and 'fwr3_start'
@@ -14,7 +14,7 @@
 ### - columns 'fwr1_start' and 'coding_frame_start' are redundant (and
 ###   column 'fwr1_start' is a dumb column anyways because it should always
 ###   be set to 1).
-NDM_COL2CLASS <- c(
+NDM_DATA_COL2CLASS <- c(
     allele_name="character",
     fwr1_start="integer",
     fwr1_end="integer",
@@ -34,7 +34,7 @@ NDM_COL2CLASS <- c(
 V_GENE_SEGMENTS <- c("fwr1", "cdr1", "fwr2", "cdr2", "fwr3")
 V_GENE_DELINEATION_COLNAMES <- paste0(rep(V_GENE_SEGMENTS, each=2L),
                                       c("_start", "_end"))
-stopifnot(all(V_GENE_DELINEATION_COLNAMES %in% names(NDM_COL2CLASS)))
+stopifnot(all(V_GENE_DELINEATION_COLNAMES %in% names(NDM_DATA_COL2CLASS)))
 
 
 ### - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@@ -46,16 +46,16 @@ check_ndm_data_col2class <- function(ndm_data, what="'ndm_data'")
 {
     if (!is.data.frame(ndm_data))
         stop(wmsg(what, " must be a data.frame"))
-    expected_colnames <- names(NDM_COL2CLASS)
+    expected_colnames <- names(NDM_DATA_COL2CLASS)
     if (!identical(colnames(ndm_data), expected_colnames)) {
         in1string <- paste(expected_colnames, collapse=", ")
         stop(wmsg(what, " must have the following columns ",
                   "(in this order): ", in1string))
     }
     col2class <- vapply(ndm_data, function(x) class(x)[[1L]], character(1))
-    if (!identical(col2class, NDM_COL2CLASS)) {
-        in1string <- paste0("    ", names(NDM_COL2CLASS), " -> ",
-                            NDM_COL2CLASS, collapse="\n")
+    if (!identical(col2class, NDM_DATA_COL2CLASS)) {
+        in1string <- paste0("    ", names(NDM_DATA_COL2CLASS), " -> ",
+                            NDM_DATA_COL2CLASS, collapse="\n")
         stop(wmsg(what, " must have the following column types:"), "\n",
              in1string)
     }
@@ -107,7 +107,7 @@ validate_ndm_rows <- function(ndm_data, allow.repeated.rows=FALSE)
 
 read_ndm_data <- function(filepath)
 {
-    read_broken_table(filepath, NDM_COL2CLASS)
+    read_broken_table(filepath, NDM_DATA_COL2CLASS)
 }
 
 write_ndm_data <- function(ndm_data, file="", check.data=FALSE)