From 2c8a3f3a83cf2f067048835114bdba81b7f092e5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Adri=C3=A1n=20Chaves?= <adrian@chaves.io>
Date: Thu, 25 Jul 2024 10:03:36 +0200
Subject: [PATCH 1/2] Add a minimal action I/O implementation to improve
 SearchAction support

---
 extruct/w3cmicrodata.py                        | 12 ++++++++++++
 tests/samples/schema.org/SearchAction.001.html |  8 ++++++++
 tests/samples/schema.org/SearchAction.001.json |  1 +
 tests/test_microdata.py                        | 15 +++++++++++++++
 4 files changed, 36 insertions(+)
 create mode 100644 tests/samples/schema.org/SearchAction.001.html
 create mode 100644 tests/samples/schema.org/SearchAction.001.json

diff --git a/extruct/w3cmicrodata.py b/extruct/w3cmicrodata.py
index 89a79b3c..c2696eae 100644
--- a/extruct/w3cmicrodata.py
+++ b/extruct/w3cmicrodata.py
@@ -235,6 +235,18 @@ def _extract_property_value(self, node, items_seen, base_url, itemids, force=Fal
         elif node.get("content"):
             return node.get("content")
 
+        # https://schema.org/docs/actions.html#part-4
+        elif (
+            (itemprop := node.get("itemprop"))
+            and (itemprop.endswith("-input") or itemprop.endswith("-output"))
+        ):
+            result = {}
+            if "required" in node.attrib:
+                result["valueRequired"] = True
+            if name := node.get("name"):
+                result["valueName"] = name
+            return result
+
         else:
             return self._extract_textContent(node)
 
diff --git a/tests/samples/schema.org/SearchAction.001.html b/tests/samples/schema.org/SearchAction.001.html
new file mode 100644
index 00000000..ef493125
--- /dev/null
+++ b/tests/samples/schema.org/SearchAction.001.html
@@ -0,0 +1,8 @@
+<div itemscope itemtype="https://schema.org/WebSite">
+  <meta itemprop="url" content="https://www.example.com/"/>
+  <form itemprop="potentialAction" itemscope itemtype="https://schema.org/SearchAction">
+    <meta itemprop="target" content="https://query.example.com/search?q={search_term_string}"/>
+    <input itemprop="query-input" type="text" name="search_term_string" required/>
+    <input type="submit"/>
+  </form>
+</div>
diff --git a/tests/samples/schema.org/SearchAction.001.json b/tests/samples/schema.org/SearchAction.001.json
new file mode 100644
index 00000000..bd967365
--- /dev/null
+++ b/tests/samples/schema.org/SearchAction.001.json
@@ -0,0 +1 @@
+[{"type": "https://schema.org/WebSite", "properties": {"url": "https://www.example.com/", "potentialAction": {"type": "https://schema.org/SearchAction", "properties": {"target": "https://query.example.com/search?q={search_term_string}", "query-input": {"valueRequired": true, "valueName": "search_term_string"}}}}}]
\ No newline at end of file
diff --git a/tests/test_microdata.py b/tests/test_microdata.py
index c1168a02..69e305cb 100644
--- a/tests/test_microdata.py
+++ b/tests/test_microdata.py
@@ -10,6 +10,18 @@ class TestMicrodata(unittest.TestCase):
 
     maxDiff = None
 
+    def _test_schemaorg(self, schema, indexes=None):
+        indexes = indexes or [1]
+        for i in indexes:
+            body = get_testdata("schema.org", f"{schema}.{i:03d}.html")
+            expected = json.loads(
+                get_testdata("schema.org", f"{schema}.{i:03d}.json").decode()
+            )
+            mde = MicrodataExtractor()
+            data = mde.extract(body)
+            self.assertEqual(data, expected)
+
+
     def test_schemaorg_CreativeWork(self):
         for i in [1]:
             body = get_testdata("schema.org", "CreativeWork.{:03d}.html".format(i))
@@ -63,6 +75,9 @@ def test_schemaorg_Event(self):
 
             self.assertEqual(data, expected)
 
+    def test_schemaorg_SearchAction(self):
+        self._test_schemaorg("SearchAction")
+
     def test_w3c_textContent_values(self):
         body = get_testdata("w3c", "microdata.4.2.strings.html")
         expected = json.loads(

From 1000c838a921372d77733bbddd6189f0296eee14 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Adri=C3=A1n=20Chaves?= <adrian@chaves.io>
Date: Thu, 25 Jul 2024 10:22:32 +0200
Subject: [PATCH 2/2] Run pre-commit

---
 extruct/w3cmicrodata.py | 5 ++---
 tests/test_microdata.py | 1 -
 2 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/extruct/w3cmicrodata.py b/extruct/w3cmicrodata.py
index c2696eae..0a7dbe6a 100644
--- a/extruct/w3cmicrodata.py
+++ b/extruct/w3cmicrodata.py
@@ -236,9 +236,8 @@ def _extract_property_value(self, node, items_seen, base_url, itemids, force=Fal
             return node.get("content")
 
         # https://schema.org/docs/actions.html#part-4
-        elif (
-            (itemprop := node.get("itemprop"))
-            and (itemprop.endswith("-input") or itemprop.endswith("-output"))
+        elif (itemprop := node.get("itemprop")) and (
+            itemprop.endswith("-input") or itemprop.endswith("-output")
         ):
             result = {}
             if "required" in node.attrib:
diff --git a/tests/test_microdata.py b/tests/test_microdata.py
index 69e305cb..eda2d26c 100644
--- a/tests/test_microdata.py
+++ b/tests/test_microdata.py
@@ -21,7 +21,6 @@ def _test_schemaorg(self, schema, indexes=None):
             data = mde.extract(body)
             self.assertEqual(data, expected)
 
-
     def test_schemaorg_CreativeWork(self):
         for i in [1]:
             body = get_testdata("schema.org", "CreativeWork.{:03d}.html".format(i))