simple compactify test, rename function, fix warn

R/archive.R

@@ -318,11 +318,13 @@ new_epi_archive <- function(
   nrow_before_compactify <- nrow(data_table)
   # Runs compactify on data frame
   if (is.null(compactify) || compactify == TRUE) {
-    data_table <- compactify(data_table, key_vars, compactify_tol)
+    compactified <- apply_compactify(data_table, key_vars, compactify_tol)
+  } else{
+    compactified <- data_table
   }
   # Warns about redundant rows if the number of rows decreased, and we didn't
   # explicitly say to compactify
-  if (is.null(compactify) && nrow(data_table) < nrow_before_compactify) {
+  if (is.null(compactify) && nrow(compactified) < nrow_before_compactify) {
     elim <- removed_by_compactify(data_table, key_vars, compactify_tol)
     warning_intro <- cli::format_inline(
       "Found rows that appear redundant based on
@@ -345,7 +347,7 @@ new_epi_archive <- function(
 
   structure(
     list(
-      DT = data_table,
+      DT = compactified,
       geo_type = geo_type,
       time_type = time_type,
       additional_metadata = additional_metadata,
@@ -364,7 +366,7 @@ new_epi_archive <- function(
 #'   changed, and so is kept.
 #' @keywords internal
 #' @importFrom dplyr filter
-compactify <- function(df, keys, tolerance = .Machine$double.eps^.5) {
+apply_compactify <- function(df, keys, tolerance = .Machine$double.eps^.5) {
   df %>%
     arrange(!!!keys) %>%
     filter(if_any(
@@ -379,7 +381,7 @@ compactify <- function(df, keys, tolerance = .Machine$double.eps^.5) {
 removed_by_compactify <- function(df, keys, tolerance) {
   df %>%
     arrange(!!!keys) %>%
-    filter(if_any(
+    filter(if_all(
       c(everything(), -version),
       ~ is_locf(., tolerance)
     )) # nolint: object_usage_linter

R/revision_analysis.R

@@ -103,7 +103,7 @@ revision_summary <- function(epi_arch,
   if (should_compactify) {
     revision_behavior <- revision_behavior %>%
       arrange(across(c(geo_value, time_value, all_of(keys), version))) %>% # need to sort before compactifying
-      compactify(c(keys, version), compactify_tol)
+      apply_compactify(c(keys, version), compactify_tol)
   }
   revision_behavior <-
     revision_behavior %>%

tests/testthat/_snaps/revision-latency-functions.md

@@ -58,16 +58,16 @@
       * 2 out of 19 (10.53%)
       Fraction of epi_key+time_values with
       No revisions:
-      * 1 out of 7 (14.29%)
+      * 2 out of 7 (28.57%)
       Quick revisions (last revision within 3 days of the `time_value`):
-      * 3 out of 7 (42.86%)
+      * 4 out of 7 (57.14%)
       Few revisions (At most 3 revisions for that `time_value`):
       * 6 out of 7 (85.71%)
       Fraction of revised epi_key+time_values which have:
       Less than 0.1 spread in relative value:
-      * 3 out of 6 (50%)
+      * 2 out of 5 (40%)
       Spread of more than 5.1 in actual value (when revised):
-      * 3 out of 6 (50%)
+      * 3 out of 5 (60%)
       days until within 20% of the latest value:
     Output
            min median     mean     max
@@ -76,10 +76,10 @@
         time_value geo_value n_revisions min_lag max_lag spread rel_spread
         <date>     <chr>           <dbl> <drtn>  <drtn>   <dbl>      <dbl>
       1 2020-01-01 ak                  6 2 days  19 days    101      0.990
-      2 2020-01-01 al                  2 0 days  19 days     99      0.99 
+      2 2020-01-01 al                  1 0 days  19 days     99      0.99 
       3 2020-01-02 ak                  1 4 days   5 days      9      0.09 
       4 2020-01-02 al                  0 0 days   0 days      0      0    
-      5 2020-01-03 ak                  1 3 days   4 days      0    NaN    
+      5 2020-01-03 ak                  0 3 days   3 days      0    NaN    
       6 2020-01-03 al                  1 1 days   2 days      3      0.75 
       7 2020-01-04 al                  1 0 days   1 days      0      0    
         time_near_latest

tests/testthat/test-archive.R

@@ -48,6 +48,16 @@ test_that("as_epi_archive custom name mapping works correctly", {
   )
 })
 
+dumb_ex <- data.frame(
+  geo_value = c("ca", "ca"),
+  time_value = as.Date(c("2020-01-01", "2020-01-01")),
+  value = c(1,1),
+  version = as.Date(c("2020-01-01", "2020-01-02"))
+)
+test_that("new_epi_archive correctly detects and warns about compactification", {
+  expect_snapshot(res <- dumb_ex %>% as_epi_archive())
+})
+
 test_that("other_keys can only contain names of the data.frame columns", {
   expect_error(as_epi_archive(archive_data, other_keys = "xyz", compactify = FALSE),
     regexp = "`other_keys` must be contained in the column names of `x`."

tests/testthat/test-revision-latency-functions.R

@@ -31,9 +31,6 @@ dummy_ex <- tibble::tribble(
   as_epi_archive(compactify = FALSE)
 
 test_that("revision_summary works for a dummy dataset", {
-  dummy_ex %>%
-    revision_summary() %>%
-    print(n = 10, width = 300)
   expect_snapshot(dummy_ex %>% revision_summary() %>% print(n = 10, width = 300))
   expect_snapshot(dummy_ex %>% revision_summary(drop_nas = FALSE) %>% print(n = 10, width = 300))
 })