Add details opt

[tgerke]

Add an option for validation details which will print a gt table within the original gt table, showing the data which failed validations inline. This can be convenient when taking quick looks at data validations without downloading individual csv files and opening separately.

[shannonpileggi]

this is cute, thanks! will leave to @Meghansaha for further review.

[shannonpileggi]

this link doesn't work, the original one does.

[Meghansaha]

Hey @tgerke. This is pretty neat!

I suggested some changes to correct things, recommended refactoring that you can take or leave, and an extra test to get affirm_report.R back to 100%.

Let me know if you have any questions or want to chat about anything!

[Meghansaha]

@shannonpileggi, FYI, leaving this in as it's mac-specific. Any contributors on a mac might have this as well.

[Meghansaha]

Suggested change

	The [{pointblank}](https://rich-iannone.github.io/pointblank/) is
	The [{pointblank}](https://rstudio.github.io/pointblank/) is

Posit officially sponsors pointblank for R and Python now. It's on Rstudio and Posit GHs respectively now.

[Meghansaha]

I think this can be refactored.

You're recreating the base of the gt table twice in both workflows (inline_data == TRUE vs FALSE)

We can make a base gt table first, raw_data_prepared then modify it based on inline_data's value:

Suggested change

	affirm_report_gt <- function(inline_data = FALSE) {
	raw_data <- affirm_report_raw_data()
	if (inline_data) {
	# Create gt table with embedded expandable details using HTML
	# The details will be in a dedicated column that spans most of the visual space
	# First, create the CSV download links
	raw_data_with_csv <- raw_data |>
	dplyr::mutate(
	csv_download_link =
	mapply(
	FUN = .as_csv_encoded_html_download_link,
	# these two args are the ones being passed to FUN
	.data$data,
	paste0("extract_", dplyr::row_number(), ".csv"),
	# additional mapply args
	SIMPLIFY = TRUE,
	USE.NAMES = FALSE
	)
	)
	gt_table <- raw_data_with_csv |>
	dplyr::mutate(status_color = NA_character_, .before = 1L) |>
	dplyr::select(-"data") |>
	gt::gt() |>
	.affirm_report_gt_stylings() |>
	# Add a new column for expandable details
	gt::cols_add(
	details = NA_character_,
	.after = "status_color"
	) |>
	gt::text_transform(
	locations = gt::cells_body(columns = "details"),
	fn = function(x) {
	mapply(
	FUN = .create_gt_expandable_details_fullwidth,
	data = raw_data_with_csv$data,
	id = raw_data_with_csv$id,
	label = raw_data_with_csv$label,
	priority = raw_data_with_csv$priority,
	data_frames = raw_data_with_csv$data_frames,
	columns = raw_data_with_csv$columns,
	error_n = raw_data_with_csv$error_n,
	total_n = raw_data_with_csv$total_n,
	error_rate = raw_data_with_csv$error_rate,
	csv_download_link = raw_data_with_csv$csv_download_link,
	SIMPLIFY = FALSE,
	USE.NAMES = FALSE
	)
	}
	) |>
	gt::cols_label(details = "") |>
	# Make the details column span the full width
	gt::cols_width(
	status_color ~ gt::px(6),
	details ~ gt::pct(100)
	) |>
	# Hide the redundant columns since they're shown in the details
	gt::cols_hide(columns = c("id", "label", "priority", "data_frames", "columns", "error_n", "total_n", "error_rate", "csv_download_link"))
	return(gt_table)
	} else {
	# Original gt table behavior
	gt_table <- raw_data |>
	dplyr::mutate(status_color = NA_character_, .before = 1L) |>
	dplyr::mutate(
	csv_download_link =
	mapply(
	FUN = .as_csv_encoded_html_download_link,
	# these two args are the ones being passed to FUN
	.data$data,
	paste0("extract_", dplyr::row_number(), ".csv"),
	# additional mapply args
	SIMPLIFY = TRUE,
	USE.NAMES = FALSE
	)
	) |>
	dplyr::select(-"data") |>
	gt::gt() |>
	.affirm_report_gt_stylings()
	return(gt_table)
	}
	}
	affirm_report_gt <- function(inline_data = FALSE) {
	raw_data <- affirm_report_raw_data()
	# Add status_color and CSV links to all rows upfront
	raw_data_prepared <- raw_data |>
	dplyr::mutate(status_color = NA_character_, .before = 1L) |>
	dplyr::mutate(
	csv_download_link = mapply(
	FUN = .as_csv_encoded_html_download_link,
	.data$data,
	paste0("extract_", dplyr::row_number(), ".csv"),
	SIMPLIFY = TRUE,
	USE.NAMES = FALSE
	)
	)
	if (inline_data) {
	# Full-width expandable details mode: hides individual columns,
	# shows all details in a single expandable "Details" column
	gt_table <- raw_data_prepared |>
	dplyr::select(-"data") |>
	gt::gt() |>
	.affirm_report_gt_stylings() |>
	gt::cols_add(details = NA_character_, .after = "status_color") |>
	gt::text_transform(
	locations = gt::cells_body(columns = "details"),
	fn = function(x) {
	mapply(
	FUN = .create_gt_expandable_details_fullwidth,
	data = raw_data_prepared$data,
	id = raw_data_prepared$id,
	label = raw_data_prepared$label,
	priority = raw_data_prepared$priority,
	data_frames = raw_data_prepared$data_frames,
	columns = raw_data_prepared$columns,
	error_n = raw_data_prepared$error_n,
	total_n = raw_data_prepared$total_n,
	error_rate = raw_data_prepared$error_rate,
	csv_download_link = raw_data_prepared$csv_download_link,
	SIMPLIFY = FALSE,
	USE.NAMES = FALSE
	)
	}
	) |>
	gt::cols_label(details = "") |>
	gt::cols_width(
	status_color ~ gt::px(6),
	details ~ gt::pct(100)
	) |>
	gt::cols_hide(
	columns = c(
	"id",
	"label",
	"priority",
	"data_frames",
	"columns",
	"error_n",
	"total_n",
	"error_rate",
	"csv_download_link"
	)
	)
	} else {
	# Summary mode: displays all columns with CSV download links
	gt_table <- raw_data_prepared |>
	dplyr::select(-"data") |>
	gt::gt() |>
	.affirm_report_gt_stylings()
	}
	gt_table
	}

But your code honestly works as-is, so up to you.

[Meghansaha]

Note

I think this is fine as-is, but wondering if we'd ever need to refactor in the future for styling. Probably not, but if so, making a note for myself that this may be a spot to pay attention to.

[Meghansaha]

These test don't cover the gt text transformations fully:

Here's a minimal one that hits each line once for 100% coverage. There's probably more that could be tested, but wanted your thoughts on how you'd tackle.

Suggested change

	# Test inline_data parameter
	# Test inline_data parameter
	test_that("affirm_report_gt(inline_data = TRUE) executes text_transform with mapply", {
	affirm_init(replace = TRUE)
	affirm_true(
	mtcars,
	label = "mpg > 10",
	condition = mpg > 10,
	data_frames = "mtcars"
	)
	affirm_true(
	mtcars,
	label = "cyl in 4,6,8",
	condition = cyl %in% c(4, 6, 8),
	data_frames = "mtcars"
	)
	gt_result <- affirm_report_gt(inline_data = TRUE)
	expect_true("details" %in% colnames(gt_result[["_data"]]))
	# Render to HTML to trigger text_transform execution
	html_output <- gt::as_raw_html(gt_result)
	# Need at least one error to generate details tags
	# Verify the HTML contains text from the inline data
	expect_true(grepl("Affirmation:", html_output, fixed = TRUE))
	expect_true(grepl("No. Errors:", html_output, fixed = TRUE))
	# Check that mapply was executed by verifying styled content
	expect_true(grepl("status_color|details|padding|csv", html_output, fixed = FALSE))
	affirm_close()
	})
	# Test inline_data parameter