tmp save

cua_lite/data/interfaces/_qwen3_vl.py

@@ -0,0 +1,256 @@
+import dataclasses
+import json
+import copy
+from typing import Callable, Any
+
+from cua_lite.data.interfaces.base import UnrolledContextDataInterface
+
+
+DESCRIPTION_PROMPT = """Use a mouse and keyboard to interact with a computer, and take screenshots.
+* This is an interface to a desktop GUI. You do not have access to a terminal or applications menu. You must click on desktop icons to start applications.
+* Some applications may take time to start or process actions, so you may need to wait and take successive screenshots to see the results of your actions. For example, if you click on Firefox and a window does not open, try waiting and then taking another screenshot.
+* The screen resolution is 1000x1000.
+* Whenever you intend to move the cursor to click on an element such as an icon, consult a screenshot first to determine the element’s coordinates before moving the cursor.
+* If you tried clicking on a program or link but it failed to load even after waiting, adjust your cursor position so that the tip of the cursor visually falls on the element you want to click.
+* Make sure to click any buttons, links, icons, or other elements with the cursor tip in the center of the element. Do not click on edges unless explicitly instructed.\
+"""
+
+ACTION_DESCRIPTION_PROMPT = """\
+* `key`: Performs key down presses on the arguments passed in order, then performs key releases in reverse order.
+* `type`: Type a string of text on the keyboard.
+* `mouse_move`: Move the cursor to a specified (x, y) pixel coordinate on the screen.
+* `left_click`: Click the left mouse button at a specified (x, y) pixel coordinate on the screen.
+* `left_click_drag`: Click and drag the cursor to a specified (x, y) pixel coordinate on the screen.
+* `right_click`: Click the right mouse button at a specified (x, y) pixel coordinate on the screen.
+* `middle_click`: Click the middle mouse button at a specified (x, y) pixel coordinate on the screen.
+* `double_click`: Double-click the left mouse button at a specified (x, y) pixel coordinate on the screen.
+* `triple_click`: Triple-click the left mouse button at a specified (x, y) pixel coordinate on the screen (simulated as double-click since it's the closest action).
+* `scroll`: Performs a scroll of the mouse scroll wheel.
+* `hscroll`: Performs a horizontal scroll (mapped to regular scroll).
+* `wait`: Wait specified seconds for the change to happen.
+* `terminate`: Terminate the current task and report its completion status.
+* `answer`: Answer a question.\
+"""
+
+TOOLS_DEF = {
+    "type": "function",
+    "function": {
+        "name_for_human": "computer_use",
+        "name": "computer_use",
+        "description": DESCRIPTION_PROMPT,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": ACTION_DESCRIPTION_PROMPT,
+                    "enum": [
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "scroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": "Required only by `action=key`.",
+                    "type": "array",
+                },
+                "text": {
+                    "description": "Required only by `action=type`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The x,y coordinates for mouse actions.",
+                    "type": "array",
+                },
+                "pixels": {"description": "The amount of scrolling.", "type": "number"},
+                "time": {"description": "The seconds to wait.", "type": "number"},
+                "status": {
+                    "description": "The status of the task.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+
+SYSTEM_PROMPT = (
+    """# Tools
+
+You may call one or more functions to assist with the user query.
+
+You are provided with function signatures within <tools></tools> XML tags:
+<tools>
+"""
+    + json.dumps(TOOLS_DEF)
+    + """
+</tools>
+
+For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:
+<tool_call>
+{"name": <function-name>, "arguments": <args-json-object>}
+</tool_call>
+
+# Response format
+
+Response format for every step:
+1) Action: a short imperative describing what to do in the UI.
+2) A single <tool_call>...</tool_call> block containing only the JSON: {"name": <function-name>, "arguments": <args-json-object>}.
+
+Rules:
+- Output exactly in the order: Action, <tool_call>.
+- Be brief: one sentence for Action.
+- Do not output anything else outside those parts.
+- If finishing, use action=terminate in the tool call."""
+)
+
+
+INSTRUCTION_PROMPT = """
+Please generate the next move according to the UI screenshot, instruction and previous actions.
+
+Instruction: {instruction}
+
+Previous actions:
+{previous_actions}"""
+
+
+@dataclasses.dataclass
+class Qwen3VLDataInterface(UnrolledContextDataInterface):
+
+    history_n: int = 4
+    add_system_prompt: bool = True
+
+    def process_context(self, row: dict[str, Any]) -> dict[str, Any]:
+        # 1. 基础拆分
+        messages = row["messages"]
+
+        # 2. 按 Step 分组 (User + Assistant)
+        steps = []
+        for i in range(0, len(messages), 2):
+            step = {
+                "user": messages[i],
+                "assistant": messages[i + 1] if i + 1 < len(messages) else None,
+            }
+            steps.append(step)
+
+        total_steps = len(steps)
+
+        # 3. 计算保留逻辑
+        steps_to_keep = self.history_n + 1
+
+        # ======== FIX 1: 不要在这里 return ========
+        # 如果总步数不够截断，就令 truncate_count=0（anchor=Step1），但仍然重写 INSTR_PROMPT
+        truncate_count = max(0, total_steps - steps_to_keep)
+        # =========================================
+
+        # 4. 提取 Step 1 的指令
+        step1_content = steps[0]["user"]["content"]
+        instruction_text = next(
+            item["text"] for item in step1_content if item.get("type") == "text"
+        )
+
+        # if "Previous actions" in instruction_text:
+        #     instruction_text = instruction_text.split("Previous actions")[0].strip()
+
+        # # 规范成 “纯 instruction 内容”
+        # if instruction_text.startswith("Instruction:"):
+        #     instruction_text = instruction_text[len("Instruction:"):].strip()
+
+        # 5. 生成历史摘要（被截断掉的那些步）
+        summary_lines = []
+        for i in range(truncate_count):
+            action = (
+                steps[i]["assistant"]["content"]
+                if isinstance(steps[i]["assistant"]["content"], str)
+                else steps[i]["assistant"]["content"][0]["text"]
+            )
+            summary_lines.append(f"Step {i+1}: {action}")
+
+        summary_block = "\n".join(summary_lines) if summary_lines else "None"
+
+        # 生成与 QwenAgent 一致的 INSTR_PROMPT
+        full_text_prompt = (
+            "\nPlease generate the next move according to the UI screenshot, instruction and previous actions.\n\n"
+            f"Instruction: {instruction_text}\n\n"
+            "Previous actions:\n"
+            f"{summary_block}"
+        )
+
+        # 6. 构建新消息列表
+        processed_messages = []
+        if self.add_system_prompt:
+            processed_messages.append(
+                {"role": "system", "content": [{"type": "text", "text": SYSTEM_PROMPT}]}
+            )
+
+        # --- Anchor Step (新起点) ---
+        anchor_step = steps[truncate_count]
+        new_anchor_user = copy.deepcopy(anchor_step["user"])
+
+        # 找到 User 消息里的 text 字段并修改，或者追加
+        text_item = next(
+            (item for item in new_anchor_user["content"] if item.get("type") == "text"),
+            None,
+        )
+        if text_item:
+            text_item["text"] = full_text_prompt
+        else:
+            new_anchor_user["content"].append(
+                {"type": "text", "text": full_text_prompt}
+            )
+
+        processed_messages.append(new_anchor_user)
+        if anchor_step["assistant"]:
+            asst_msg = copy.deepcopy(anchor_step["assistant"])
+            # Prepend "Action: " using a clean reference
+            content_node = asst_msg["content"][0]
+            content_node["text"] = "Action: " + content_node["text"]
+            processed_messages.append(asst_msg)
+
+        # --- 追加剩余步骤 ---
+        for i in range(truncate_count + 1, total_steps):
+            # ======== FIX 2: 对齐 QwenAgent：非 anchor 的 user 只保留 image 部分 ========
+            u = copy.deepcopy(steps[i]["user"])
+            if isinstance(u.get("content"), list):
+                u["content"] = [it for it in u["content"] if it.get("type") != "text"]
+            processed_messages.append(u)
+            # ================================================================
+            if steps[i]["assistant"]:
+                asst_msg = copy.deepcopy(steps[i]["assistant"])
+                # Prepend "Action: " using a clean reference
+                content_node = asst_msg["content"][0]
+                content_node["text"] = "Action: " + content_node["text"]
+                processed_messages.append(asst_msg)
+
+        return {"messages": processed_messages}
+
+
+if __name__ == "__main__":
+    from datasets import load_from_disk
+    from transformers import AutoProcessor
+    from cua_lite.data.utils import clean_nones
+
+    processor = AutoProcessor.from_pretrained(
+        "/mnt/lustrenew/mllm_aligned/shared/models/huggingface/Qwen/Qwen3-VL-2B-Thinking"
+    )
+
+    dataset = load_from_disk(".data/unzipped/scalecua/ubuntu/shard-00000-of-00256")
+    dataset.set_transform(clean_nones)
+
+    qwen3vl_interface = Qwen3VLDataInterface(history_n=0)
+    dataset_mapped = dataset.map(
+        qwen3vl_interface.process_context_batch, batched=True, batch_size=1000
+    )
+
+    inputs = processor.apply_chat_template(dataset_mapped["messages"][1], return_dict=True)
+    breakpoint()

cua_lite/data/interfaces/base.py

@@ -4,31 +4,80 @@ from typing import Any
 from cua_lite.data.utils import batch_proc
 
 
+# @dataclasses.dataclass
+# class BaseDataInterface:
+
+#     # -- Helpers (shared between grounding and trajectory) --
+#     def _add_system_prompt(self, row: dict[str, Any], system_prompt: str | None = None) -> dict[str, Any]:
+#         if system_prompt and row["messages"][0]["role"] != "system":
+#             row["messages"].insert(0, {"role": "system", "content": [{"type": "text", "text": system_prompt}]})
+#         return row
+
+#     def _process_action(self, row: dict[str, Any], **kwargs) -> dict[str, Any]:
+#         return row
+
+#     # -- Grounding ---
+#     def process_grounding(self, row: dict[str, Any], system_prompt: str | None = None, **kwargs) -> dict[str, Any]:
+#         return self._add_system_prompt(
+#             self._process_action(row, **kwargs), 
+#             system_prompt=system_prompt
+#         )
+
+#     def process_grounding_batch(self, batch: dict[str, list[Any]], **kwargs) -> dict[str, list[Any]]:
+#         return batch_proc(self.process_grounding, batch, **kwargs)
+
+#     # --- Trajectory ---
+#     def _process_context(self, row: dict[str, Any], **kwargs) -> dict[str, Any]:
+#         return row
+
+#     def process_trajectory(self, row: dict[str, Any], system_prompt: str | None = None, **kwargs) -> dict[str, Any]:
+#         return self._add_system_prompt(
+#             self._process_context(self._process_action(row, **kwargs), **kwargs), 
+#             system_prompt=system_prompt
+#         )
+
+#     def process_trajectory_batch(self, batch: dict[str, list[Any]], **kwargs) -> dict[str, list[Any]]:
+#         return batch_proc(self.process_trajectory, batch, **kwargs)
+
 @dataclasses.dataclass
 class BaseDataInterface:
 
-    # --- Grounding ---
-    # def process_grounding(self, row: dict[str, Any]) -> dict[str, Any]:
-    #     return row
+    grounding_system_prompt: str | None = None
+    trajectory_system_prompt: str | None = None
 
-    # def process_grounding_batch(self, batch: dict[str, list[Any]]) -> dict[str, list[Any]]:
-    #     return batch_proc(self.process_grounding, batch)
+    # -- Helpers (shared between grounding and trajectory) --
+    def _add_system_prompt(self, row: dict[str, Any], system_prompt: str | None = None) -> dict[str, Any]:
+        if system_prompt and row["messages"][0]["role"] != "system":
+            row["messages"].insert(0, {"role": "system", "content": [{"type": "text", "text": system_prompt}]})
+        return row
 
-    # --- Planning ---
-    def process_action(self, row: dict[str, Any]) -> dict[str, Any]:
+    def _process_action(self, row: dict[str, Any], **kwargs) -> dict[str, Any]:
+        # TODO: overwrite this in subclass 
         return row
 
-    def process_action_batch(self, batch: dict[str, list[Any]]) -> dict[str, list[Any]]:
-        return batch_proc(self.process_action, batch)
+    # -- Grounding ---
+    def process_grounding(self, row: dict[str, Any], **kwargs) -> dict[str, Any]:
+        return self._add_system_prompt(
+            self._process_action(row, **kwargs), 
+            system_prompt=self.grounding_system_prompt
+        )
 
-    # --- Context ---
-    def process_context(self, row: dict[str, Any]) -> dict[str, Any]:
+    def process_grounding_batch(self, batch: dict[str, list[Any]], **kwargs) -> dict[str, list[Any]]:
+        return batch_proc(self.process_grounding, batch, **kwargs)
+
+    # --- Trajectory ---
+    def _process_context(self, row: dict[str, Any], **kwargs) -> dict[str, Any]:
+        # TODO: overwrite this in subclass
         return row
 
-    def process_context_batch(
-        self, batch: dict[str, list[Any]]
-    ) -> dict[str, list[Any]]:
-        return batch_proc(self.process_context, batch)
+    def process_trajectory(self, row: dict[str, Any], **kwargs) -> dict[str, Any]:
+        return self._add_system_prompt(
+            self._process_context(self._process_action(row, **kwargs), **kwargs), 
+            system_prompt=self.trajectory_system_prompt
+        )
+
+    def process_trajectory_batch(self, batch: dict[str, list[Any]], **kwargs) -> dict[str, list[Any]]:
+        return batch_proc(self.process_trajectory, batch, **kwargs)
 
 
 @dataclasses.dataclass
@@ -36,9 +85,8 @@ class UnrolledContextDataInterface(BaseDataInterface):
 
     """Useful for reasoning models"""
 
-    def process_context_batch(
-        self, batch: dict[str, list[Any]]
-    ) -> dict[str, list[Any]]:
+    def process_trajectory_batch(
+        self, batch: dict[str, list[Any]], **kwargs) -> dict[str, list[Any]]:
         """
         Handles 1-to-N data expansion (Unrolling conversation history).
         Note: This changes the number of rows.
@@ -51,10 +99,9 @@ class UnrolledContextDataInterface(BaseDataInterface):
             ]
             for assistant_index in assistant_indices:
                 # Slicing includes the assistant message
-                context = messages[: assistant_index + 1]
-                processed_context = self.process_context({"messages": context})[
-                    "messages"
-                ]
-                messages_list.append(processed_context)
+                processed_messages = self.process_trajectory({"messages": messages[: assistant_index + 1]}, **kwargs)["messages"]
+                messages_list.append(processed_messages)
+
+        batch = {"messages": messages_list}
 
-        return {"messages": messages_list}
+        return batch

cua_lite/data/interfaces/qwen3_vl/interfaces.py

@@ -11,121 +11,123 @@ from cua_lite.data.interfaces.qwen3_vl.prompts import (
     MOBILE_SYSTEM_PROMPT,
 )
 
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1 = ""
+
+MOBILE_GROUNDING_SYSTEM_PROMPT_V1 = ""
+MOBILE_SYSTEM_PROMPT_V1 = ""
+
+MOBILE_GROUNDING_SYSTEM_PROMPT_V0 = ""
+MOBILE_SYSTEM_PROMPT_V0 = ""
+
 
 @dataclasses.dataclass
-class Qwen3VLDataInterface(UnrolledContextDataInterface):
+class Qwen3VLBaseDataInterface(UnrolledContextDataInterface):
 
     history_n: int = 4
 
-    def process_context(self, row: dict[str, Any]) -> dict[str, Any]:
-        # 1. 基础拆分
-        messages = row["messages"]
+    def _process_context(self, row: dict[str, Any], **kwargs) -> dict[str, Any]:
+        messages = copy.deepcopy(row["messages"])
 
-        # 2. 按 Step 分组 (User + Assistant)
-        steps = []
-        for i in range(0, len(messages), 2):
-            step = {
-                "user": messages[i],
-                "assistant": messages[i + 1] if i + 1 < len(messages) else None,
-            }
-            steps.append(step)
+        # 1. Group User/Assistant pairs
+        steps = [
+            {"user": messages[i], "assistant": messages[i + 1] if i + 1 < len(messages) else None}
+            for i in range(0, len(messages), 2)
+        ]
 
-        total_steps = len(steps)
-
-        # 3. 计算保留逻辑
+        # 2. Calculate truncation
         steps_to_keep = self.history_n + 1
+        truncate_count = max(0, len(steps) - steps_to_keep)
 
-        # ======== FIX 1: 不要在这里 return ========
-        # 如果总步数不够截断，就令 truncate_count=0（anchor=Step1），但仍然重写 INSTR_PROMPT
-        truncate_count = max(0, total_steps - steps_to_keep)
-        # =========================================
-
-        # 4. 提取 Step 1 的指令
-        step1_content = steps[0]["user"]["content"]
-        instruction_text = next(
-            item["text"] for item in step1_content if item.get("type") == "text"
+        # 3. Extract Instruction (from Step 1) and create Summary
+        step1_text = next(
+            item["text"] for item in steps[0]["user"]["content"] if item.get("type") == "text"
         )
 
-        # if "Previous actions" in instruction_text:
-        #     instruction_text = instruction_text.split("Previous actions")[0].strip()
-
-        # # 规范成 “纯 instruction 内容”
-        # if instruction_text.startswith("Instruction:"):
-        #     instruction_text = instruction_text[len("Instruction:"):].strip()
-
-        # 5. 生成历史摘要（被截断掉的那些步）
-        summary_lines = []
-        for i in range(truncate_count):
-            action = (
-                steps[i]["assistant"]["content"]
-                if isinstance(steps[i]["assistant"]["content"], str)
-                else steps[i]["assistant"]["content"][0]["text"]
-            )
-            summary_lines.append(f"Step {i+1}: {action}")
-
+        summary_lines = [
+            f"Step {i+1}: {step['assistant']['content'][0]['text']}"
+            for i, step in enumerate(steps[:truncate_count])
+            if step["assistant"]
+        ]
         summary_block = "\n".join(summary_lines) if summary_lines else "None"
 
-        # 生成与 QwenAgent 一致的 INSTR_PROMPT
-        # full_text_prompt = (
-        #     "\nPlease generate the next move according to the UI screenshot, instruction and previous actions.\n\n"
-        #     f"Instruction: {instruction_text}\n\n"
-        #     "Previous actions:\n"
-        #     f"{summary_block}"
-        # )
-        full_text_prompt = INSTRUCTION_PROMPT.format(
-            instruction=instruction_text,
-            previous_actions=summary_block,
+        full_prompt = INSTRUCTION_PROMPT.format(
+            instruction=step1_text, previous_actions=summary_block
         )
 
-        # 6. 构建新消息列表
+        # 4. Build new message list from remaining steps
         processed_messages = []
-        if self.add_system_prompt:
-            processed_messages.append(
-                {"role": "system", "content": [{"type": "text", "text": SYSTEM_PROMPT}]}
-            )
-
-        # --- Anchor Step (新起点) ---
-        anchor_step = steps[truncate_count]
-        new_anchor_user = copy.deepcopy(anchor_step["user"])
-
-        # 找到 User 消息里的 text 字段并修改，或者追加
-        text_item = next(
-            (item for item in new_anchor_user["content"] if item.get("type") == "text"),
-            None,
-        )
-        if text_item:
-            text_item["text"] = full_text_prompt
-        else:
-            new_anchor_user["content"].append(
-                {"type": "text", "text": full_text_prompt}
-            )
-
-        processed_messages.append(new_anchor_user)
-        if anchor_step["assistant"]:
-            asst_msg = copy.deepcopy(anchor_step["assistant"])
-            # Prepend "Action: " using a clean reference
-            content_node = asst_msg["content"][0]
-            content_node["text"] = "Action: " + content_node["text"]
-            processed_messages.append(asst_msg)
-
-        # --- 追加剩余步骤 ---
-        for i in range(truncate_count + 1, total_steps):
-            # ======== FIX 2: 对齐 QwenAgent：非 anchor 的 user 只保留 image 部分 ========
-            u = copy.deepcopy(steps[i]["user"])
-            if isinstance(u.get("content"), list):
-                u["content"] = [it for it in u["content"] if it.get("type") != "text"]
-            processed_messages.append(u)
-            # ================================================================
-            if steps[i]["assistant"]:
-                asst_msg = copy.deepcopy(steps[i]["assistant"])
-                # Prepend "Action: " using a clean reference
-                content_node = asst_msg["content"][0]
-                content_node["text"] = "Action: " + content_node["text"]
-                processed_messages.append(asst_msg)
+        for i, step in enumerate(steps[truncate_count:]):
+            u_msg, a_msg = step["user"], step["assistant"]
+
+            # Handle User Message
+            if i == 0:  # Anchor step: Inject full prompt
+                text_item = next((x for x in u_msg["content"] if x.get("type") == "text"), None)
+                if text_item:
+                    text_item["text"] = full_prompt
+                else:
+                    u_msg["content"].append({"type": "text", "text": full_prompt})
+            else:  # Subsequent steps: Remove text, keep images only
+                u_msg["content"] = [x for x in u_msg["content"] if x.get("type") != "text"]
+            
+            processed_messages.append(u_msg)
+
+            # Handle Assistant Message
+            if a_msg:
+                # Prepend "Action: " to the first content block
+                a_msg["content"][0]["text"] = "Action: " + a_msg["content"][0]["text"]
+                processed_messages.append(a_msg)
 
         return {"messages": processed_messages}
 
 
+@dataclasses.dataclass
+class Qwen3VLDesktopDataInterfaceV0(Qwen3VLBaseDataInterface):
+    grounding_system_prompt: str = DESKTOP_GROUNDING_SYSTEM_PROMPT_V0
+    trajectory_system_prompt: str = DESKTOP_TRAJECTORY_SYSTEM_PROMPT_V0
+
+    def _process_action(
+        self,
+        row: dict[str, Any],
+    ) -> dict[str, Any]:
+        # TODO: debug
+        messages = copy.deepcopy(row["messages"])
+        for message in messages:
+            if message["role"] == "assistant":
+                # TODO: pass tool calling results
+                if message["tool_calls"]["function"]["arguments"]["action"] == "hscroll":
+                    message["tool_calls"]["function"]["arguments"]["action"] = "scroll"
+                elif message["tool_calls"]["function"]["arguments"]["action"] == "triple_click":
+                    message["tool_calls"]["function"]["arguments"]["action"] = "double_click"
+        return {"messages": messages}
+
+
+@dataclasses.dataclass
+class Qwen3VLMobileDataInterfaceV0(Qwen3VLBaseDataInterface):
+
+    grounding_system_prompt: str = DESKTOP_GROUNDING_SYSTEM_PROMPT_V1
+    trajectory_system_prompt: str = DESKTOP_TRAJECTORY_SYSTEM_PROMPT_V1
+
+    def process_action(
+        self,
+        row: dict[str, Any],
+    ) -> dict[str, Any]:
+        # TODO: implement
+        pass
+
+
+@dataclasses.dataclass
+class Qwen3VLDesktopDataInterfaceV1(Qwen3VLBaseDataInterface):
+    grounding_system_prompt: str = DESKTOP_GROUNDING_SYSTEM_PROMPT_V1
+    trajectory_system_prompt: str = DESKTOP_SYSTEM_PROMPT_V1
+
+
+@dataclasses.dataclass
+class Qwen3VLMobileDataInterfaceV1(Qwen3VLBaseDataInterface):
+    grounding_system_prompt: str = MOBILE_GROUNDING_SYSTEM_PROMPT_V1
+    trajectory_system_prompt: str = MOBILE_SYSTEM_PROMPT_V1
+
+
+
 if __name__ == "__main__":
     from datasets import load_from_disk
     from transformers import AutoProcessor

cua_lite/data/interfaces/qwen3_vl/prompts/_desktop.py

@@ -0,0 +1,431 @@
+import copy
+import json
+
+# from .common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
+from cua_lite.data.interfaces.qwen3_vl.prompts.common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
+
+DESKTOP_TRAJECTORY_DESCRIPTION = """Use a mouse and keyboard to interact with a computer, and take screenshots.
+* This is an interface to a desktop GUI. You do not have access to a terminal or applications menu. You must click on desktop icons to start applications.
+* Some applications may take time to start or process actions, so you may need to wait and take successive screenshots to see the results of your actions. For example, if you click on Firefox and a window does not open, try waiting and then taking another screenshot.
+* The screen resolution is 1000x1000.
+* Whenever you intend to move the cursor to click on an element such as an icon, consult a screenshot first to determine the element's coordinates before moving the cursor.
+* If you tried clicking on a program or link but it failed to load even after waiting, adjust your cursor position so that the tip of the cursor visually falls on the element you want to click.
+* Make sure to click any buttons, links, icons, or other elements with the cursor tip in the center of the element. Do not click on edges unless explicitly instructed."""
+
+# mention point and bbox
+DESKTOP_GROUNDING_DESCRIPTION = """Given a computer screenshot, map the natural language instruction to an action or map the natural language reference to a point or a bounding box  in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_ACTIONS = """Given a computer screenshot, map the natural language action to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_POINT = """Given a computer screenshot, map the natural language reference to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_BBOX = """Given a computer screenshot, map the natural language reference to a bounding box in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+
+# v0: trajectory (https://github.com/xlang-ai/OSWorld/blob/main/mm_agents/qwen3vl_agent.py)
+# DESKTOP_TRAJECTORY_ACTION_DESCRIPTION_V0 = """\
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_use",
+        "name": "desktop_use",
+        "description": DESKTOP_TRAJECTORY_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `key`: Performs key down presses on the arguments passed in order, then performs key releases in reverse order.\n"
+                        + "* `type`: Type a string of text on the keyboard.\n"
+                        + "* `mouse_move`: Move the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click`: Click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click_drag`: Click and drag the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `right_click`: Click the right mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `middle_click`: Click the middle mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `double_click`: Double-click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `triple_click`: Triple-click the left mouse button at a specified (x, y) pixel coordinate on the screen (simulated as double-click since it's the closest action).\n"
+                        + "* `scroll`: Performs a scroll of the mouse scroll wheel.\n"
+                        + "* `hscroll`: Performs a horizontal scroll (mapped to regular scroll).\n"
+                        + "* `wait`: Wait specified seconds for the change to happen.\n"
+                        + "* `terminate`: Terminate the current task and report its completion status.\n"
+                        + "* `answer`: Answer a question."
+                    ),
+                    "enum": [
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "scroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": "Required only by `action=key`.",
+                    "type": "array",
+                },
+                "text": {
+                    "description": "Required only by `action=type`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The x,y coordinates for mouse actions.",
+                    "type": "array",
+                },
+                "pixels": {"description": "The amount of scrolling.", "type": "number"},
+                "time": {"description": "The seconds to wait.", "type": "number"},
+                "status": {
+                    "description": "The status of the task.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_TRAJECTORY_SYSTEM_PROMPT_V0 = TRAJECTORY_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0))
+
+# v0: grounding
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0_ACTIONS = copy.deepcopy(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0)
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0_ACTIONS["function"]["description"] = DESKTOP_GROUNDING_DESCRIPTION_ACTIONS
+
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0_ACTIONS = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_POINT = {
+    "type": "function",
+    "function": {
+        "name_for_human": "point",
+        "name": "point",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_POINT,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 2-element list [x, y] representing the screen coordinates for the center of the described object.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0_POINT = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_POINT))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_BBOX = {
+    "type": "function",
+    "function": {
+        "name_for_human": "bbox",
+        "name": "bbox",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_BBOX,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 4-element list [x_min, y_min, x_max, y_max] defining the bounding box, where (x_min, y_min) is the top-left corner and (x_max, y_max) is the bottom-right corner.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0_BBOX = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_BBOX))
+
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V0 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_grounding",
+        "name": "desktop_grounding",
+        "description": DESKTOP_GROUNDING_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `point`: Return the [x, y] center coordinates of the object described.\n"
+                        + "* `bbox`: Return the [x_min, y_min, x_max, y_max] bounding box of the object described.\n"
+                        + "* `key`: Performs key down presses on the arguments passed in order, then performs key releases in reverse order.\n"
+                        + "* `type`: Type a string of text on the keyboard.\n"
+                        + "* `mouse_move`: Move the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click`: Click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click_drag`: Click and drag the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `right_click`: Click the right mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `middle_click`: Click the middle mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `double_click`: Double-click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `triple_click`: Triple-click the left mouse button at a specified (x, y) pixel coordinate on the screen (simulated as double-click since it's the closest action).\n"
+                        + "* `scroll`: Performs a scroll of the mouse scroll wheel.\n"
+                        + "* `hscroll`: Performs a horizontal scroll (mapped to regular scroll).\n"
+                        + "* `wait`: Wait specified seconds for the change to happen.\n"
+                        + "* `terminate`: Terminate the current task and report its completion status.\n"
+                        + "* `answer`: Answer a question."
+                    ),
+                    "enum": [
+                        "point",
+                        "bbox",
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "scroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": "Required only by `action=key`.",
+                    "type": "array",
+                },
+                "text": {
+                    "description": "Required only by `action=type`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The [x, y] coordinates for mouse actions/point, or [x_min, y_min, x_max, y_max] for bbox.",
+                    "type": "array",
+                },
+                "pixels": {"description": "The amount of scrolling.", "type": "number"},
+                "time": {"description": "The seconds to wait.", "type": "number"},
+                "status": {
+                    "description": "The status of the task.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0 = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V0))
+
+
+
+# v1
+
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_use",
+        "name": "desktop_use",
+        "description": DESKTOP_TRAJECTORY_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `key(keys=[\"key1\", \"key2\"])`: Press the keys in order (key down), then release them in reverse order (key up).\n"
+                        + "* `type(text=\"string\")`: Type the given text.\n"
+                        + "* `mouse_move(coordinate=[x, y])`: Move the cursor to (x, y).\n"
+                        + "* `left_click(coordinate=[x, y][, keys=[\"key\"]])`: Left click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `left_click_drag(coordinate=[x, y][, keys=[\"key\"]])`: Drag to (x, y) with the left mouse button. If `keys` is provided, hold those keys while dragging.\n"
+                        + "* `right_click(coordinate=[x, y][, keys=[\"key\"]])`: Right click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `middle_click(coordinate=[x, y][, keys=[\"key\"]])`: Middle click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `double_click(coordinate=[x, y][, keys=[\"key\"]])`: Double left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `triple_click(coordinate=[x, y][, keys=[\"key\"]])`: Triple left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `scroll(pixels=number[, keys=[\"key\"]])`: Scroll vertically. `pixels > 0` scrolls up, `pixels < 0` scrolls down. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `hscroll(pixels=number[, keys=[\"key\"]])`: Scroll horizontally. `pixels > 0` scrolls right, `pixels < 0` scrolls left. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `wait(time=seconds)`: Wait for the given number of seconds.\n"
+                        + "* `terminate(status=\"success\"|\"failure\")`: End the task and report the status.\n"
+                    ),
+                    "enum": [
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "triple_click",
+                        "scroll",
+                        "hscroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": (
+                        "Keys to press/hold.\n"
+                        "Required only for action: `key` (press in order, release in reverse).\n"
+                        "Optional only for actions: `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`, `scroll`, `hscroll` (hold these keys while performing the action).\n"
+                    ),
+                    "type": "array",
+                },
+                "text": {
+                    "description": "The text content to type or the answer to return. Required only for actions: `type`, `answer`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The (x, y) pixel coordinate on the screen. Required only for actions: `mouse_move`, `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`.",
+                    "type": "array",
+                },
+                "pixels": {
+                    "description": "The amount of scrolling to perform. Required only for actions: `scroll`, `hscroll`.",
+                    "type": "number"
+                },
+                "time": {
+                    "description": "The duration to wait in seconds. Required only for action: `wait`.",
+                    "type": "number"
+                },
+                "status": {
+                    "description": "The completion status of the task. Required only for action: `terminate`.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+
+DESKTOP_TRAJECTORY_SYSTEM_PROMPT_V1 = TRAJECTORY_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1))
+
+# v1: grounding
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1_ACTIONS = copy.deepcopy(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1)
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1_ACTIONS["function"]["description"] = DESKTOP_GROUNDING_DESCRIPTION_ACTIONS
+
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1_ACTIONS = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_POINT = {
+    "type": "function",
+    "function": {
+        "name_for_human": "point",
+        "name": "point",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_POINT,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 2-element list [x, y] representing the screen coordinates for the center of the described object.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1_POINT = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_POINT))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_BBOX = {
+    "type": "function",
+    "function": {
+        "name_for_human": "bbox",
+        "name": "bbox",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_BBOX,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 4-element list [x_min, y_min, x_max, y_max] defining the bounding box, where (x_min, y_min) is the top-left corner and (x_max, y_max) is the bottom-right corner.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1_BBOX = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_BBOX))
+
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V1 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_grounding",
+        "name": "desktop_grounding",
+        "description": DESKTOP_GROUNDING_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `point`: Return the [x, y] center coordinates of the object described.\n"
+                        + "* `bbox`: Return the [x_min, y_min, x_max, y_max] bounding box coordinates of the object described.\n"
+                        + "* `key(keys=[\"key1\", \"key2\"])`: Press the keys in order (key down), then release them in reverse order (key up).\n"
+                        + "* `type(text=\"string\")`: Type the given text.\n"
+                        + "* `mouse_move(coordinate=[x, y])`: Move the cursor to (x, y).\n"
+                        + "* `left_click(coordinate=[x, y][, keys=[\"key\"]])`: Left click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `left_click_drag(coordinate=[x, y][, keys=[\"key\"]])`: Drag to (x, y) with the left mouse button. If `keys` is provided, hold those keys while dragging.\n"
+                        + "* `right_click(coordinate=[x, y][, keys=[\"key\"]])`: Right click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `middle_click(coordinate=[x, y][, keys=[\"key\"]])`: Middle click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `double_click(coordinate=[x, y][, keys=[\"key\"]])`: Double left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `triple_click(coordinate=[x, y][, keys=[\"key\"]])`: Triple left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `scroll(pixels=number[, keys=[\"key\"]])`: Scroll vertically. `pixels > 0` scrolls up, `pixels < 0` scrolls down. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `hscroll(pixels=number[, keys=[\"key\"]])`: Scroll horizontally. `pixels > 0` scrolls right, `pixels < 0` scrolls left. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `wait(time=seconds)`: Wait for the given number of seconds.\n"
+                        + "* `terminate(status=\"success\"|\"failure\")`: End the task and report the status.\n"
+                        + "* `answer(text=\"answer\")`: Return a final answer in text."
+                    ),
+                    "enum": [
+                        "point",
+                        "bbox",
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "scroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": (
+                        "Keys to press/hold.\n"
+                        "Required only for action: `key` (press in order, release in reverse).\n"
+                        "Optional only for actions: `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`, `scroll`, `hscroll` (hold these keys while performing the action).\n"
+                    ),
+                    "type": "array",
+                },
+                "text": {
+                    "description": "The text content to type or the answer to return. Required only for actions: `type`, `answer`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The [x, y] coordinates for mouse actions/point, or [x_min, y_min, x_max, y_max] for bbox. Required only for actions: `point`, `bbox`, `mouse_move`, `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`.",
+                    "type": "array",
+                },
+                "pixels": {
+                    "description": "The amount of scrolling to perform. Required only for actions: `scroll`, `hscroll`.",
+                    "type": "number"
+                },
+                "time": {
+                    "description": "The duration to wait in seconds. Required only for action: `wait`.",
+                    "type": "number"
+                },
+                "status": {
+                    "description": "The completion status of the task. Required only for action: `terminate`.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1 = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V1))
+
+
+
+if __name__ == "__main__":
+    pass
+

cua_lite/data/interfaces/qwen3_vl/prompts/common.py

@@ -1,38 +1,83 @@
-SYSTEM_PROMPT = (
-    """# Tools
+# TRAJECTORY_SYSTEM_PROMPT = (
+#     """# Tools
+
+# You may call one or more functions to assist with the user query.
+
+# You are provided with function signatures within <tools></tools> XML tags:
+# <tools>
+# """
+#     + "{tools_definition}"
+#     + """
+# </tools>
+
+# For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:
+# <tool_call>
+# {"name": <function-name>, "arguments": <args-json-object>}
+# </tool_call>
+
+# # Response format
+
+# Response format for every step:
+# 1) Action: a short imperative describing what to do in the UI.
+# 2) A single <tool_call>...</tool_call> block containing only the JSON: {"name": <function-name>, "arguments": <args-json-object>}.
+
+# Rules:
+# - Output exactly in the order: Action, <tool_call>.
+# - Be brief: one sentence for Action.
+# - Do not output anything else outside those parts.
+# - If finishing, use action=terminate in the tool call."""
+# )
+
+TRAJECTORY_SYSTEM_PROMPT_TEMPLATE = """# Tools
 
 You may call one or more functions to assist with the user query.
 
 You are provided with function signatures within <tools></tools> XML tags:
 <tools>
-"""
-    + "{tools_definition}"
-    + """
+{tools_definition}
 </tools>
 
 For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:
 <tool_call>
-{"name": <function-name>, "arguments": <args-json-object>}
+{{"name": <function-name>, "arguments": <args-json-object>}}
 </tool_call>
 
 # Response format
 
 Response format for every step:
 1) Action: a short imperative describing what to do in the UI.
-2) A single <tool_call>...</tool_call> block containing only the JSON: {"name": <function-name>, "arguments": <args-json-object>}.
+2) A single <tool_call>...</tool_call> block containing only the JSON: {{"name": <function-name>, "arguments": <args-json-object>}}.
 
 Rules:
 - Output exactly in the order: Action, <tool_call>.
 - Be brief: one sentence for Action.
 - Do not output anything else outside those parts.
 - If finishing, use action=terminate in the tool call."""
-)
 
 
-INSTRUCTION_PROMPT = """
+
+
+TRAJECTORY_INSTRUCTION_PROMPT_TEMPLATE = """
 Please generate the next move according to the UI screenshot, instruction and previous actions.
 
 Instruction: {instruction}
 
 Previous actions:
 {previous_actions}"""
+
+
+
+
+GROUNDING_SYSTEM_PROMPT_TEMPLATE = """# Tools
+
+You may call one or more functions to assist with the user query.
+
+You are provided with function signatures within <tools></tools> XML tags:
+<tools>
+{tools_definition}
+</tools>
+
+For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:
+<tool_call>
+{{"name": <function-name>, "arguments": <args-json-object>}}
+</tool_call>""" 

cua_lite/data/interfaces/qwen3_vl/prompts/desktop.py

@@ -1,45 +1,47 @@
+import copy
 import json
 
-from .common import SYSTEM_PROMPT
+# from .common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
+from cua_lite.data.interfaces.qwen3_vl.prompts.common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
 
-
-DESKTOP_DESCRIPTION_PROMPT = """Use a mouse and keyboard to interact with a computer, and take screenshots.
+DESKTOP_TRAJECTORY_DESCRIPTION = """Use a mouse and keyboard to interact with a computer, and take screenshots.
 * This is an interface to a desktop GUI. You do not have access to a terminal or applications menu. You must click on desktop icons to start applications.
 * Some applications may take time to start or process actions, so you may need to wait and take successive screenshots to see the results of your actions. For example, if you click on Firefox and a window does not open, try waiting and then taking another screenshot.
 * The screen resolution is 1000x1000.
-* Whenever you intend to move the cursor to click on an element such as an icon, consult a screenshot first to determine the element’s coordinates before moving the cursor.
+* Whenever you intend to move the cursor to click on an element such as an icon, consult a screenshot first to determine the element's coordinates before moving the cursor.
 * If you tried clicking on a program or link but it failed to load even after waiting, adjust your cursor position so that the tip of the cursor visually falls on the element you want to click.
-* Make sure to click any buttons, links, icons, or other elements with the cursor tip in the center of the element. Do not click on edges unless explicitly instructed.\
-"""
+* Make sure to click any buttons, links, icons, or other elements with the cursor tip in the center of the element. Do not click on edges unless explicitly instructed."""
 
-# v0 (https://github.com/xlang-ai/OSWorld/blob/main/mm_agents/qwen3vl_agent.py)
-DESKTOP_ACTION_DESCRIPTION_PROMPT_V0 = """\
-* `key`: Performs key down presses on the arguments passed in order, then performs key releases in reverse order.
-* `type`: Type a string of text on the keyboard.
-* `mouse_move`: Move the cursor to a specified (x, y) pixel coordinate on the screen.
-* `left_click`: Click the left mouse button at a specified (x, y) pixel coordinate on the screen.
-* `left_click_drag`: Click and drag the cursor to a specified (x, y) pixel coordinate on the screen.
-* `right_click`: Click the right mouse button at a specified (x, y) pixel coordinate on the screen.
-* `middle_click`: Click the middle mouse button at a specified (x, y) pixel coordinate on the screen.
-* `double_click`: Double-click the left mouse button at a specified (x, y) pixel coordinate on the screen.
-* `triple_click`: Triple-click the left mouse button at a specified (x, y) pixel coordinate on the screen (simulated as double-click since it's the closest action).
-* `scroll`: Performs a scroll of the mouse scroll wheel.
-* `hscroll`: Performs a horizontal scroll (mapped to regular scroll).
-* `wait`: Wait specified seconds for the change to happen.
-* `terminate`: Terminate the current task and report its completion status.
-* `answer`: Answer a question.\
-"""
+# mention point and bbox
+DESKTOP_GROUNDING_DESCRIPTION = """Given a computer screenshot, map the natural language instruction to an action or map the natural language reference to a point or a bounding box  in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+# DESKTOP_GROUNDING_DESCRIPTION_ACTIONS = """Given a computer screenshot, map the natural language action to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+# DESKTOP_GROUNDING_DESCRIPTION_POINT = """Given a computer screenshot, map the natural language reference to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+# DESKTOP_GROUNDING_DESCRIPTION_BBOX = """Given a computer screenshot, map the natural language reference to a bounding box in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
 
-DESKTOP_TOOLS_DEFINITION_V0 = {
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION = {
     "type": "function",
     "function": {
         "name_for_human": "desktop_use",
         "name": "desktop_use",
-        "description": DESKTOP_DESCRIPTION_PROMPT,
+        "description": DESKTOP_TRAJECTORY_DESCRIPTION,
         "parameters": {
             "properties": {
                 "action": {
-                    "description": DESKTOP_ACTION_DESCRIPTION_PROMPT_V0,
+                    "description": (
+                        "* `key(keys=[\"key1\", \"key2\"])`: Press the keys in order (key down), then release them in reverse order (key up).\n"
+                        + "* `type(text=\"string\")`: Type the given text.\n"
+                        + "* `mouse_move(coordinate=[x, y])`: Move the cursor to (x, y).\n"
+                        + "* `left_click(coordinate=[x, y][, keys=[\"key\"]])`: Left click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `left_click_drag(coordinate=[x, y][, keys=[\"key\"]])`: Drag to (x, y) with the left mouse button. If `keys` is provided, hold those keys while dragging.\n"
+                        + "* `right_click(coordinate=[x, y][, keys=[\"key\"]])`: Right click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `middle_click(coordinate=[x, y][, keys=[\"key\"]])`: Middle click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `double_click(coordinate=[x, y][, keys=[\"key\"]])`: Double left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `triple_click(coordinate=[x, y][, keys=[\"key\"]])`: Triple left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `scroll(pixels=number[, keys=[\"key\"]])`: Scroll vertically. `pixels > 0` scrolls up, `pixels < 0` scrolls down. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `hscroll(pixels=number[, keys=[\"key\"]])`: Scroll horizontally. `pixels > 0` scrolls right, `pixels < 0` scrolls left. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `wait(time=seconds)`: Wait for the given number of seconds.\n"
+                        + "* `terminate(status=\"success\"|\"failure\")`: End the task and report the status.\n"
+                    ),
                     "enum": [
                         "key",
                         "type",
@@ -49,28 +51,40 @@ DESKTOP_TOOLS_DEFINITION_V0 = {
                         "right_click",
                         "middle_click",
                         "double_click",
+                        "triple_click",
                         "scroll",
+                        "hscroll",
                         "wait",
                         "terminate",
                     ],
                     "type": "string",
                 },
                 "keys": {
-                    "description": "Required only by `action=key`.",
+                    "description": (
+                        "Keys to press/hold.\n"
+                        "Required only for action: `key` (press in order, release in reverse).\n"
+                        "Optional only for actions: `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`, `scroll`, `hscroll` (hold these keys while performing the action).\n"
+                    ),
                     "type": "array",
                 },
                 "text": {
-                    "description": "Required only by `action=type`.",
+                    "description": "The text content to type or the answer to return. Required only for actions: `type`, `answer`.",
                     "type": "string",
                 },
                 "coordinate": {
-                    "description": "The x,y coordinates for mouse actions.",
+                    "description": "The (x, y) pixel coordinate on the screen. Required only for actions: `mouse_move`, `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`.",
                     "type": "array",
                 },
-                "pixels": {"description": "The amount of scrolling.", "type": "number"},
-                "time": {"description": "The seconds to wait.", "type": "number"},
+                "pixels": {
+                    "description": "The amount of scrolling to perform. Required only for actions: `scroll`, `hscroll`.",
+                    "type": "number"
+                },
+                "time": {
+                    "description": "The duration to wait in seconds. Required only for action: `wait`.",
+                    "type": "number"
+                },
                 "status": {
-                    "description": "The status of the task.",
+                    "description": "The completion status of the task. Required only for action: `terminate`.",
                     "type": "string",
                     "enum": ["success", "failure"],
                 },
@@ -81,39 +95,87 @@ DESKTOP_TOOLS_DEFINITION_V0 = {
         "args_format": "Format the arguments as a JSON object.",
     },
 }
+DESKTOP_TRAJECTORY_SYSTEM_PROMPT = TRAJECTORY_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION ))
+
+# grounding
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_ACTIONS = copy.deepcopy(DESKTOP_TRAJECTORY_TOOLS_DEFINITION)
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_ACTIONS["function"]["description"] = DESKTOP_GROUNDING_DESCRIPTION_ACTIONS
+
+DESKTOP_GROUNDING_SYSTEM_PROMPT_ACTIONS = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION))
 
-DESKTOP_SYSTEM_PROMPT_V0 = SYSTEM_PROMPT.format(tools_definition=json.dumps(DESKTOP_TOOLS_DEFINITION_V0))
+DESKTOP_GROUNDING_TOOLS_DEFINITION_POINT = {
+    "type": "function",
+    "function": {
+        "name_for_human": "point",
+        "name": "point",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_POINT,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 2-element list [x, y] representing the screen coordinates for the center of the described object.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_POINT = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_POINT))
 
+DESKTOP_GROUNDING_TOOLS_DEFINITION_BBOX = {
+    "type": "function",
+    "function": {
+        "name_for_human": "bbox",
+        "name": "bbox",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_BBOX,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 4-element list [x_min, y_min, x_max, y_max] defining the bounding box, where (x_min, y_min) is the top-left corner and (x_max, y_max) is the bottom-right corner.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_BBOX = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_BBOX))
 
-# v1
-DESKTOP_ACTION_DESCRIPTION_PROMPT_V1 = """\
-* `key(keys=["key1", "key2"])`: Press the keys in order (key down), then release them in reverse order (key up).
-* `type(text="string")`: Type the given text.
-* `mouse_move(coordinate=[x, y])`: Move the cursor to (x, y).
-* `left_click(coordinate=[x, y][, keys=["key"]])`: Left click at (x, y). If `keys` is provided, hold those keys while clicking.
-* `left_click_drag(coordinate=[x, y][, keys=["key"]])`: Drag to (x, y) with the left mouse button. If `keys` is provided, hold those keys while dragging.
-* `right_click(coordinate=[x, y][, keys=["key"]])`: Right click at (x, y). If `keys` is provided, hold those keys while clicking.
-* `middle_click(coordinate=[x, y][, keys=["key"]])`: Middle click at (x, y). If `keys` is provided, hold those keys while clicking.
-* `double_click(coordinate=[x, y][, keys=["key"]])`: Double left click at (x, y). If `keys` is provided, hold those keys while double-clicking.
-* `triple_click(coordinate=[x, y][, keys=["key"]])`: Triple left click at (x, y). If `keys` is provided, hold those keys while double-clicking.
-* `scroll(pixels=number[, keys=["key"]])`: Scroll vertically. `pixels > 0` scrolls up, `pixels < 0` scrolls down. If `keys` is provided, hold those keys while scrolling.
-* `hscroll(pixels=number[, keys=["key"]])`: Scroll horizontally. `pixels > 0` scrolls right, `pixels < 0` scrolls left. If `keys` is provided, hold those keys while scrolling.
-* `wait(time=seconds)`: Wait for the given number of seconds.
-* `terminate(status="success"|"failure")`: End the task and report the status.
-* `answer(text="answer")`: Return a final answer in text.\
-"""
 
-DESKTOP_TOOLS_DEFINITION_V1 = {
+DESKTOP_GROUNDING_TOOLS_DEFINITION = {
     "type": "function",
     "function": {
-        "name_for_human": "desktop_use",
-        "name": "desktop_use",
-        "description": DESKTOP_DESCRIPTION_PROMPT,
+        "name_for_human": "desktop_grounding",
+        "name": "desktop_grounding",
+        "description": DESKTOP_GROUNDING_DESCRIPTION,
         "parameters": {
             "properties": {
                 "action": {
-                    "description": DESKTOP_ACTION_DESCRIPTION_PROMPT_V1,
+                    "description": (
+                        "* `point`: Return the [x, y] center coordinates of the object described.\n"
+                        + "* `bbox`: Return the [x_min, y_min, x_max, y_max] bounding box coordinates of the object described.\n"
+                        + "* `key(keys=[\"key1\", \"key2\"])`: Press the keys in order (key down), then release them in reverse order (key up).\n"
+                        + "* `type(text=\"string\")`: Type the given text.\n"
+                        + "* `mouse_move(coordinate=[x, y])`: Move the cursor to (x, y).\n"
+                        + "* `left_click(coordinate=[x, y][, keys=[\"key\"]])`: Left click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `left_click_drag(coordinate=[x, y][, keys=[\"key\"]])`: Drag to (x, y) with the left mouse button. If `keys` is provided, hold those keys while dragging.\n"
+                        + "* `right_click(coordinate=[x, y][, keys=[\"key\"]])`: Right click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `middle_click(coordinate=[x, y][, keys=[\"key\"]])`: Middle click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `double_click(coordinate=[x, y][, keys=[\"key\"]])`: Double left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `triple_click(coordinate=[x, y][, keys=[\"key\"]])`: Triple left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `scroll(pixels=number[, keys=[\"key\"]])`: Scroll vertically. `pixels > 0` scrolls up, `pixels < 0` scrolls down. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `hscroll(pixels=number[, keys=[\"key\"]])`: Scroll horizontally. `pixels > 0` scrolls right, `pixels < 0` scrolls left. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `wait(time=seconds)`: Wait for the given number of seconds.\n"
+                        + "* `terminate(status=\"success\"|\"failure\")`: End the task and report the status.\n"
+                        + "* `answer(text=\"answer\")`: Return a final answer in text."
+                    ),
                     "enum": [
+                        "point",
+                        "bbox",
                         "key",
                         "type",
                         "mouse_move",
@@ -127,7 +189,6 @@ DESKTOP_TOOLS_DEFINITION_V1 = {
                         "hscroll",
                         "wait",
                         "terminate",
-                        "answer"
                     ],
                     "type": "string",
                 },
@@ -135,7 +196,7 @@ DESKTOP_TOOLS_DEFINITION_V1 = {
                     "description": (
                         "Keys to press/hold.\n"
                         "Required only for action: `key` (press in order, release in reverse).\n"
-                        "Optional only for actions: `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `double_click`, `scroll`, `hscroll` (hold these keys while performing the action).\n"
+                        "Optional only for actions: `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`, `scroll`, `hscroll` (hold these keys while performing the action).\n"
                     ),
                     "type": "array",
                 },
@@ -144,11 +205,11 @@ DESKTOP_TOOLS_DEFINITION_V1 = {
                     "type": "string",
                 },
                 "coordinate": {
-                    "description": "The (x, y) pixel coordinate on the screen. Required only for actions: `mouse_move`, `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`.",
+                    "description": "The [x, y] coordinates for mouse actions/point, or [x_min, y_min, x_max, y_max] for bbox. Required only for actions: `point`, `bbox`, `mouse_move`, `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`.",
                     "type": "array",
                 },
                 "pixels": {
-                    "description": "The amount of scrolling to perform. Required only fors actions: `scroll`, `hscroll`.",
+                    "description": "The amount of scrolling to perform. Required only for actions: `scroll`, `hscroll`.",
                     "type": "number"
                 },
                 "time": {
@@ -167,5 +228,9 @@ DESKTOP_TOOLS_DEFINITION_V1 = {
         "args_format": "Format the arguments as a JSON object.",
     },
 }
+DESKTOP_GROUNDING_SYSTEM_PROMPT = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION))
+
+
+if __name__ == "__main__":
+    pass
 
-DESKTOP_SYSTEM_PROMPT_V1 = SYSTEM_PROMPT.format(tools_definition=json.dumps(DESKTOP_TOOLS_DEFINITION_V1))

cua_lite/data/interfaces/qwen3_vl/prompts/version_0/desktop_v0.py

@@ -0,0 +1,209 @@
+import copy
+import json
+
+# from .common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
+from cua_lite.data.interfaces.qwen3_vl.prompts.common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
+
+DESKTOP_TRAJECTORY_DESCRIPTION = """Use a mouse and keyboard to interact with a computer, and take screenshots.
+* This is an interface to a desktop GUI. You do not have access to a terminal or applications menu. You must click on desktop icons to start applications.
+* Some applications may take time to start or process actions, so you may need to wait and take successive screenshots to see the results of your actions. For example, if you click on Firefox and a window does not open, try waiting and then taking another screenshot.
+* The screen resolution is 1000x1000.
+* Whenever you intend to move the cursor to click on an element such as an icon, consult a screenshot first to determine the element's coordinates before moving the cursor.
+* If you tried clicking on a program or link but it failed to load even after waiting, adjust your cursor position so that the tip of the cursor visually falls on the element you want to click.
+* Make sure to click any buttons, links, icons, or other elements with the cursor tip in the center of the element. Do not click on edges unless explicitly instructed."""
+
+# mention point and bbox
+DESKTOP_GROUNDING_DESCRIPTION = """Given a computer screenshot, map the natural language instruction to an action or map the natural language reference to a point or a bounding box  in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_ACTIONS = """Given a computer screenshot, map the natural language action to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_POINT = """Given a computer screenshot, map the natural language reference to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_BBOX = """Given a computer screenshot, map the natural language reference to a bounding box in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+
+# v0: trajectory (https://github.com/xlang-ai/OSWorld/blob/main/mm_agents/qwen3vl_agent.py)
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_use",
+        "name": "desktop_use",
+        "description": DESKTOP_TRAJECTORY_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `key`: Performs key down presses on the arguments passed in order, then performs key releases in reverse order.\n"
+                        + "* `type`: Type a string of text on the keyboard.\n"
+                        + "* `mouse_move`: Move the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click`: Click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click_drag`: Click and drag the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `right_click`: Click the right mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `middle_click`: Click the middle mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `double_click`: Double-click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `triple_click`: Triple-click the left mouse button at a specified (x, y) pixel coordinate on the screen (simulated as double-click since it's the closest action).\n"
+                        + "* `scroll`: Performs a scroll of the mouse scroll wheel.\n"
+                        + "* `hscroll`: Performs a horizontal scroll (mapped to regular scroll).\n"
+                        + "* `wait`: Wait specified seconds for the change to happen.\n"
+                        + "* `terminate`: Terminate the current task and report its completion status.\n"
+                        + "* `answer`: Answer a question."
+                    ),
+                    "enum": [
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "scroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": "Required only by `action=key`.",
+                    "type": "array",
+                },
+                "text": {
+                    "description": "Required only by `action=type`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The x,y coordinates for mouse actions.",
+                    "type": "array",
+                },
+                "pixels": {"description": "The amount of scrolling.", "type": "number"},
+                "time": {"description": "The seconds to wait.", "type": "number"},
+                "status": {
+                    "description": "The status of the task.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_TRAJECTORY_SYSTEM_PROMPT_V0 = TRAJECTORY_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0))
+
+# v0: grounding
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0_ACTIONS = copy.deepcopy(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0)
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0_ACTIONS["function"]["description"] = DESKTOP_GROUNDING_DESCRIPTION_ACTIONS
+
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0_ACTIONS = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V0))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_POINT = {
+    "type": "function",
+    "function": {
+        "name_for_human": "point",
+        "name": "point",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_POINT,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 2-element list [x, y] representing the screen coordinates for the center of the described object.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0_POINT = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_POINT))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_BBOX = {
+    "type": "function",
+    "function": {
+        "name_for_human": "bbox",
+        "name": "bbox",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_BBOX,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 4-element list [x_min, y_min, x_max, y_max] defining the bounding box, where (x_min, y_min) is the top-left corner and (x_max, y_max) is the bottom-right corner.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0_BBOX = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V0_BBOX))
+
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V0 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_grounding",
+        "name": "desktop_grounding",
+        "description": DESKTOP_GROUNDING_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `point`: Return the [x, y] center coordinates of the object described.\n"
+                        + "* `bbox`: Return the [x_min, y_min, x_max, y_max] bounding box of the object described.\n"
+                        + "* `key`: Performs key down presses on the arguments passed in order, then performs key releases in reverse order.\n"
+                        + "* `type`: Type a string of text on the keyboard.\n"
+                        + "* `mouse_move`: Move the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click`: Click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `left_click_drag`: Click and drag the cursor to a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `right_click`: Click the right mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `middle_click`: Click the middle mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `double_click`: Double-click the left mouse button at a specified (x, y) pixel coordinate on the screen.\n"
+                        + "* `triple_click`: Triple-click the left mouse button at a specified (x, y) pixel coordinate on the screen (simulated as double-click since it's the closest action).\n"
+                        + "* `scroll`: Performs a scroll of the mouse scroll wheel.\n"
+                        + "* `hscroll`: Performs a horizontal scroll (mapped to regular scroll).\n"
+                        + "* `wait`: Wait specified seconds for the change to happen.\n"
+                        + "* `terminate`: Terminate the current task and report its completion status.\n"
+                        + "* `answer`: Answer a question."
+                    ),
+                    "enum": [
+                        "point",
+                        "bbox",
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "scroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": "Required only by `action=key`.",
+                    "type": "array",
+                },
+                "text": {
+                    "description": "Required only by `action=type`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The [x, y] coordinates for mouse actions/point, or [x_min, y_min, x_max, y_max] for bbox.",
+                    "type": "array",
+                },
+                "pixels": {"description": "The amount of scrolling.", "type": "number"},
+                "time": {"description": "The seconds to wait.", "type": "number"},
+                "status": {
+                    "description": "The status of the task.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V0 = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V0))

cua_lite/data/interfaces/qwen3_vl/prompts/version_1/desktop_v1.py

@@ -0,0 +1,238 @@
+import copy
+import json
+
+# from .common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
+from cua_lite.data.interfaces.qwen3_vl.prompts.common import TRAJECTORY_SYSTEM_PROMPT_TEMPLATE, GROUNDING_SYSTEM_PROMPT_TEMPLATE
+
+DESKTOP_TRAJECTORY_DESCRIPTION = """Use a mouse and keyboard to interact with a computer, and take screenshots.
+* This is an interface to a desktop GUI. You do not have access to a terminal or applications menu. You must click on desktop icons to start applications.
+* Some applications may take time to start or process actions, so you may need to wait and take successive screenshots to see the results of your actions. For example, if you click on Firefox and a window does not open, try waiting and then taking another screenshot.
+* The screen resolution is 1000x1000.
+* Whenever you intend to move the cursor to click on an element such as an icon, consult a screenshot first to determine the element's coordinates before moving the cursor.
+* If you tried clicking on a program or link but it failed to load even after waiting, adjust your cursor position so that the tip of the cursor visually falls on the element you want to click.
+* Make sure to click any buttons, links, icons, or other elements with the cursor tip in the center of the element. Do not click on edges unless explicitly instructed."""
+
+# mention point and bbox
+DESKTOP_GROUNDING_DESCRIPTION = """Given a computer screenshot, map the natural language instruction to an action or map the natural language reference to a point or a bounding box  in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_ACTIONS = """Given a computer screenshot, map the natural language action to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_POINT = """Given a computer screenshot, map the natural language reference to a point in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+DESKTOP_GROUNDING_DESCRIPTION_BBOX = """Given a computer screenshot, map the natural language reference to a bounding box in the screenshot. This is an interface to a desktop GUI. The screen resolution is 1000x1000."""
+
+# v1
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_use",
+        "name": "desktop_use",
+        "description": DESKTOP_TRAJECTORY_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `key(keys=[\"key1\", \"key2\"])`: Press the keys in order (key down), then release them in reverse order (key up).\n"
+                        + "* `type(text=\"string\")`: Type the given text.\n"
+                        + "* `mouse_move(coordinate=[x, y])`: Move the cursor to (x, y).\n"
+                        + "* `left_click(coordinate=[x, y][, keys=[\"key\"]])`: Left click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `left_click_drag(coordinate=[x, y][, keys=[\"key\"]])`: Drag to (x, y) with the left mouse button. If `keys` is provided, hold those keys while dragging.\n"
+                        + "* `right_click(coordinate=[x, y][, keys=[\"key\"]])`: Right click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `middle_click(coordinate=[x, y][, keys=[\"key\"]])`: Middle click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `double_click(coordinate=[x, y][, keys=[\"key\"]])`: Double left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `triple_click(coordinate=[x, y][, keys=[\"key\"]])`: Triple left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `scroll(pixels=number[, keys=[\"key\"]])`: Scroll vertically. `pixels > 0` scrolls up, `pixels < 0` scrolls down. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `hscroll(pixels=number[, keys=[\"key\"]])`: Scroll horizontally. `pixels > 0` scrolls right, `pixels < 0` scrolls left. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `wait(time=seconds)`: Wait for the given number of seconds.\n"
+                        + "* `terminate(status=\"success\"|\"failure\")`: End the task and report the status.\n"
+                    ),
+                    "enum": [
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "triple_click",
+                        "scroll",
+                        "hscroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": (
+                        "Keys to press/hold.\n"
+                        "Required only for action: `key` (press in order, release in reverse).\n"
+                        "Optional only for actions: `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`, `scroll`, `hscroll` (hold these keys while performing the action).\n"
+                    ),
+                    "type": "array",
+                },
+                "text": {
+                    "description": "The text content to type or the answer to return. Required only for actions: `type`, `answer`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The (x, y) pixel coordinate on the screen. Required only for actions: `mouse_move`, `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`.",
+                    "type": "array",
+                },
+                "pixels": {
+                    "description": "The amount of scrolling to perform. Required only for actions: `scroll`, `hscroll`.",
+                    "type": "number"
+                },
+                "time": {
+                    "description": "The duration to wait in seconds. Required only for action: `wait`.",
+                    "type": "number"
+                },
+                "status": {
+                    "description": "The completion status of the task. Required only for action: `terminate`.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_TRAJECTORY_SYSTEM_PROMPT_V1 = TRAJECTORY_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1))
+
+# v1: grounding
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1_ACTIONS = copy.deepcopy(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1)
+DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1_ACTIONS["function"]["description"] = DESKTOP_GROUNDING_DESCRIPTION_ACTIONS
+
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1_ACTIONS = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_TRAJECTORY_TOOLS_DEFINITION_V1))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_POINT = {
+    "type": "function",
+    "function": {
+        "name_for_human": "point",
+        "name": "point",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_POINT,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 2-element list [x, y] representing the screen coordinates for the center of the described object.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1_POINT = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_POINT))
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_BBOX = {
+    "type": "function",
+    "function": {
+        "name_for_human": "bbox",
+        "name": "bbox",
+        "description": DESKTOP_GROUNDING_DESCRIPTION_BBOX,
+        "parameters": {
+            "properties": {
+                "coordinate": {
+                    "description": "A 4-element list [x_min, y_min, x_max, y_max] defining the bounding box, where (x_min, y_min) is the top-left corner and (x_max, y_max) is the bottom-right corner.",
+                    "type": "array",
+                },
+            },
+            "required": ["coordinate"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1_BBOX = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V1_BBOX))
+
+
+DESKTOP_GROUNDING_TOOLS_DEFINITION_V1 = {
+    "type": "function",
+    "function": {
+        "name_for_human": "desktop_grounding",
+        "name": "desktop_grounding",
+        "description": DESKTOP_GROUNDING_DESCRIPTION,
+        "parameters": {
+            "properties": {
+                "action": {
+                    "description": (
+                        "* `point`: Return the [x, y] center coordinates of the object described.\n"
+                        + "* `bbox`: Return the [x_min, y_min, x_max, y_max] bounding box coordinates of the object described.\n"
+                        + "* `key(keys=[\"key1\", \"key2\"])`: Press the keys in order (key down), then release them in reverse order (key up).\n"
+                        + "* `type(text=\"string\")`: Type the given text.\n"
+                        + "* `mouse_move(coordinate=[x, y])`: Move the cursor to (x, y).\n"
+                        + "* `left_click(coordinate=[x, y][, keys=[\"key\"]])`: Left click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `left_click_drag(coordinate=[x, y][, keys=[\"key\"]])`: Drag to (x, y) with the left mouse button. If `keys` is provided, hold those keys while dragging.\n"
+                        + "* `right_click(coordinate=[x, y][, keys=[\"key\"]])`: Right click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `middle_click(coordinate=[x, y][, keys=[\"key\"]])`: Middle click at (x, y). If `keys` is provided, hold those keys while clicking.\n"
+                        + "* `double_click(coordinate=[x, y][, keys=[\"key\"]])`: Double left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `triple_click(coordinate=[x, y][, keys=[\"key\"]])`: Triple left click at (x, y). If `keys` is provided, hold those keys while double-clicking.\n"
+                        + "* `scroll(pixels=number[, keys=[\"key\"]])`: Scroll vertically. `pixels > 0` scrolls up, `pixels < 0` scrolls down. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `hscroll(pixels=number[, keys=[\"key\"]])`: Scroll horizontally. `pixels > 0` scrolls right, `pixels < 0` scrolls left. If `keys` is provided, hold those keys while scrolling.\n"
+                        + "* `wait(time=seconds)`: Wait for the given number of seconds.\n"
+                        + "* `terminate(status=\"success\"|\"failure\")`: End the task and report the status.\n"
+                        + "* `answer(text=\"answer\")`: Return a final answer in text."
+                    ),
+                    "enum": [
+                        "point",
+                        "bbox",
+                        "key",
+                        "type",
+                        "mouse_move",
+                        "left_click",
+                        "left_click_drag",
+                        "right_click",
+                        "middle_click",
+                        "double_click",
+                        "triple_click",
+                        "scroll",
+                        "hscroll",
+                        "wait",
+                        "terminate",
+                    ],
+                    "type": "string",
+                },
+                "keys": {
+                    "description": (
+                        "Keys to press/hold.\n"
+                        "Required only for action: `key` (press in order, release in reverse).\n"
+                        "Optional only for actions: `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`, `scroll`, `hscroll` (hold these keys while performing the action).\n"
+                    ),
+                    "type": "array",
+                },
+                "text": {
+                    "description": "The text content to type or the answer to return. Required only for actions: `type`, `answer`.",
+                    "type": "string",
+                },
+                "coordinate": {
+                    "description": "The [x, y] coordinates for mouse actions/point, or [x_min, y_min, x_max, y_max] for bbox. Required only for actions: `point`, `bbox`, `mouse_move`, `left_click`, `left_click_drag`, `right_click`, `middle_click`, `double_click`, `triple_click`.",
+                    "type": "array",
+                },
+                "pixels": {
+                    "description": "The amount of scrolling to perform. Required only for actions: `scroll`, `hscroll`.",
+                    "type": "number"
+                },
+                "time": {
+                    "description": "The duration to wait in seconds. Required only for action: `wait`.",
+                    "type": "number"
+                },
+                "status": {
+                    "description": "The completion status of the task. Required only for action: `terminate`.",
+                    "type": "string",
+                    "enum": ["success", "failure"],
+                },
+            },
+            "required": ["action"],
+            "type": "object",
+        },
+        "args_format": "Format the arguments as a JSON object.",
+    },
+}
+DESKTOP_GROUNDING_SYSTEM_PROMPT_V1 = GROUNDING_SYSTEM_PROMPT_TEMPLATE.format(tools_definition=json.dumps(DESKTOP_GROUNDING_TOOLS_DEFINITION_V1))
+
+
+
+if __name__ == "__main__":
+    pass
+

cua_lite/data/preproc/opencua/opencua.py

@@ -65,16 +65,16 @@ class ScriptArguments:
 # -----------------------------
 
 
-def _make_computer_use_tool_call(arguments: Dict[str, Any]) -> Dict[str, Any]:
-    """Wrap a computer_use tool call in Qwen3-VL tool_call structure."""
+def _make_desktop_use_tool_call(arguments: Dict[str, Any]) -> Dict[str, Any]:
+    """Wrap a desktop_use tool call in Qwen3-VL tool_call structure."""
     if "action" not in arguments:
         raise ValueError(
-            f"computer_use arguments must include 'action'. Got: {arguments}"
+            f"desktop_use arguments must include 'action'. Got: {arguments}"
         )
     return {
         "type": "function",
         "function": {
-            "name": "computer_use",
+            "name": "desktop_use",
             "arguments": arguments,
         },
     }
@@ -153,7 +153,7 @@ def _norm01_to_0_1000(x: float, y: float) -> List[int]:
 
 
 def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
-    """Convert AgentNet pyautogui/computer code string into Qwen3-VL computer_use tool calls."""
+    """Convert AgentNet pyautogui/computer code string into Qwen3-VL desktop_use tool calls."""
     if not isinstance(code, str) or not code.strip():
         raise AgentNetCodeParseError(f"Expected non-empty code string. Got: {code!r}")
 
@@ -177,7 +177,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
         if fname == "pyautogui.click":
             x, y = _extract_xy(call)
             tool_calls.append(
-                _make_computer_use_tool_call(
+                _make_desktop_use_tool_call(
                     {"action": "left_click", "coordinate": _norm01_to_0_1000(x, y)}
                 )
             )
@@ -186,7 +186,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
         if fname == "pyautogui.rightClick":
             x, y = _extract_xy(call)
             tool_calls.append(
-                _make_computer_use_tool_call(
+                _make_desktop_use_tool_call(
                     {"action": "right_click", "coordinate": _norm01_to_0_1000(x, y)}
                 )
             )
@@ -195,7 +195,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
         if fname == "pyautogui.middleClick":
             x, y = _extract_xy(call)
             tool_calls.append(
-                _make_computer_use_tool_call(
+                _make_desktop_use_tool_call(
                     {"action": "middle_click", "coordinate": _norm01_to_0_1000(x, y)}
                 )
             )
@@ -204,7 +204,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
         if fname == "pyautogui.doubleClick":
             x, y = _extract_xy(call)
             tool_calls.append(
-                _make_computer_use_tool_call(
+                _make_desktop_use_tool_call(
                     {"action": "double_click", "coordinate": _norm01_to_0_1000(x, y)}
                 )
             )
@@ -213,7 +213,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
         if fname in {"pyautogui.tripleClick", "computer.tripleClick"}:
             x, y = _extract_xy(call)
             tool_calls.append(
-                _make_computer_use_tool_call(
+                _make_desktop_use_tool_call(
                     {"action": "triple_click", "coordinate": _norm01_to_0_1000(x, y)}
                 )
             )
@@ -223,7 +223,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
         if fname == "pyautogui.moveTo":
             x, y = _extract_xy(call)
             tool_calls.append(
-                _make_computer_use_tool_call(
+                _make_desktop_use_tool_call(
                     {"action": "mouse_move", "coordinate": _norm01_to_0_1000(x, y)}
                 )
             )
@@ -238,7 +238,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
                 )
             x, y = _extract_xy(call)
             tool_calls.append(
-                _make_computer_use_tool_call(
+                _make_desktop_use_tool_call(
                     {"action": "left_click_drag", "coordinate": _norm01_to_0_1000(x, y)}
                 )
             )
@@ -253,11 +253,11 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
             pixels = int(_literal_eval(call.args[0]))
             if fname == "pyautogui.scroll":
                 tool_calls.append(
-                    _make_computer_use_tool_call({"action": "scroll", "pixels": pixels})
+                    _make_desktop_use_tool_call({"action": "scroll", "pixels": pixels})
                 )
             else:
                 tool_calls.append(
-                    _make_computer_use_tool_call({"action": "hscroll", "pixels": pixels})
+                    _make_desktop_use_tool_call({"action": "hscroll", "pixels": pixels})
                 )
             continue
 
@@ -274,7 +274,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
                 )
             keys = [str(k).lower() for k in keys_val]
             tool_calls.append(
-                _make_computer_use_tool_call({"action": "key", "keys": keys})
+                _make_desktop_use_tool_call({"action": "key", "keys": keys})
             )
             continue
 
@@ -287,13 +287,13 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
             if isinstance(key_val, (list, tuple)):
                 for k in key_val:
                     tool_calls.append(
-                        _make_computer_use_tool_call(
+                        _make_desktop_use_tool_call(
                             {"action": "key", "keys": [str(k).lower()]}
                         )
                     )
             else:
                 tool_calls.append(
-                    _make_computer_use_tool_call(
+                    _make_desktop_use_tool_call(
                         {"action": "key", "keys": [str(key_val).lower()]}
                     )
                 )
@@ -309,7 +309,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
                 )
             text = str(_literal_eval(msg_node))
             tool_calls.append(
-                _make_computer_use_tool_call({"action": "type", "text": text})
+                _make_desktop_use_tool_call({"action": "type", "text": text})
             )
             continue
 
@@ -324,7 +324,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
                     f"Unsupported wait signature.\ncode=\n{code}"
                 )
             tool_calls.append(
-                _make_computer_use_tool_call({"action": "wait", "time": t})
+                _make_desktop_use_tool_call({"action": "wait", "time": t})
             )
             continue
 
@@ -340,7 +340,7 @@ def agentnet_code_to_qwen_tool_calls(code: str) -> List[Dict[str, Any]]:
                     f"Unsupported terminate status={status!r}.\ncode=\n{code}"
                 )
             tool_calls.append(
-                _make_computer_use_tool_call({"action": "terminate", "status": status})
+                _make_desktop_use_tool_call({"action": "terminate", "status": status})
             )
             continue
 
@@ -713,7 +713,7 @@ messages = [
         "content": [{"type": "text", "text": "{traj[0].value.action}"}],
         # Convert raw `{traj[0].value.code}` (pyauto) -> Qwen3-VL action space
         "tool_calls": [
-            {"type": "function", "function": {"name": "computer_use", "arguments": "..."}}
+            {"type": "function", "function": {"name": "desktop_use", "arguments": "..."}}
         ],
     },
 
@@ -732,7 +732,7 @@ messages = [
         "content": [{"type": "text", "text": "{traj[1].value.action}"}],
         # Convert raw `{traj[1].value.code}` (pyauto) -> Qwen3-VL action space
         "tool_calls": [
-            {"type": "function", "function": {"name": "computer_use", "arguments": "..."}}
+            {"type": "function", "function": {"name": "desktop_use", "arguments": "..."}}
         ],
     },
 
@@ -750,7 +750,7 @@ messages = [
         "content": [{"type": "text", "text": "{traj[n].value.action}"}],
         # Convert raw `{traj[n].value.code}` (pyauto) -> Qwen3-VL action space
         "tool_calls": [
-            {"type": "function", "function": {"name": "computer_use", "arguments": "..."}}
+            {"type": "function", "function": {"name": "desktop_use", "arguments": "..."}}
         ],
     },
 ]
@@ -840,7 +840,7 @@ tools_def = {
 
 # AgentNet → Qwen3-VL Action Mapping
 This file documents how to convert AgentNet GUI actions (`traj[n].value.code`)
-into Qwen3-VL `computer_use` tool calls.
+into Qwen3-VL `desktop_use` tool calls.
 
 Coordinates rule:
 - AgentNet uses normalized floats `[0, 1]`
@@ -858,7 +858,7 @@ pyautogui.click(x=0.018, y=0.508)
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "left_click",
         "coordinate": [18, 508]
@@ -877,7 +877,7 @@ pyautogui.rightClick(x=0.304, y=0.684)
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "right_click",
         "coordinate": [304, 684]
@@ -896,7 +896,7 @@ pyautogui.doubleClick(x=0.153, y=0.283)
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "double_click",
         "coordinate": [153, 283]
@@ -915,7 +915,7 @@ computer.tripleClick(x=0.226, y=0.311)
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "triple_click",
         "coordinate": [226, 311]
@@ -934,7 +934,7 @@ pyautogui.moveTo(x=0.394, y=0.604)
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "mouse_move",
         "coordinate": [394, 604]
@@ -955,7 +955,7 @@ pyautogui.dragTo(x=0.085, y=0.361, button="left")
     {
       "type": "function",
       "function": {
-        "name": "computer_use",
+        "name": "desktop_use",
         "arguments": {
           "action": "mouse_move",
           "coordinate": [78, 381]
@@ -965,7 +965,7 @@ pyautogui.dragTo(x=0.085, y=0.361, button="left")
     {
       "type": "function",
       "function": {
-        "name": "computer_use",
+        "name": "desktop_use",
         "arguments": {
           "action": "left_click_drag",
           "coordinate": [85, 361]
@@ -985,7 +985,7 @@ pyautogui.scroll(-8)
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "scroll",
         "pixels": -8
@@ -1004,7 +1004,7 @@ pyautogui.hotkey(["ctrl", "o"])
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "key",
         "keys": ["ctrl", "o"]
@@ -1023,7 +1023,7 @@ pyautogui.press("enter")
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "key",
         "keys": ["enter"]
@@ -1042,7 +1042,7 @@ pyautogui.write(message="cd ./Pictures")
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "type",
         "text": "cd ./Pictures"
@@ -1061,7 +1061,7 @@ computer.wait()
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "wait",
         "time": 1.0
@@ -1080,7 +1080,7 @@ computer.terminate(status="success")
   "tool_calls": [{
     "type": "function",
     "function": {
-      "name": "computer_use",
+      "name": "desktop_use",
       "arguments": {
         "action": "terminate",
         "status": "success"

cua_lite/data/utils.py

@@ -1,5 +1,5 @@
 from collections import defaultdict
-from typing import Callable, Any
+from typing import Any, Callable, Mapping, Iterable
 
 
 def clean_nones(item):
@@ -14,15 +14,8 @@ def clean_nones(item):
         return item
 
 
-# # 2. 定义 Transform 函数（适配 Dataset 的 batch 格式）
-# def transform_batch(batch):
-#     # batch 是一个字典，例如 {'messages': [ [...], [...] ]}
-#     # 我们需要对里面的每一个样本进行清洗
-#     return {k: [clean_nones(item) for item in v] for k, v in batch.items()}
-
-
 def batch_proc(
-    func: Callable[[dict[str, Any]], dict[str, Any]], batch: dict[str, list[Any]]
+    func: Callable[[dict[str, Any]], dict[str, Any]], batch: dict[str, list[Any]], **kwargs
 ) -> dict[str, list[Any]]:
     """
     Core reusable logic:
@@ -47,10 +40,58 @@ def batch_proc(
 
         # 3. Call the processing function
         #    Expected to return a dict, e.g., {'messages': ..., 'status': ..., 'meta': ...}
-        processed_row = func(row)
+        processed_row = func(row, **kwargs)
 
         # 4. Aggregate ALL keys from the result
         for key, value in processed_row.items():
             output_batch[key].append(value)
 
     return dict(output_batch)
+
+
+# def batch_proc(
+#     func: Callable[[dict[str, Any]], Mapping[str, Any]],
+#     batch: dict[str, list[Any]],
+#     *,
+#     flatten_list_values: bool = False,
+#     validate_batch: bool = True,
+#     **kwargs,
+# ) -> dict[str, list[Any]]:
+#     """
+#     Columnar batch -> row-wise func -> columnar aggregation.
+
+#     Assumes func(row, **kwargs) returns a dict[str, Any | list[Any]].
+#     """
+
+#     if not batch:
+#         return {}
+
+#     first_key = next(iter(batch))
+#     batch_size = len(batch[first_key])
+
+#     if validate_batch:
+#         for k, col in batch.items():
+#             if len(col) != batch_size:
+#                 raise ValueError(
+#                     f"Column length mismatch: '{first_key}' has {batch_size}, "
+#                     f"but '{k}' has {len(col)}."
+#                 )
+
+#     output_batch = defaultdict(list)
+
+#     for i in range(batch_size):
+#         row = {key: batch[key][i] for key in batch}
+
+#         processed_row = func(row, **kwargs)
+#         if not isinstance(processed_row, Mapping):
+#             raise TypeError(
+#                 f"func must return Mapping[str, Any], got {type(processed_row).__name__}"
+#             )
+
+#         for key, value in processed_row.items():
+#             if flatten_list_values and isinstance(value, (list, tuple)):
+#                 output_batch[key].extend(value)
+#             else:
+#                 output_batch[key].append(value)
+
+#     return dict(output_batch)