dplyr v. 1.0.0 compatibility

DESCRIPTION

@@ -18,7 +18,7 @@ License: MIT + file LICENSE
 Depends:
     R (>= 3.1.2)
 Imports:
-    dplyr (>= 0.7.0),
+    dplyr (>= 0.8.0),
     rlang,
     readr,
     stringr,

NAMESPACE

@@ -1,12 +1,5 @@
 # Generated by roxygen2: do not edit by hand
 
-S3method(as.tbl_genome,data.frame)
-S3method(as.tbl_genome,tbl_df)
-S3method(as.tbl_interval,GRanges)
-S3method(as.tbl_interval,data.frame)
-S3method(as.tbl_interval,tbl_df)
-export(as.tbl_genome)
-export(as.tbl_interval)
 export(bed12_to_exons)
 export(bed_absdist)
 export(bed_closest)
@@ -32,6 +25,8 @@ export(bed_sort)
 export(bed_subtract)
 export(bed_window)
 export(bound_intervals)
+export(check_genome)
+export(check_interval)
 export(concat)
 export(create_introns)
 export(create_tss)
@@ -40,20 +35,15 @@ export(create_utrs5)
 export(db_ensembl)
 export(db_ucsc)
 export(flip_strands)
+export(gr_to_bed)
 export(interval_spacing)
-export(is.tbl_genome)
-export(is.tbl_interval)
 export(read_bed)
 export(read_bed12)
 export(read_bedgraph)
 export(read_broadpeak)
 export(read_genome)
 export(read_narrowpeak)
 export(read_vcf)
-export(tbl_genome)
-export(tbl_interval)
-export(trbl_genome)
-export(trbl_interval)
 export(valr_example)
 export(values)
 export(values_unique)

NEWS.md

@@ -1,13 +1,24 @@
 # valr 0.5.0.9000
 
+## Major changes
+
+* `trbl_interval()` and `trbl_genome()` custom `tibble` subclasses have been deemed unnecessary and have been removed from the package. 
+
+* coercing `GRanges` to a `valr` compatible data.frame now uses the `gr_to_bed()` function rather than `as.trbl_interal()` methods. 
+
+
 ## Minor changes
 
+* dplyr version < 0.8.0 is no longer supported due to unnecessary code bloat and challenges with handling multiple grouping structures (#359).
+
 * The `sort_by` argument of `bed_random()` has been changed to `sorted`, and will now by default
 use `bed_sort()` to sort the output, rather than rely on naming the sorting columns. Sorting can
 be suppressed by using `sorted = FALSE`. 
 
 * `bed_sort()` now uses base R sorting with the `radix` method for increased speed. (#353)
 
+* `tbls` processed by `bed_merge()`or `bed_sort()` no longer store either `merged` or `sorted` as attributes, due to these attributes being rarely checked in the codebase and potential sources of unexpected behavior.
+
 ## Bug fixes
 
 * Fixed `bed_closest()` to prevent erroneous intervals being reported when adjacent closest intervals are present in the `y` table. (#348)

R/bed12_to_exons.r

@@ -4,7 +4,7 @@
 #' number, with respect to strand (i.e., the first exon for `-` strand
 #' genes will have larger start and end coordinates).
 #'
-#' @param x [tbl_interval()]
+#' @param x [ivl_df]
 #'
 #' @family utilities
 #'
@@ -15,7 +15,7 @@
 #'
 #' @export
 bed12_to_exons <- function(x) {
-  if (!is.tbl_interval(x)) x <- as.tbl_interval(x)
+  x <- check_interval(x)
 
   if (!ncol(x) == 12) {
     stop("expected 12 column input", call. = FALSE)

R/bed_absdist.r

@@ -15,12 +15,12 @@
 #'   Both absolute and scaled distances are reported as `.absdist` and
 #'   `.absdist_scaled`.
 #'
-#' @param x [tbl_interval()]
-#' @param y [tbl_interval()]
-#' @param genome [tbl_genome()]
+#' @param x [ivl_df]
+#' @param y [ivl_df]
+#' @param genome [genome_df]
 #'
 #' @return
-#' [tbl_interval()] with `.absdist` and `.absdist_scaled` columns.
+#' [ivl_df] with `.absdist` and `.absdist_scaled` columns.
 #'
 #' @template stats
 #'
@@ -39,9 +39,9 @@
 #'
 #' @export
 bed_absdist <- function(x, y, genome) {
-  if (!is.tbl_interval(x)) x <- as.tbl_interval(x)
-  if (!is.tbl_interval(y)) y <- as.tbl_interval(y)
-  if (!is.tbl_genome(genome)) genome <- as.tbl_genome(genome)
+  x <- check_interval(x)
+  y <- check_interval(y)
+  genome <- check_genome(genome)
 
   # establish grouping with shared groups (and chrom)
   groups_xy <- shared_groups(x, y)
@@ -55,20 +55,11 @@ bed_absdist <- function(x, y, genome) {
   x <- group_by(x, !!! groups_vars)
   y <- group_by(y, !!! groups_vars)
 
-  if (utils::packageVersion("dplyr") < "0.7.99.9000"){
-    x_cpp <- update_groups(x)
-    y_cpp <- update_groups(y)
-    grp_indexes <- shared_group_indexes(x_cpp, y_cpp)
-    res <- dist_impl(x_cpp, y_cpp,
-                     grp_indexes$x,
-                     grp_indexes$y,
-                     distcalc = "absdist")
-  } else {
-    grp_indexes <- shared_group_indexes(x, y)
-    res <- dist_impl(x, y,
-                     grp_indexes$x, grp_indexes$y,
-                     distcalc = "absdist")
-  }
+
+  grp_indexes <- shared_group_indexes(x, y)
+  res <- dist_impl(x, y,
+                   grp_indexes$x, grp_indexes$y,
+                   distcalc = "absdist")
 
   # convert groups_xy to character vector
   if (!is.null(groups_xy)) {
@@ -80,18 +71,18 @@ bed_absdist <- function(x, y, genome) {
   genome <- inner_join(genome, get_labels(y), by = c("chrom"))
 
   ref_points <- summarize(y, .ref_points = n())
-  genome <- inner_join(genome, ref_points, by = c("chrom", groups_xy))
+  genome <- inner_join(genome, ref_points, by = c(groups_xy))
 
   genome <- mutate(genome, .ref_gap = .ref_points / size)
   genome <- select(genome, -size, -.ref_points)
 
   # calculate scaled reference sizes
-  res <- full_join(res, genome, by = c("chrom", groups_xy))
+  res <- full_join(res, genome, by = c(groups_xy))
   res <- mutate(res, .absdist_scaled = .absdist * .ref_gap)
   res <- select(res, -.ref_gap)
 
   # report back original x intervals not found
-  x_missing <- anti_join(x, res, by = c("chrom", groups_xy))
+  x_missing <- anti_join(x, res, by = c(groups_xy))
   x_missing <- ungroup(x_missing)
   x_missing <- mutate(x_missing, .absdist = NA, .absdist_scaled = NA)
   res <- bind_rows(res, x_missing)

R/bed_closest.r

@@ -1,14 +1,14 @@
 #' Identify closest intervals.
 #'
-#' @param x [tbl_interval()]
-#' @param y [tbl_interval()]
+#' @param x [ivl_df]
+#' @param y [ivl_df]
 #' @param overlap report overlapping intervals
 #' @param suffix colname suffixes in output
 #'
 #' @template groups
 #'
 #' @return
-#' [tbl_interval()] with additional columns:
+#' [ivl_df] with additional columns:
 #'   - `.dist` distance to closest interval. Negative distances
 #'     denote upstream intervals.
 #'   - `.overlap` overlap with closest interval
@@ -18,26 +18,26 @@
 #' @seealso \url{http://bedtools.readthedocs.io/en/latest/content/tools/closest.html}
 #'
 #' @examples
-#' x <- trbl_interval(
+#' x <- tibble::tribble(
 #'   ~chrom, ~start, ~end,
 #'   'chr1', 100,    125
 #' )
 #'
-#' y <- trbl_interval(
+#' y <- tibble::tribble(
 #'   ~chrom, ~start, ~end,
 #'   'chr1', 25,     50,
 #'   'chr1', 140,    175
 #' )
 #'
 #' bed_glyph(bed_closest(x, y))
 #'
-#' x <- trbl_interval(
+#' x <- tibble::tribble(
 #'   ~chrom, ~start, ~end,
 #'   "chr1", 500,    600,
 #'   "chr2", 5000,   6000
 #' )
 #'
-#' y <- trbl_interval(
+#' y <- tibble::tribble(
 #'   ~chrom, ~start, ~end,
 #'   "chr1", 100,    200,
 #'   "chr1", 150,    200,
@@ -50,12 +50,12 @@
 #' bed_closest(x, y, overlap = FALSE)
 #'
 #' # Report distance based on strand
-#' x <- trbl_interval(
+#' x <- tibble::tribble(
 #'   ~chrom, ~start, ~end, ~name, ~score, ~strand,
 #'   "chr1", 10,	   20,   "a",   1,      "-"
 #' )
 #'
-#' y <- trbl_interval(
+#' y <- tibble::tribble(
 #'   ~chrom, ~start, ~end, ~name, ~score, ~strand,
 #'   "chr1", 8,	     9,	   "b",   1,      "+",
 #'   "chr1", 21,	   22,	 "b",   1,      "-"
@@ -74,8 +74,8 @@
 #' @export
 bed_closest <- function(x, y, overlap = TRUE,
                         suffix = c(".x", ".y")) {
-  if (!is.tbl_interval(x)) x <- as.tbl_interval(x)
-  if (!is.tbl_interval(y)) y <- as.tbl_interval(y)
+  x <- check_interval(x)
+  y <- check_interval(y)
 
   check_suffix(suffix)
 
@@ -96,10 +96,6 @@ bed_closest <- function(x, y, overlap = TRUE,
 
   suffix <- list(x = suffix[1], y = suffix[2])
 
-  if (utils::packageVersion("dplyr") < "0.7.99.9000"){
-    x <- update_groups(x)
-    y <- update_groups(y)
-  }
   grp_indexes <- shared_group_indexes(x, y)
 
   res <- closest_impl(x, y,

R/bed_cluster.r

@@ -4,20 +4,20 @@
 #' `max_dist = 0` means that both overlapping and book-ended intervals will be
 #' clustered.
 #'
-#' @param x [tbl_interval()]
+#' @param x [ivl_df]
 #' @param max_dist maximum distance between clustered intervals.
 #'
 #' @template groups
 #'
-#' @return [tbl_interval()] with `.id` column specifying sets of clustered intervals.
+#' @return [ivl_df] with `.id` column specifying sets of clustered intervals.
 #'
 #' @family single set operations
 #'
 #' @seealso
 #' \url{http://bedtools.readthedocs.org/en/latest/content/tools/cluster.html}
 #'
 #' @examples
-#' x <- trbl_interval(
+#' x <- tibble::tribble(
 #'   ~chrom, ~start, ~end,
 #'   'chr1', 100,    200,
 #'   'chr1', 180,    250,
@@ -30,7 +30,7 @@
 #' bed_cluster(x)
 #'
 #' # glyph illustrating clustering of overlapping and book-ended intervals
-#' x <- trbl_interval(
+#' x <- tibble::tribble(
 #'   ~chrom, ~start, ~end,
 #'   'chr1', 1,      10,
 #'   'chr1', 5,      20,
@@ -43,14 +43,12 @@
 #'
 #' @export
 bed_cluster <- function(x, max_dist = 0) {
-  if (!is.tbl_interval(x)) x <- as.tbl_interval(x)
+  x <- check_interval(x)
 
-  res <- group_by(x, chrom, add = TRUE)
-  res <- bed_sort(res)
+  groups <- rlang::syms(unique(c("chrom", group_vars(x))))
+  res <- group_by(x, !!! groups)
 
-  if (utils::packageVersion("dplyr") < "0.7.99.9000"){
-    res <- update_groups(res)
-  }
+  res <- bed_sort(res)
 
   res <- merge_impl(res, max_dist, collapse = FALSE)
 

R/bed_complement.r

@@ -1,34 +1,34 @@
 #' Identify intervals in a genome not covered by a query.
 #'
-#' @param x [tbl_interval()]
-#' @param genome [tbl_genome()]
+#' @param x [ivl_df]
+#' @param genome [ivl_df]
 #'
 #' @family single set operations
 #'
-#' @return [tbl_interval()]
+#' @return [ivl_df]
 #'
 #' @examples
-#' x <- trbl_interval(
+#' x <- tibble::tribble(
 #'   ~chrom, ~start, ~end,
 #'   'chr1', 0,      10,
 #'   'chr1', 75,     100
 #' )
 #'
-#' genome <- trbl_genome(
+#' genome <- tibble::tribble(
 #'   ~chrom, ~size,
 #'   'chr1', 200
 #' )
 #'
 #' bed_glyph(bed_complement(x, genome))
 #'
-#' genome <- trbl_genome(
+#' genome <- tibble::tribble(
 #'    ~chrom,  ~size,
 #'    'chr1',  500,
 #'    'chr2',  600,
 #'    'chr3',  800
 #' )
 #'
-#' x <- trbl_interval(
+#' x <- tibble::tribble(
 #'    ~chrom, ~start, ~end,
 #'    'chr1', 100,    300,
 #'    'chr1', 200,    400,
@@ -42,8 +42,8 @@
 #'
 #' @export
 bed_complement <- function(x, genome) {
-  if (!is.tbl_interval(x)) x <- as.tbl_interval(x)
-  if (!is.tbl_genome(genome)) genome <- as.tbl_genome(genome)
+  x <- check_interval(x)
+  genome <- check_genome(genome)
 
   res <- bed_merge(x)
 
@@ -57,13 +57,9 @@ bed_complement <- function(x, genome) {
 
   res <- group_by(res, chrom)
 
-  if (utils::packageVersion("dplyr") < "0.7.99.9000"){
-    res <- update_groups(res)
-  }
-
   res <- complement_impl(res, genome)
 
-  res <- bind_rows(res, chroms_no_overlaps)
+  res <- bind_rows(res, as_tibble(chroms_no_overlaps))
   res <- bed_sort(res)
 
   res