test_that("msigdbdf() human", {
  mdb <- msigdbdf(target_species = "HS")
  expect_s3_class(mdb, "tbl_df")
  expect_gt(nrow(mdb), 4400000)
  expect_lt(nrow(mdb), 4500000)
  expect_identical(names(mdb)[1:6], c("db_gene_symbol", "db_ncbi_gene", "db_ensembl_gene", "source_gene", "gs_id", "gs_name"))
  expect_equal(n_distinct(mdb$db_version), 1)
  expect_identical(unique(mdb$db_target_species), "HS")
  expect_equal(nrow(filter(mdb, db_gene_symbol == "")), 0)
  expect_equal(nrow(filter(mdb, db_ncbi_gene == "")), 0)
  expect_equal(nrow(filter(mdb, db_ensembl_gene == "")), 0)
  expect_equal(nrow(filter(mdb, gs_collection == "")), 0)
  # overall dimensions
  expect_equal(n_distinct(mdb$gs_collection), 9)
  expect_equal(n_distinct(mdb$gs_subcollection), 22)
  expect_gt(n_distinct(mdb$gs_id), 34000)
  expect_gt(n_distinct(mdb$db_gene_symbol), 40000)
  expect_gt(n_distinct(mdb$db_ncbi_gene), 40000)
  expect_gt(n_distinct(mdb$db_ensembl_gene), 40000)
  # gene set sizes
  expect_equal(min(table(mdb$gs_id)), 5)
  expect_lt(max(table(mdb$gs_id)), 2500)
  expect_lt(quantile(table(mdb$gs_id), 0.999), 2000)
  expect_lt(quantile(table(mdb$gs_id), 0.98), 1000)
  expect_lt(quantile(table(mdb$gs_id), 0.9), 300)
  expect_gt(quantile(table(mdb$gs_id), 0.9), 200)
  expect_gt(quantile(table(mdb$gs_id), 0.5), 40)
  expect_gt(quantile(table(mdb$gs_id), 0.2), 10)
  mdb_sym <- distinct(mdb, gs_id, db_gene_symbol)
  expect_equal(min(table(mdb_sym$gs_id)), 5)
  expect_lt(max(table(mdb_sym$gs_id)), 2001)
  mdb_ens <- distinct(mdb, gs_id, db_ensembl_gene)
  expect_equal(min(table(mdb_ens$gs_id)), 5)
  expect_lt(max(table(mdb_ens$gs_id)), 2001)
  # specific gene set sizes
  expect_equal(nrow(filter(mdb_sym, gs_id == "M39207")), 5)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M39207")), 5)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M32651")), 5)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M32651")), 5)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M22010")), 8)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M22010")), 8)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M5902")), 161)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M5902")), 161)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M5903")), 32)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M5903")), 32)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M40020")), 12)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M40020")), 12)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M40180")), 91)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M40180")), 91)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M6220")), 33)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M6220")), 33)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M30053")), 1957)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M30053")), 1957)
  expect_equal(nrow(filter(mdb_sym, gs_id == "M3458")), 1902)
  expect_equal(nrow(filter(mdb_ens, gs_id == "M3458")), 1920)
  # specific genes present in specific gene sets
  expect_equal(nrow(filter(mdb, gs_id == "M30055", db_gene_symbol == "FOS")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M30055", db_ncbi_gene == "2353")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M30055", db_ensembl_gene == "ENSG00000170345")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M40827", db_gene_symbol == "ABCA11P")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M40827", db_ncbi_gene == "79963")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M40827", db_ensembl_gene == "ENSG00000251595")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M8918", db_gene_symbol == "NEPNP")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M8918", db_ncbi_gene == "442253")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "M8918", db_ensembl_gene == "ENSG00000218233")), 1)
  # specific Ensembl mappings
  mdb_map <- distinct(mdb, db_gene_symbol, db_ensembl_gene)
  expect_lt(nrow(mdb_map), n_distinct(mdb_map$db_gene_symbol) * 1.05)
  expect_identical(filter(mdb, db_gene_symbol == "CDK1"), filter(mdb, db_ensembl_gene == "ENSG00000170312"))
  expect_identical(filter(mdb, db_gene_symbol == "TP53"), filter(mdb, db_ensembl_gene == "ENSG00000141510"))
  expect_identical(filter(mdb, db_gene_symbol == "SRSF1"), filter(mdb, db_ensembl_gene == "ENSG00000136450"))
})

test_that("msigdbdf() mouse", {
  mdb <- msigdbdf(target_species = "MM")
  expect_s3_class(mdb, "tbl_df")
  expect_gt(nrow(mdb), 1600000)
  expect_lt(nrow(mdb), 1700000)
  expect_equal(n_distinct(mdb$db_version), 1)
  expect_identical(unique(mdb$db_target_species), "MM")
  expect_equal(nrow(filter(mdb, db_gene_symbol == "")), 0)
  expect_equal(nrow(filter(mdb, db_ncbi_gene == "")), 0)
  expect_equal(nrow(filter(mdb, db_ensembl_gene == "")), 0)
  expect_equal(nrow(filter(mdb, gs_collection == "")), 0)
  # overall dimensions
  expect_equal(n_distinct(mdb$gs_collection), 6)
  expect_equal(n_distinct(mdb$gs_subcollection), 11)
  expect_gt(n_distinct(mdb$gs_id), 16000)
  expect_gt(n_distinct(mdb$db_gene_symbol), 40000)
  expect_gt(n_distinct(mdb$db_ncbi_gene), 40000)
  expect_gt(n_distinct(mdb$db_ensembl_gene), 40000)
  # gene set sizes
  expect_equal(min(table(mdb$gs_id)), 5)
  expect_lt(max(table(mdb$gs_id)), 2900)
  expect_lt(quantile(table(mdb$gs_id), 0.999), 2000)
  expect_lt(quantile(table(mdb$gs_id), 0.98), 1000)
  expect_lt(quantile(table(mdb$gs_id), 0.9), 300)
  expect_gt(quantile(table(mdb$gs_id), 0.9), 200)
  expect_gt(quantile(table(mdb$gs_id), 0.5), 20)
  expect_gt(quantile(table(mdb$gs_id), 0.1), 5)
  mdb_sym <- distinct(mdb, gs_id, db_gene_symbol)
  expect_equal(min(table(mdb_sym$gs_id)), 5)
  expect_lt(max(table(mdb_sym$gs_id)), 2001)
  mdb_ens <- distinct(mdb, gs_id, db_ensembl_gene)
  expect_equal(min(table(mdb_ens$gs_id)), 5)
  expect_lt(max(table(mdb_ens$gs_id)), 2001)
  # specific gene set sizes
  expect_equal(nrow(filter(mdb_sym, gs_id == "MM3871")), 200)
  expect_equal(nrow(filter(mdb_ens, gs_id == "MM3871")), 200)
  expect_equal(nrow(filter(mdb_sym, gs_id == "MM3634")), 294)
  expect_equal(nrow(filter(mdb_ens, gs_id == "MM3634")), 294)
  expect_equal(nrow(filter(mdb_sym, gs_id == "MM10270")), 12)
  expect_equal(nrow(filter(mdb_ens, gs_id == "MM10270")), 12)
  # specific genes present in specific gene sets
  expect_equal(nrow(filter(mdb, gs_id == "MM1067", db_gene_symbol == "Zfas1")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "MM1067", db_ncbi_gene == "68949")), 1)
  expect_equal(nrow(filter(mdb, gs_id == "MM1067", db_ensembl_gene == "ENSMUSG00000074578")), 1)
  # specific Ensembl mappings
  mdb_map <- distinct(mdb, db_gene_symbol, db_ensembl_gene)
  expect_lt(nrow(mdb_map), n_distinct(mdb_map$db_gene_symbol) * 1.05)
  expect_identical(filter(mdb, db_gene_symbol == "Cdk1"), filter(mdb, db_ensembl_gene == "ENSMUSG00000019942"))
  expect_identical(filter(mdb, db_gene_symbol == "Trp53"), filter(mdb, db_ensembl_gene == "ENSMUSG00000059552"))
  expect_identical(filter(mdb, db_gene_symbol == "Srsf1"), filter(mdb, db_ensembl_gene == "ENSMUSG00000018379"))
})

test_that("msigdbdf() target_species variants", {
  expect_identical(msigdbdf(target_species = "HS"), msigdbdf(target_species = "Hs"))
  expect_identical(msigdbdf(target_species = "MM"), msigdbdf(target_species = "Mm"))
})