dsROCrate v0.2.2

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: dsROCrate
 Title: 'DataSHIELD' RO-Crate Governance Functions
-Version: 0.2.1
+Version: 0.2.2
 Authors@R: c(
     person(given = "Roberto",
            family = "Villegas-Diaz",

NEWS.md

@@ -1,3 +1,12 @@
+# dsROCrate 0.2.2
+
+## Bug fixes
+
+* Fixed an issue in `safe_output()` where Opal audit logs containing no `AGGREGATE` events could cause the audit pipeline to fail when constructing the calls table.
+* `safe_output()` now handles audit sessions with logs but no `AGGREGATE` calls gracefully, returning an empty calls table rather than attempting to operate on `NULL`.
+* Fixed an issue where an RO-Crate without an explicit Opal profile attribute could cause safe_output() to fail when filtering audit logs. The default "default" Opal profile is now used when no profile is supplied.
+* Improved robustness of provenance extraction for audit sessions with incomplete or limited audit activity.
+
 # dsROCrate 0.2.1
 
 ## Bug Fixes

R/audit.R

@@ -29,8 +29,15 @@
 #' @param ... Additional arguments.
 #' @param intent Additional object with governance bundle/specification of the
 #'     intent of a project. It takes the same types as `x`.
-#' @param project String with project name(s) from which to extra Safe Project
+#' @param profile String with profile name (used for OBiBa's Opal backend).
+#' @param project String with project name(s) from which to extract Safe Project
 #'     details.
+#' @param resources Vector of strings with the names of the resources, part of
+#'     `project`. Optional, if not provided, all the resources associated to
+#'     `project` will be included in the RO-Crate.
+#' @param tables Vector of strings with the names of the tables/datasets, part
+#'     of `project`. Optional, if not provided, all the tables/datasets
+#'     associated to `project` will be included in the RO-Crate.
 #' @param user String with the user name for which to extract Safe People
 #'     details.
 #' @param logs_from Lower limit timestamp to filter out the outputs generated
@@ -142,7 +149,10 @@ audit.opal <- function(
   x,
   ...,
   intent = NULL,
+  profile = "default",
   project = NULL,
+  resources = NULL,
+  tables = NULL,
   user = NULL,
   logs_from = -Inf,
   logs_to = Inf,
@@ -154,7 +164,10 @@ audit.opal <- function(
   # call next method
   main_audit <- audit_engine(
     x,
+    profile = profile,
     project = c(intent_lst$main_audit_args$project, project),
+    resources = resources,
+    tables = tables,
     user = c(intent_lst$main_audit_args$user, user),
     logs_from = logs_from,
     logs_to = logs_to,

R/audit_engine.R

@@ -7,8 +7,15 @@
 #'     governance archive file, representing the intent of a project and
 #'     associated governance details.
 #' @param ... Other optional arguments, see full documentation for details.
-#' @param project String with project name(s) from which to extra Safe Project
+#' @param profile String with profile name (used for OBiBa's Opal backend).
+#' @param project String with project name(s) from which to extract Safe Project
 #'     details.
+#' @param resources Vector of strings with the names of the resources, part of
+#'     `project`. Optional, if not provided, all the resources associated to
+#'     `project` will be included in the RO-Crate.
+#' @param tables Vector of strings with the names of the tables/datasets, part
+#'     of `project`. Optional, if not provided, all the tables/datasets
+#'     associated to `project` will be included in the RO-Crate.
 #' @param user String with the user name for which to extract Safe People
 #'     details.
 #' @param logs_from Lower limit timestamp to filter out the outputs generated
@@ -61,7 +68,10 @@ audit_engine.cr8tor <- function(x, ...) {
 audit_engine.opal <- function(
   x,
   ...,
+  profile = "default",
   project = NULL,
+  resources = NULL,
+  tables = NULL,
   user = NULL,
   logs_from = -Inf,
   logs_to = Inf,
@@ -70,11 +80,16 @@ audit_engine.opal <- function(
   # local bindings
   name <- NULL
 
-  # create RO-Create with the 5 safes profile
-  crate <- rocrateR::rocrate_5s()
-
-  # validate backend
-  validate_backend(x, ...)
+  # initialise empty RO-Crate with audit settings
+  crate <- x |>
+    dsROCrate::init(
+      profile = profile,
+      project = project,
+      resources = resources,
+      tables = tables,
+      user = user,
+      path = path
+    )
 
   # if `project` is missing, then error
   if (is.null(project)) {
@@ -165,7 +180,7 @@ audit_engine.opal <- function(
     purrr::reduce(rocrateR::remove_entity, .init = crate)
 
   # Safe Settings ----
-  crate <- safe_setting(x, rocrate = crate)
+  crate <- safe_setting(crate, connection = x)
 
   # Safe Outputs ----
   crate <- safe_people_tbl$name |>

R/backend-opal.R

@@ -9,8 +9,8 @@ backend_options.opal <- function(x, ...) {
 }
 
 #' @export
-backend_packages.opal <- function(x, ...) {
-  opalr::dsadmin.package_descriptions(x, ...)
+backend_packages.opal <- function(x, ..., profile = "default") {
+  opalr::dsadmin.package_descriptions(x, profile = profile, ...)
 }
 
 #' @export

R/safe_output.R

@@ -104,7 +104,11 @@ safe_output.opal <- function(
   `@timestamp` <- backend <- logger_name <- safe_people_id <- username <- NULL
   ds_action <- ds_eval <- ds_id <- ds_function <- ds_symbol <- ds_table <- NULL
   asset <- action <- is_placeholder <- kind <- symbol_id <- timestamp <- NULL
-  expr <- fx <- log_id <- r_cmd <- session <- symbol <- NULL
+  ds_profile <- expr <- fx <- log_id <- r_cmd <- session <- symbol <- NULL
+
+  if (is.null(profile)) {
+    profile <- "default"
+  }
 
   # create formatted versions of input dates
   logs_from_is_valid <- FALSE
@@ -217,12 +221,14 @@ safe_output.opal <- function(
   # parse logs
   userlogs_tbl <- backend_logs(x) |>
     tibble::as_tibble() |>
+    dplyr::bind_rows(tibble::tibble(ds_profile = "-999999")) |>
     dplyr::mutate(
       `@timestamp` = as.POSIXct(`@timestamp`, format = "%Y-%m-%dT%H:%M:%S")
     ) |>
     # filter logs
     dplyr::filter(`@timestamp` >= logs_from, `@timestamp` <= logs_to) |>
     dplyr::filter(logger_name == "datashield.user") |>
+    dplyr::filter(ds_profile == profile) |>
     dplyr::filter(username %in% user)
 
   userlogs <- NULL
@@ -350,27 +356,7 @@ safe_output.opal <- function(
     })
 
   # convert list of calls into tibble
-  calls_tbl <- purrr::map(calls_lst, \(x) {
-    tibble::tibble(
-      timestamp = format(x$created_at, '%Y-%m-%dT%H:%M:%S'),
-      action = "AGGREGATE",
-      user = x$user,
-      r_cmd = x$original,
-      fx = paste0(x$package, x$namespace, x$fx),
-      args = list(x$args),
-      symbol = NA,
-      table = x$args |>
-        purrr::map(function(x) {
-          if (!inherits(x, "safe_reference")) {
-            return(NA_character_)
-          }
-          resolve_symbol_asset(x$symbol_id, registry)
-        }),
-      session = x$session,
-      profile = x$profile
-    )
-  }) |>
-    purrr::list_c()
+  calls_tbl <- calls_to_tbl(calls_lst, registry)
 
   # combine function calls with symbol's registry
   calls_symbols_tbl <- calls_tbl |>

R/safe_setting.R

@@ -200,7 +200,7 @@ safe_setting.opal <- function(
 
   # computational environment ----
   # extract information about R packages installed in the environment
-  pkg_tbl <- backend_packages(x) |>
+  pkg_tbl <- backend_packages(x, profile = profile) |>
     tibble::as_tibble()
   pkg_entities <- pkg_tbl |>
     purrr::pmap(function(Package, Version, Description, Author, ...) {

R/utils-safe_output.R

@@ -1,3 +1,53 @@
+#' Convert log calls into tibble
+#'
+#' @param calls List with log calls.
+#' @param registry Symbol registry.
+#'
+#' @returns Tibble with parsed log calls.
+#' @keywords internal
+#'
+#' @noRd
+calls_to_tbl <- function(calls, registry) {
+  if (length(calls) == 0L) {
+    return(
+      tibble::tibble(
+        timestamp = character(),
+        action = character(),
+        user = character(),
+        r_cmd = character(),
+        fx = character(),
+        args = list(),
+        symbol = character(),
+        table = character(),
+        session = character(),
+        profile = character()
+      )
+    )
+  }
+
+  purrr::map(calls, \(x) {
+    tibble::tibble(
+      timestamp = format(x$created_at, "%Y-%m-%dT%H:%M:%S"),
+      action = "AGGREGATE",
+      user = x$user,
+      r_cmd = x$original,
+      fx = paste0(x$package, x$namespace, x$fx),
+      args = list(x$args),
+      symbol = NA_character_,
+      table = x$args |>
+        purrr::map(function(x) {
+          if (!inherits(x, "safe_reference")) {
+            return(NA_character_)
+          }
+          resolve_symbol_asset(x$symbol_id, registry)
+        }),
+      session = x$session,
+      profile = x$profile
+    )
+  }) |>
+    purrr::list_rbind()
+}
+
 #' Extract Safe Output entity(ies)
 #'
 #' @inheritParams safe_data