test_that("gutenberg_download works with real API - single book", {
  skip_if_not_integration()
  skip_on_cran()
  skip_if_offline()

  # 1: Declaration of Independence
  result <- gutenberg_download(
    1,
    strip = TRUE,
    verbose = FALSE,
    use_cache = FALSE
  )

  expect_s3_class(result, "tbl_df")
  expect_named(result, c("gutenberg_id", "text"))
  expect_true(nrow(result) > 0)
  expect_equal(unique(result$gutenberg_id), 1)
  expect_type(result$text, "character")

  all_text <- paste(result$text, collapse = " ")
  expect_true(nchar(all_text) > 100)
})

test_that("gutenberg_download works with real API - multiple books", {
  skip_if_not_integration()
  skip_on_cran()
  skip_if_offline()

  # 1: Declaration of Independence
  # 2: US Bill of Rights
  result <- gutenberg_download(
    c(1, 2),
    strip = TRUE,
    verbose = FALSE,
    use_cache = FALSE
  )

  expect_s3_class(result, "tbl_df")
  expect_named(result, c("gutenberg_id", "text"))
  expect_true(nrow(result) > 0)
  expect_setequal(unique(result$gutenberg_id), c(1, 2))

  counts <- table(result$gutenberg_id)
  expect_true(all(counts > 0))
})

test_that("gutenberg_download with meta_fields works with real API", {
  skip_if_not_integration()
  skip_on_cran()
  skip_if_offline()

  result <- gutenberg_download(
    1,
    meta_fields = c("title", "author"),
    strip = TRUE,
    verbose = FALSE,
    use_cache = FALSE
  )

  expect_s3_class(result, "tbl_df")
  expect_true(all(
    c("gutenberg_id", "text", "title", "author") %in% names(result)
  ))
  expect_true(nrow(result) > 0)

  expect_equal(length(unique(result$title)), 1)
  expect_true(!is.na(unique(result$title)))
})


test_that("gutenberg_download strip parameter works with real API", {
  skip_if_not_integration()
  skip_on_cran()
  skip_if_offline()

  result_stripped <- gutenberg_download(
    1,
    strip = TRUE,
    verbose = FALSE,
    use_cache = FALSE
  )

  result_unstripped <- gutenberg_download(
    1,
    strip = FALSE,
    verbose = FALSE,
    use_cache = FALSE
  )

  expect_true(nrow(result_unstripped) > nrow(result_stripped))
  expect_true(nrow(result_stripped) > 0)
  expect_true(nrow(result_unstripped) > 0)
})

test_that("gutenberg_download caching works with real API", {
  skip_if_not_integration()
  skip_on_cran()
  skip_if_offline()

  with_gutenberg_cache({
    # First download - should hit the API
    network_result <- gutenberg_download(
      1,
      strip = TRUE,
      verbose = FALSE,
      use_cache = TRUE
    )

    cache_files <- gutenberg_cache_list(verbose = FALSE)
    expect_true(nrow(cache_files) > 0)
    expect_true(cache_files$file %in% c("1.rds"))

    # Second download - should use cache
    # We can't directly test if it used cache, but we can verify results match
    cache_result <- gutenberg_download(
      1,
      strip = TRUE,
      verbose = FALSE,
      use_cache = TRUE
    )

    expect_identical(network_result, cache_result)
  })
})

test_that("gutenberg_download works with data frame input", {
  skip_if_not_integration()
  skip_on_cran()
  skip_if_offline()

  books_df <- data.frame(gutenberg_id = c(1, 2))

  result <- gutenberg_download(
    books_df,
    strip = TRUE,
    verbose = FALSE,
    use_cache = FALSE
  )

  expect_s3_class(result, "tbl_df")
  expect_true(nrow(result) > 0)
  expect_setequal(unique(result$gutenberg_id), c(1, 2))
})