Fix UI-TARS scroll parsing (scroll actions were silently dropped)

src/open_apps/agent/utils.py

@@ -303,22 +303,26 @@ def uitars_parser(result):
     # type(content=text) -> keyboard_type(text=text)
     if result["action"].startswith("type(content="):
         result["action"] = translate_uitars_type_action(result["action"])
-    # scroll(direction='down', point='(906,509)') -> scroll(dx, dy)
-    if result["action"].startswith("scroll(direction=d"):
-        direction = re.findall(
-            r"scroll\(direction='(.*?)', point='\((\d+),(\d+)\)'\)", result["action"]
+    # scroll(direction='down'|'up'|'left'|'right', point='(x, y)') -> scroll(dx, dy)
+    if result["action"].startswith("scroll("):
+        dir_match = re.search(
+            r"direction\s*=\s*['\"]?(down|up|left|right)",
+            result["action"],
+            re.IGNORECASE,
         )
-        if direction:
-            result["action"] = f"scroll({int(direction[0][1])}, {int(direction[0][2])})"
-    # scroll(direction='up', point='(906,509)') -> scroll(dx, dy)
-    if result["action"].startswith("scroll(direction=u"):
-        direction = re.findall(
-            r"scroll\(direction='(.*?)', point='\((\d+),(\d+)\)'\)", result["action"]
-        )
-        if direction:
-            result["action"] = (
-                f"scroll({-int(direction[0][1])}, {-int(direction[0][2])})"
-            )
+        nums = re.findall(r"-?\d+", result["action"])
+        if dir_match and len(nums) >= 2:
+            direction = dir_match.group(1).lower()
+            x, y = int(nums[0]), int(nums[1])
+            if direction == "down":
+                dx, dy = 0, y
+            elif direction == "up":
+                dx, dy = 0, -y
+            elif direction == "right":
+                dx, dy = x, 0
+            else:
+                dx, dy = -x, 0
+            result["action"] = f"scroll({dx}, {dy})"
     # right_single(point='(531,256)') -> mouse_click(x, y, button='right')
     if result["action"].startswith("right_single(point="):
         coords = re.findall(r"\d+", result["action"])

tests/test_uitars_parser.py

@@ -0,0 +1,30 @@
+from open_apps.agent.utils import flexible_parser
+
+
+def _action(native: str) -> str:
+    return flexible_parser(f"<think>t</think><action>{native}</action>")["action"]
+
+
+def test_scroll_down_translates_to_positive_dy():
+    assert _action("scroll(direction='down', point='(612,455)')") == "scroll(0, 455)"
+
+
+def test_scroll_up_translates_to_negative_dy():
+    assert _action("scroll(direction='up', point='(1920,536)')") == "scroll(0, -536)"
+
+
+def test_scroll_left_and_right_move_along_x_axis():
+    assert _action("scroll(direction='right', point='(300,400)')") == "scroll(300, 0)"
+    assert _action("scroll(direction='left', point='(300,400)')") == "scroll(-300, 0)"
+
+
+def test_scroll_tolerates_whitespace_in_point():
+    assert _action("scroll(direction='down', point='(612, 455)')") == "scroll(0, 455)"
+
+
+def test_click_point_regression():
+    assert _action("click(point='(100,200)')") == "mouse_click(x=100, y=200)"
+
+
+def test_type_regression():
+    assert _action("type(content='hello\\n')") == "keyboard_type(text='hello\\n')"