test_bdotsFit.R
In bdots: Bootstrapped Differences of Time Series

# library(bdots)
#
# groups <- c("Group", "LookType")
# res <- bdotsFit(data = cohort_unrelated,
#                 subject = "Subject",
#                 time = "Time",
#                 y = "Fixations",
#                 group = groups,
#                 curveType = doubleGauss(concave = TRUE),
#                 cor = TRUE,
#                 numRefits = 2,
#                 cores = 0,
#                 verbose = FALSE)
#
# ### Verify components of data.table
#
# ## Verify class correct, and dt info
# expect_equal(class(res), c("bdotsObj", "data.table", "data.frame"))
#
# ## Correct column
# # Subject, y, fit, ar1, r2, fitcode
# expect_equal(ncol(res), 5 + length(groups))
#
# ## Verify formula is fine
# ff <- attr(res, "formula")
#
# expect_equal(class(ff), "call")
# expect_equal(names(nn <- attributes(ff)), "parnames")
#
# # par names present in formula?
# ss <- Reduce(`+`, lapply(unlist(nn), grep, x = deparse1(ff)))
# expect_equal(ss, length(unlist(nn)))
#
# ## Grouping var match dataset?
# grp_vals <- do.call(function(...) paste(..., sep = "."), unique(cohort_unrelated[, ..groups]))
# grp_attr <- attr(res, "groups")
#
# expect_equal(grp_attr[['groups']], groups)
# expect_equal(grp_attr[['vals']], grp_vals)