tests/testthat/test-S4Vectors-behaviors.R

In TnT: Interactive Visualization for Genomic Features

# This test aims to monitor some special behaviors involving conversion of
# nested data frame in S4Vectors.
#
# It seems that currently S4Vectors does not really consider nested data.frame nor DataFrame
# a valid usage.

context("S4Vectors-behaviors")

suppressPackageStartupMessages({
    library(testthat)
    library(S4Vectors)
})

# Behaviors of S4Vectors have been changed, nested DataFrame will be flatten in as.data.frame
# 
# test_that("as.data.frame(DF) with nested df will have the nested df wrapped in AsIs class", {
#     DF <- DataFrame(x = c(1:5))
#     DF$nested <- data.frame(
#         nested_a = runif(5),
#         nested_b = paste0("aa", 1:5)
#     )
#     converted.df <- as.data.frame(DF)
#     expect_identical(unclass(converted.df$nested), unclass(DF$nested))
#     
#     # Error is associated with printing method of the converted data frame
#     expect_error(print(converted.df), "dims .* do not match the length of object")
#     # Reason is that the column of nested data frame has an AsIs class
#     expect_identical(class(converted.df$x), "integer")
#     expect_identical(class(converted.df$nested), c("AsIs", "data.frame"))
#     # Unwrap the AsIs class
#     cl <- class(converted.df$nested)
#     class(converted.df$nested) <- cl[cl != "AsIs"]
#     # No error now
#     capture_output(print(converted.df))
# })
# 
# test_that("DF with nested DF can not be converted to data.frame", {
#     DF <- DataFrame(x = c(1:5))
#     DF$nested <- DataFrame(
#         nested_a = runif(5),
#         nested_b = paste0("aa", 1:5)
#     )
#     
#     expect_error(as.data.frame(DF))
# })

# Has been fixed in S4Vectors version 0.15.10
#
# test_that("Error printing GRanges with a meta-column of nested df which has only one column", {
#     suppressPackageStartupMessages(library(GenomicRanges))
#     
#     gr <- GRanges("chr21", IRanges(1:5, width = 1))
#     
#     gr$df <- data.frame(x = 1:5)
#     expect_error(capture.output(show(gr)),
#                  "number of rows of matrices must match")
#     
#     gr$col <- 1:5
#     expect_warning(capture.output(show(gr)),
#                    "row names were found from a short variable and have been discarded")
#     
#     gr$df <- data.frame(x = 1:5, y = 1:5)
#     # Now the printing is correct
#     capture.output(show(gr))
# })