Count distinct scientific names per group in summarize_observations()

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: camtraptor
 Title: Explore and Visualize Camera Trap Data Packages
-Version: 1.0.0.9000
+Version: 1.0.1
 Date: 2026-09-09
 Authors@R: c(
     person("Damiano", "Oldoni", , "damiano.oldoni@inbo.be", role = c("aut", "cre"),
@@ -19,6 +19,7 @@ Authors@R: c(
            comment = c(ORCID = "0000-0001-6788-5876", affiliation = "Research Institute for Nature and Forest (INBO)")),
     person("Nele", "Mullens", role = "ctb",
            comment = c(ORCID = "0000-0002-3054-9665", affiliation = "Regionaal Landschap Dijleland vzw")),
+    person("northfox", , , "vulpes.engineer@gmail.com", role = "ctb"),
     person("Raphaël", "Nussbaumer", , "rafnuss@gmail.com", role = "ctb",
            comment = c(ORCID = "0000-0002-8185-1020", affilication = "Swiss Ornithological Institute")),
     person("Thierry", "Onkelinx", , "thierry.onkelinx@inbo.be", role = "ctb",

NEWS.md

@@ -1,4 +1,12 @@
-# camtraptor (development version)
+# camtraptor 1.0.1
+
+## Bug fixes
+
+- Fixed an issue where `summarize_observations()` counted a scientific name once per deployment when `group_by` did not contain `deploymentID` (#432).
+
+## Other changes
+
+-  Add [northfox](https://github.com/northfox) as contributor. Welcome northfox!
 
 # camtraptor 1.0.0
 

R/summarize_observations.R

@@ -110,18 +110,36 @@ summarize_observations <- function(
   formula_n_species <- rlang::expr(
     "n_scientificName" := sum(.data$n_scientificName, na.rm = TRUE)
   )
-  # Calculate n_scientificName
+  # Calculate n_scientificName. The number of distinct scientific names is not
+  # additive across deployments: a scientific name observed in more than one
+  # deployment would be counted more than once. So, retain the scientific names
+  # by grouping by `scientificName` as well and count the distinct scientific
+  # names over all deployments afterwards.
   n_species_df <- calc_obs_feature(
     deployment_ids = deployment_ids,
     deployments = deployments,
     observations = observations,
     group_by_deployments = group_by_deployments,
-    group_by_observations = group_by_observations,
+    group_by_observations = union(group_by_observations, "scientificName"),
     group_time_by = group_time_by,
     col_obs_for_feature = "scientificName",
     formula_per_deployment = formula_n_species_per_dep,
     formula_total = formula_n_species
-  )
+  ) %>%
+    dplyr::group_by(
+      dplyr::across(dplyr::all_of(c(
+        group_by_deployments,
+        group_by_observations,
+        group_time_by
+      )))
+    ) %>%
+    dplyr::summarise(
+      "n_scientificName" := dplyr::n_distinct(
+        .data$scientificName,
+        na.rm = TRUE
+      ),
+      .groups = "keep"
+    )
   # Define the formula for the number of events per deployment
   formula_n_events_per_dep <- rlang::expr(
     "n_events" := dplyr::n_distinct(.data$eventID, na.rm = TRUE)

tests/testthat/test-summarize_observations.R

@@ -217,6 +217,58 @@ test_that("summarize_observations() returns correct summary for grouping by
                    rai_sum_count_df$rai_count)
 })
 
+test_that(
+  paste0("summarize_observations() returns n_scientificName equal to 1 or 0 ",
+         "when grouping by scientificName"
+  ), {
+    skip_if_offline()
+    x <- example_dataset()
+    summary <- summarize_observations(x, group_by = "scientificName")
+    # A scientific name observed in more than one deployment is counted once
+    expect_identical(
+      summary$n_scientificName,
+      dplyr::if_else(is.na(summary$scientificName), 0L, 1L)
+    )
+  })
+
+test_that(
+  paste0("summarize_observations() counts distinct scientific names over ",
+         "all deployments of a group"
+  ), {
+    skip_if_offline()
+    x <- example_dataset()
+    # Group all deployments in one location
+    deployments(x) <- deployments(x) %>%
+      dplyr::mutate(locationName = "location")
+    summary <- summarize_observations(x, group_by = "locationName")
+
+    # Scientific names observed in more than one deployment are counted once
+    n_species <- x %>%
+      filter_observations(.data$observationLevel == "event") %>%
+      observations() %>%
+      dplyr::pull("scientificName") %>%
+      dplyr::n_distinct(na.rm = TRUE)
+    expect_identical(summary$n_scientificName, n_species)
+
+    # Same with time grouping
+    summary_year <- summarize_observations(
+      x, group_by = "locationName", group_time_by = "year"
+    )
+    n_species_year <- x %>%
+      filter_observations(.data$observationLevel == "event") %>%
+      observations() %>%
+      dplyr::group_by(
+        year = lubridate::floor_date(.data$eventStart, "year")
+      ) %>%
+      dplyr::summarise(
+        n_scientificName = dplyr::n_distinct(.data$scientificName, na.rm = TRUE)
+      )
+    expect_identical(
+      summary_year$n_scientificName,
+      n_species_year$n_scientificName
+    )
+  })
+
 testthat::test_that(
   "Deployments without observations are not included in the summary", {
     skip_if_offline()