Merge pull request #34 from valentinogagliardi/select-role

Implement select role lookup

Author: valentinogagliardi

docs/conf.py

@@ -10,7 +10,7 @@
 project = "unbrowsed"
 copyright = "2025, Valentino Gagliardi"
 author = "Valentino Gagliardi"
-release = "0.1.0a19"
+release = "0.1.0a20"
 
 # -- General configuration ---------------------------------------------------
 extensions = [

pyproject.toml

@@ -5,7 +5,7 @@ requires = [ "setuptools>=61" ]
 
 [project]
 name = "unbrowsed"
-version = "0.1.0a19"
+version = "0.1.0a20"
 description = "A browserless HTML testing library for Python"
 readme = "README.md"
 license = { file = "LICENSE" }

src/unbrowsed/queries.py

@@ -1,6 +1,6 @@
 """unbrowsed queries."""
 
-from typing import Any, Optional, Union
+from typing import Any, Optional
 
 from selectolax.lexbor import LexborHTMLParser as Parser
 from selectolax.lexbor import LexborNode
@@ -243,7 +243,7 @@ def get_by_text(dom: Parser, text: str, exact=True) -> QueryResult:
 def query_by_role(
     dom: Parser,
     role: AriaRoles,
-    current: Optional[Union[bool, str]] = None,
+    current: Optional[bool | str] = None,
     name: Optional[str] = None,
     description: Optional[str] = None,
 ) -> Optional[QueryResult]:
@@ -315,7 +315,7 @@ def query_by_role(
 def get_by_role(
     dom: Parser,
     role: AriaRoles,
-    current: Optional[Union[bool, str]] = None,
+    current: Optional[bool | str] = None,
     name: Optional[str] = None,
     description: Optional[str] = None,
 ) -> QueryResult:
@@ -367,7 +367,7 @@ def get_by_role(
 
 
 def query_all_by_role(
-    dom: Parser, role: AriaRoles, current: Optional[Union[bool, str]] = None
+    dom: Parser, role: AriaRoles, current: Optional[bool | str] = None
 ) -> list[QueryResult]:
     """
     Queries the DOM for all elements with the specified ARIA role.
@@ -402,7 +402,7 @@ def query_all_by_role(
 
 
 def get_all_by_role(
-    dom: Parser, role: AriaRoles, current: Optional[Union[bool, str]] = None
+    dom: Parser, role: AriaRoles, current: Optional[bool | str] = None
 ) -> list[QueryResult]:
     """
     Retrieves all elements from the DOM by their ARIA role.

src/unbrowsed/resolvers.py

@@ -113,9 +113,7 @@ def __init__(
     @staticmethod
     def get_implicit_role_mapping() -> ImplicitRoleMapping:
         return {
-            "a": lambda node: (
-                "link" if "href" in node.attributes else "generic"
-            ),
+            "a": RoleResolver.get_a_role,
             "article": "article",
             "aside": "complementary",
             "address": "group",
@@ -133,7 +131,7 @@ def get_implicit_role_mapping() -> ImplicitRoleMapping:
                 "password": "textbox",
             },
             "textarea": "textbox",
-            "select": "combobox",
+            "select": RoleResolver.get_select_role,
             "nav": "navigation",
             "main": "main",
             "meter": "meter",
@@ -195,14 +193,14 @@ def get_implicit_role(node: LexborNode):
         return handler if isinstance(handler, str) else None
 
     @staticmethod
-    def get_td_role(node: LexborNode) -> Optional[str]:
+    def get_td_role(node: LexborNode) -> str:
         """Determine the role of a td element."""
         ancestor = node.parent
         while ancestor and ancestor.tag != "table":
             ancestor = ancestor.parent
 
         if not ancestor:
-            return None
+            return ""
 
         table_role = ancestor.attributes.get("role")
 
@@ -215,10 +213,10 @@ def get_td_role(node: LexborNode) -> Optional[str]:
         elif table_role in ["grid", "treegrid"]:
             return "gridcell"
 
-        return None
+        return ""
 
     @staticmethod
-    def get_img_role(node: LexborNode) -> Optional[str]:
+    def get_img_role(node: LexborNode) -> str:
         """Determine the role of an img element."""
         if "alt" in node.attributes:
             alt = node.attributes.get("alt")
@@ -231,3 +229,23 @@ def get_img_role(node: LexborNode) -> Optional[str]:
         ):
             return "presentation"
         return "img"
+
+    @staticmethod
+    def get_select_role(node: LexborNode) -> str:
+        """Determine the role of a select element."""
+        if "multiple" in node.attributes:
+            return "listbox"
+        if (
+            "size" in node.attributes
+            and node.attributes.get("size", None) is not None
+        ):
+            if int(node.attributes.get("size")) > 1:  # type: ignore
+                return "listbox"
+        return "combobox"
+
+    @staticmethod
+    def get_a_role(node: LexborNode) -> str:
+        """Determine the role of an element"""
+        if "href" in node.attributes:
+            return "link"
+        return "generic"

src/unbrowsed/types.py

@@ -1,4 +1,6 @@
-from typing import Literal, TypedDict, Callable
+from typing import Literal, TypedDict
+from collections.abc import Callable
+from selectolax.lexbor import LexborNode
 
 Alert = Literal["alert"]
 Article = Literal["article"]
@@ -104,7 +106,7 @@ class InputType(TypedDict):
 
 
 class ImplicitRoleMapping(TypedDict, total=False):
-    a: Callable
+    a: Callable[[LexborNode], str]
     article: Article
     address: Group
     aside: Complementary
@@ -128,7 +130,7 @@ class ImplicitRoleMapping(TypedDict, total=False):
     h5: Heading
     h6: Heading
     hr: Separator
-    img: Callable
+    img: Callable[[LexborNode], str]
     input: InputType
     li: ListItem
     main: Main
@@ -141,11 +143,11 @@ class ImplicitRoleMapping(TypedDict, total=False):
     output: Status
     progress: ProgressBar
     section: Region
-    select: ComboBox
+    select: Callable[[LexborNode], str]
     summary: Button
     table: Table
     tbody: RowGroup
-    td: Callable
+    td: Callable[[LexborNode], str]
     textarea: TextBox
     tfoot: RowGroup
     th: ColumnHeader

tests/test_get_by_role.py

@@ -682,3 +682,54 @@ def test_get_by_role_img_no_accessible_name():
      """
     dom = parse_html(html)
     get_by_role(dom, "presentation")
+
+
+def test_gey_by_role_select():
+    html = """
+    <label for="pet-select">Choose a pet:</label>
+
+    <select name="pets" id="pet-select">
+      <option value="">--Please choose an option--</option>
+      <option value="dog">Dog</option>
+      <option value="cat">Cat</option>
+      <option value="hamster">Hamster</option>
+      <option value="parrot">Parrot</option>
+      <option value="spider">Spider</option>
+      <option value="goldfish">Goldfish</option>
+    </select>
+    """
+    dom = parse_html(html)
+    get_by_role(dom, "combobox")
+
+    html = """
+    <label for="shakes">Which shakes would you like to order?</label>
+    <select id="shakes" name="shakes" size>
+      <option>Vanilla Shake</option>
+      <option>Strawberry Shake</option>
+      <option>Chocolate Shake</option>
+    </select>
+    """
+    dom = parse_html(html)
+    get_by_role(dom, "combobox")
+
+    html = """
+    <label for="shakes">Which shakes would you like to order?</label>
+    <select id="shakes" name="shakes" size="2">
+      <option>Vanilla Shake</option>
+      <option>Strawberry Shake</option>
+      <option>Chocolate Shake</option>
+    </select>
+    """
+    dom = parse_html(html)
+    get_by_role(dom, "listbox")
+
+    html = """
+    <label for="shakes">Which shakes would you like to order?</label>
+    <select id="shakes" name="shakes" multiple>
+      <option>Vanilla Shake</option>
+      <option>Strawberry Shake</option>
+      <option>Chocolate Shake</option>
+    </select>
+    """
+    dom = parse_html(html)
+    get_by_role(dom, "listbox")