tests/testthat/test_create_pmap.R

In twang2218/pmv: Process Map Visualization

set.seed(101)

context("create_pmap()")

test_that("create_pmap() should handle simple graph", {
  eventlog <- data.frame(
     timestamp = c(as.POSIXct("2017-10-01"), as.POSIXct("2017-10-20")),
     case_id = c("c1", "c1"),
     activity = c("a", "b"),
     category = c("campaign", "sale"),
     stringsAsFactors = FALSE
  )

  p <- create_pmap(eventlog, target_categories = c("sale"))

  ndf <- DiagrammeR::get_node_df(p)
  expect_equal(nrow(ndf), 2)

  edf <- DiagrammeR::get_edge_df(p)
  expect_equal(nrow(edf), 1)
})

test_that("create_pmap() should handle complex graph", {
  eventlog <- generate_eventlog(
    size_of_eventlog = 10000,
    number_of_cases = 1000,
    categories = c("campaign", "sale"),
    categories_size = c(10, 4)
  )

  expect_named(
    eventlog,
    c("timestamp", "case_id", "activity", "category"),
    ignore.order = TRUE,
    ignore.case = TRUE)
  expect_equal(nrow(eventlog), 10000)

  # print(str(eventlog))
  # print("create_pmap_graph()")
  p <- create_pmap(eventlog, target_categories = c("sale"))

  # print(generate_dot(p))

  # print("render_graph()")
  expect_true(!any(is.null(render_pmap(p))))
  # print(render_pmap(p))
})

test_that("create_pmap() should handle more complex graph with multiple categories without target", {
  eventlog <- generate_eventlog(
    size_of_eventlog = 10000,
    number_of_cases = 1000,
    categories = c("campaign", "visit", "phone", "sale"),
    categories_size = c(5, 3, 2, 4)
  )

  expect_named(
    eventlog,
    c("timestamp", "case_id", "activity", "category"),
    ignore.order = TRUE,
    ignore.case = TRUE)
  expect_equal(nrow(eventlog), 10000)

  # print(str(eventlog))
  # print("create_pmap_graph()")
  p <- eventlog %>% create_pmap() %>% prune_nodes(0.3) %>% prune_edges(0.3)

  # print(DiagrammeR::generate_dot(p))

  # print("render_graph()")
  expect_true(!any(is.null(render_pmap(p))))
  print(render_pmap(p))
})

test_that("create_pmap() should handle names with SPACE padding", {
  p <- create_pmap(
    data.frame(
      timestamp = c(
        "2017-07-01",
        "2017-07-02",
        "2017-07-03",
        "2017-07-04",
        "2017-07-05",
        "2017-07-06",
        "2017-07-20"
      ),
      case_id = c("c1", "c1 ", "c1 ", "c2 ", "c2", "c3", "c3 "),
      activity = c("  a", "b  ", "a  ", "b", " b", "  a", "b "),
      category = c("  campaign", " sale", "campaign  ", " sale  ", " sale", " campaign", "sale"),
      stringsAsFactors = FALSE
    ),
    target_categories = c(" sale")
  )

  nodes <- DiagrammeR::get_node_df(p)
  expect_equal(nrow(nodes), 2)
  expect_equal(nodes$name, c("a", "b"))
  expect_equal(nodes$type, c("campaign", "sale"))
  expect_equal(nodes$amount, c(3, 4))

  edges <- DiagrammeR::get_edge_df(p)
  expect_equal(nrow(edges), 1)
  expect_equal(edges$from, 1)
  expect_equal(edges$to, 2)
  expect_equal(edges$amount, 2)

  # test duration
  expect_equal(edges$mean_duration, as.difftime(7.5, units = "days"))
  expect_equal(edges$median_duration, as.difftime(7.5, units = "days"))
  expect_equal(edges$max_duration, as.difftime(14, units = "days"))
  expect_equal(edges$min_duration, as.difftime(1, units = "days"))
  expect_equal(edges$tooltip, "from: a\nto: b\namount: 2\nmean_duration: 1.07 weeks\nmedian_duration: 1.07 weeks\nmax_duration: 2 weeks\nmin_duration: 1 days")
})

test_that("create_pmap() should distinct repeated activities if `distinct_repeated_activities`", {
  eventlog <- data.frame(
    timestamp = c(
      "2017-07-01",
      "2017-07-02",
      "2017-07-03",
      "2017-07-04",
      "2017-07-05",
      "2017-07-06",
      "2017-07-20"
    ),
    case_id = c("c1", "c1", "c1", "c2", "c2", "c3", "c3"),
    activity = c("a", "b", "a", "b", "b", "a", "b"),
    category = c("campaign", "sale", "campaign", "sale", "sale", "campaign", "sale"),
    stringsAsFactors = FALSE
  )

  p <- create_pmap(
    eventlog,
    distinct_repeated_activities = TRUE
  )

  nodes <- DiagrammeR::get_node_df(p)
  expect_equal(nrow(nodes), 4)
  expect_equal(nodes$name, c("a (1)", "a (2)", "b (1)", "b (2)"))
  expect_equal(nodes$type, c("campaign", "campaign", "sale", "sale"))
  expect_equal(nodes$amount, c(2, 1, 3, 1))
})

test_that("create_pmap() should distinct repeated activities if `distinct_repeated_activities` with missing `category` column in `eventlog`", {
  eventlog <- data.frame(
    timestamp = c(
      "2017-07-01",
      "2017-07-02",
      "2017-07-03",
      "2017-07-04",
      "2017-07-05",
      "2017-07-06",
      "2017-07-20"
    ),
    case_id = c("c1", "c1 ", "c1 ", "c2 ", "c2", "c3", "c3"),
    activity = c("a", "b", "a", "b", "b", "a", "b"),
    stringsAsFactors = FALSE
  )

  p <- create_pmap(
    eventlog,
    distinct_repeated_activities = TRUE
  )

  nodes <- DiagrammeR::get_node_df(p)
  expect_equal(nrow(nodes), 4)
  expect_equal(nodes$name, c("a (1)", "a (2)", "b (1)", "b (2)"))
  expect_equal(nodes$type, c("a", "a", "b", "b"))
  expect_equal(nodes$amount, c(2, 1, 3, 1))
})

test_that("create_pmap() should handle the eventlog without timestamp", {
  p <- create_pmap(
    data.frame(
      case_id = c("c1", "c1", "c1", "c2", "c2", "c3", "c3"),
      activity = c("a", "b", "a", "b", "b", "a", "b"),
      category = c("campaign", "sale", "campaign", "sale", "sale", "campaign", "sale"),
      stringsAsFactors = FALSE
    )
  )

  edges <- DiagrammeR::get_edge_df(p)

  expect_equal(nrow(edges), 3)
  expect_equal(edges$from, c(1, 2, 2))
  expect_equal(edges$to, c(2, 1, 2))
  expect_equal(edges$amount, c(2, 1, 1))

  # test duration
  expect_equal("mean_duration" %in% colnames(edges), FALSE)
  expect_equal("median_duration" %in% colnames(edges), FALSE)
  expect_equal("max_duration" %in% colnames(edges), FALSE)
  expect_equal("min_duration" %in% colnames(edges), FALSE)
})