From 69869b02c3c0fc792c2626ef1ece61616117596b Mon Sep 17 00:00:00 2001
From: Ata B Barzegar <atabadrbarzegar@gmail.com>
Date: Mon, 24 Aug 2026 02:30:38 +0300
Subject: [PATCH 1/4] feat: extract schema from VoID dataset descriptions

- Add extract_schema_void() to query void:class, void:property,
and void:linkPredicate from a SPARQL endpoint or rdflib file and
populate the classes and properties fields of the schema.
---
 R/schema_void.R | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)
 create mode 100644 R/schema_void.R

diff --git a/R/schema_void.R b/R/schema_void.R
new file mode 100644
index 0000000..cf8ac6d
--- /dev/null
+++ b/R/schema_void.R
@@ -0,0 +1,35 @@
+#' @importFrom dplyr %>% filter transmute distinct
+NULL
+
+
+#' Extract a schema from a VoID dataset description,extracted information
+#'
+#' @param source A SPARQL endpoint URL or a parsed rdflib file.
+#' @return A schema list, see [.empty_schema()]; only `classes` and
+#'   `properties` are populated.
+#' @export
+extract_schema_void <- function(source) {
+  schema <- .empty_schema()
+
+  void_classes <- safe_query(source, '
+    PREFIX void: <http://rdfs.org/ns/void#>
+    SELECT DISTINCT ?class WHERE { [] void:class ?class . }', expected_cols = "class")
+
+  void_props <- safe_query(source, '
+    PREFIX void: <http://rdfs.org/ns/void#>
+    SELECT DISTINCT ?prop WHERE {
+      { [] void:property ?prop } UNION { [] void:linkPredicate ?prop }
+    }', expected_cols = "prop")
+
+  schema$classes <- void_classes %>%
+    filter(!is.na(class), nzchar(class)) %>%
+    transmute(uri = class, label = local_name(class)) %>%
+    distinct(uri, .keep_all = TRUE)
+
+  schema$properties <- void_props %>%
+    filter(!is.na(prop), nzchar(prop)) %>%
+    transmute(property = prop, label = local_name(prop)) %>%
+    distinct(property, .keep_all = TRUE)
+
+  schema
+}

From 3bf26a71f064d0e020e3c3f8fc29922adc1e0799 Mon Sep 17 00:00:00 2001
From: Ata B Barzegar <atabadrbarzegar@gmail.com>
Date: Mon, 24 Aug 2026 02:37:20 +0300
Subject: [PATCH 2/4] docs: add dplyr transmute to docs.

---
 NAMESPACE                  |  2 ++
 man/extract_schema_void.Rd | 18 ++++++++++++++++++
 2 files changed, 20 insertions(+)
 create mode 100644 man/extract_schema_void.Rd

diff --git a/NAMESPACE b/NAMESPACE
index 62ae34b..4548657 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -5,6 +5,7 @@ export(build_ont)
 export(extract_list_edges)
 export(extract_restrictions)
 export(extract_schema_owl)
+export(extract_schema_void)
 export(local_name)
 export(make_display_list)
 export(namespace_of)
@@ -19,3 +20,4 @@ importFrom(dplyr,distinct)
 importFrom(dplyr,filter)
 importFrom(dplyr,mutate)
 importFrom(dplyr,select)
+importFrom(dplyr,transmute)
diff --git a/man/extract_schema_void.Rd b/man/extract_schema_void.Rd
new file mode 100644
index 0000000..dd8685a
--- /dev/null
+++ b/man/extract_schema_void.Rd
@@ -0,0 +1,18 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/schema_void.R
+\name{extract_schema_void}
+\alias{extract_schema_void}
+\title{Extract a schema from a VoID dataset description,extracted information}
+\usage{
+extract_schema_void(source)
+}
+\arguments{
+\item{source}{A SPARQL endpoint URL or a parsed rdflib file.}
+}
+\value{
+A schema list, see \code{\link[=.empty_schema]{.empty_schema()}}; only \code{classes} and
+\code{properties} are populated.
+}
+\description{
+Extract a schema from a VoID dataset description,extracted information
+}

From bab452e571580e4292518dbabbe57a8e1c2ae160 Mon Sep 17 00:00:00 2001
From: Ata B Barzegar <atabadrbarzegar@gmail.com>
Date: Mon, 24 Aug 2026 02:37:52 +0300
Subject: [PATCH 3/4] tests: add tests for schema_void.R file

---
 tests/testthat/test-schema_void.R | 40 +++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)
 create mode 100644 tests/testthat/test-schema_void.R

diff --git a/tests/testthat/test-schema_void.R b/tests/testthat/test-schema_void.R
new file mode 100644
index 0000000..a3eeb8a
--- /dev/null
+++ b/tests/testthat/test-schema_void.R
@@ -0,0 +1,40 @@
+test_that("extract_schema_void reads void:class/property declarations", {
+  fake_sparql <- function(url, query, timeout = 60) {
+    if (grepl("void:class", query, fixed = TRUE)) {
+      data.frame(class = "http://ex.org/Dataset", stringsAsFactors = FALSE)
+    } else {
+      data.frame(prop = "http://purl.org/dc/terms/title", stringsAsFactors = FALSE)
+    }
+  }
+  testthat::local_mocked_bindings(sparql_query = fake_sparql)
+
+  schema <- extract_schema_void("https://example.org/sparql")
+
+  expect_equal(schema$classes$uri, "http://ex.org/Dataset")
+  expect_equal(schema$classes$label, "Dataset")
+  expect_equal(schema$properties$property, "http://purl.org/dc/terms/title")
+  expect_equal(schema$properties$label, "title")
+})
+
+test_that("extract_schema_void returns empty tables when there is no VoID data", {
+  testthat::local_mocked_bindings(
+    sparql_query = function(url, query, timeout = 60) data.frame()
+  )
+
+  schema <- extract_schema_void("https://example.org/sparql")
+
+  expect_equal(nrow(schema$classes), 0)
+  expect_equal(names(schema$classes), c("uri", "label"))
+  expect_equal(nrow(schema$properties), 0)
+})
+
+test_that("extract_schema_void tolerates a failing endpoint", {
+  testthat::local_mocked_bindings(
+    sparql_query = function(url, query, timeout = 60) stop("timeout")
+  )
+
+  schema <- suppressWarnings(extract_schema_void("https://example.org/sparql"))
+
+  expect_equal(nrow(schema$classes), 0)
+  expect_equal(nrow(schema$properties), 0)
+})

From 92cdec30d39e95b72cfe7966ce48388286d3d0b9 Mon Sep 17 00:00:00 2001
From: Ata B Barzegar <atabadrbarzegar@gmail.com>
Date: Mon, 24 Aug 2026 02:45:53 +0300
Subject: [PATCH 4/4] docs: change empty_schema function documentation into
 inline code

---
 R/schema_void.R            | 2 +-
 R/utils.R                  | 1 +
 man/extract_schema_void.Rd | 2 +-
 3 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/R/schema_void.R b/R/schema_void.R
index cf8ac6d..51216f8 100644
--- a/R/schema_void.R
+++ b/R/schema_void.R
@@ -5,7 +5,7 @@ NULL
 #' Extract a schema from a VoID dataset description,extracted information
 #'
 #' @param source A SPARQL endpoint URL or a parsed rdflib file.
-#' @return A schema list, see [.empty_schema()]; only `classes` and
+#' @return A schema list, see `.empty_schema()`; only `classes` and
 #'   `properties` are populated.
 #' @export
 extract_schema_void <- function(source) {
diff --git a/R/utils.R b/R/utils.R
index bd2abac..4ade5ce 100644
--- a/R/utils.R
+++ b/R/utils.R
@@ -44,6 +44,7 @@ local_name <- function(uri) {
 normalize_scheme <- function(x) sub("^https://", "http://", x)
 
 
+
 .empty_schema <- function() {
   list(
     classes = data.frame(uri = character(), label = character(), stringsAsFactors = FALSE),
diff --git a/man/extract_schema_void.Rd b/man/extract_schema_void.Rd
index dd8685a..673cccd 100644
--- a/man/extract_schema_void.Rd
+++ b/man/extract_schema_void.Rd
@@ -10,7 +10,7 @@ extract_schema_void(source)
 \item{source}{A SPARQL endpoint URL or a parsed rdflib file.}
 }
 \value{
-A schema list, see \code{\link[=.empty_schema]{.empty_schema()}}; only \code{classes} and
+A schema list, see \code{.empty_schema()}; only \code{classes} and
 \code{properties} are populated.
 }
 \description{