fleet-python 0.2.17__tar.gz → 0.2.18__tar.gz

{fleet_python-0.2.17 → fleet_python-0.2.18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fleet-python
-Version: 0.2.17
+Version: 0.2.18
 Summary: Python SDK for Fleet environments
 Author-email: Fleet AI <nic@fleet.so>
 License: Apache-2.0

{fleet_python-0.2.17 → fleet_python-0.2.18}/fleet/__init__.py

@@ -37,9 +37,6 @@ from .verifiers import (
     TASK_SUCCESSFUL_SCORE,
 )
 
-# Import Playwright wrapper
-from .playwright import FleetPlaywrightWrapper
-
 # Import async verifiers (default verifier is async for modern usage)
 from ._async.verifiers import (
     verifier,
@@ -75,8 +72,6 @@ __all__ = [
     "FleetAPIError", 
     "FleetTimeoutError",
     "FleetConfigurationError",
-    # Playwright wrapper
-    "FleetPlaywrightWrapper",
     # Verifiers (async is default)
     "verifier",
     "verifier_sync", 

{fleet_python-0.2.17 → fleet_python-0.2.18}/fleet/client.py

@@ -537,12 +537,6 @@ class Fleet:
         browser = new_env.browser()
         browser.start(width=snapshot.viewport_size[0], height=snapshot.viewport_size[1])
 
-        from fleet.playwright import FleetPlaywrightWrapper
-
-        playwright_wrapper = FleetPlaywrightWrapper(
-            cdp_url=browser.cdp_url(), instance_client=new_env.instance
-        )
-
         # Replay tool logs in order
         validation_errors = []
         last_timestamp = None
@@ -559,7 +553,7 @@ class Fleet:
 
                 # Replay the tool action
                 _replay_tool_action(
-                    playwright_wrapper,
+                    None,
                     tool_log,
                     new_env.instance._client,
                     replay_session_id,
@@ -588,7 +582,7 @@ class Fleet:
 
         if validate:
             validation = _validate_resumed_state(
-                new_env, snapshot, playwright_wrapper, validation_errors
+                new_env, snapshot, None, validation_errors
             )
 
         return new_env, validation
@@ -664,7 +658,7 @@ def _execute_verifier_remote(
 
 
 def _replay_tool_action(
-    playwright_wrapper: "FleetPlaywrightWrapper",
+    playwright_wrapper,
     tool_log: ToolLogEntry,
     client: "SyncWrapper",
     session_id: str,
@@ -760,7 +754,7 @@ def _replay_tool_action(
 def _validate_resumed_state(
     new_env: Environment,
     snapshot: EnvironmentSnapshot,
-    playwright_wrapper: "FleetPlaywrightWrapper",
+    playwright_wrapper,
     existing_errors: List[str],
 ) -> SnapshotValidation:
     """Validate that the resumed state matches the snapshot."""

fleet_python-0.2.18/fleet/verifiers/code.py

@@ -0,0 +1,2 @@
+TASK_SUCCESSFUL_SCORE = 1
+TASK_FAILED_SCORE = 0

{fleet_python-0.2.17 → fleet_python-0.2.18}/fleet_python.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fleet-python
-Version: 0.2.17
+Version: 0.2.18
 Summary: Python SDK for Fleet environments
 Author-email: Fleet AI <nic@fleet.so>
 License: Apache-2.0

{fleet_python-0.2.17 → fleet_python-0.2.18}/fleet_python.egg-info/SOURCES.txt

@@ -24,14 +24,12 @@ fleet/client.py
 fleet/config.py
 fleet/exceptions.py
 fleet/models.py
-fleet/playwright.py
 fleet/tasks.py
 fleet/types.py
 fleet/_async/__init__.py
 fleet/_async/base.py
 fleet/_async/client.py
 fleet/_async/exceptions.py
-fleet/_async/playwright.py
 fleet/_async/tasks.py
 fleet/_async/env/__init__.py
 fleet/_async/env/client.py

{fleet_python-0.2.17 → fleet_python-0.2.18}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "fleet-python"
-version = "0.2.17"
+version = "0.2.18"
 description = "Python SDK for Fleet environments"
 authors = [
     {name = "Fleet AI", email = "nic@fleet.so"},

fleet_python-0.2.17/fleet/_async/playwright.py

@@ -1,291 +0,0 @@
-import base64
-from typing import List, Dict, Any
-from playwright.async_api import async_playwright, Browser, Page
-from .client import AsyncEnv
-
-
-# Key mapping for computer use actions
-CUA_KEY_TO_PLAYWRIGHT_KEY = {
-    "/": "Divide",
-    "\\": "Backslash",
-    "alt": "Alt",
-    "arrowdown": "ArrowDown",
-    "arrowleft": "ArrowLeft",
-    "arrowright": "ArrowRight",
-    "arrowup": "ArrowUp",
-    "backspace": "Backspace",
-    "capslock": "CapsLock",
-    "cmd": "Meta",
-    "ctrl": "Control",
-    "delete": "Delete",
-    "end": "End",
-    "enter": "Enter",
-    "esc": "Escape",
-    "home": "Home",
-    "insert": "Insert",
-    "option": "Alt",
-    "pagedown": "PageDown",
-    "pageup": "PageUp",
-    "shift": "Shift",
-    "space": " ",
-    "super": "Meta",
-    "tab": "Tab",
-    "win": "Meta",
-}
-
-
-class AsyncFleetPlaywrightWrapper:
-    """
-    A wrapper that adds Playwright browser automation to Fleet environment instances.
-
-    This class handles:
-    - Browser connection via CDP
-    - Computer actions (click, scroll, type, etc.)
-    - Screenshot capture
-    - Integration with OpenAI computer use API
-
-    Usage:
-        instance = await fleet.env.make(env_key="hubspot", version="v1.2.7")
-        browser = AsyncFleetPlaywrightWrapper(instance)
-        await browser.start()
-
-        # Use browser methods
-        screenshot = await browser.screenshot()
-        tools = [browser.openai_cua_tool]
-
-        # Clean up when done
-        await browser.close()
-    """
-
-    def get_environment(self):
-        return "browser"
-
-    def get_dimensions(self):
-        return (1920, 1080)
-
-    def __init__(
-        self,
-        env: AsyncEnv,
-        display_width: int = 1920,
-        display_height: int = 1080,
-    ):
-        """
-        Initialize the Fleet Playwright wrapper.
-
-        Args:
-            env: Fleet environment instance
-            display_width: Browser viewport width
-            display_height: Browser viewport height
-        """
-        self.env = env
-        self.display_width = display_width
-        self.display_height = display_height
-
-        self._playwright = None
-        self._browser: Browser | None = None
-        self._page: Page | None = None
-        self._started = False
-
-    async def start(self):
-        """Start the browser and establish connection."""
-        if self._started:
-            return
-
-        # Start Playwright
-        self._playwright = await async_playwright().start()
-
-        # Start browser on the Fleet instance
-        print("Starting browser...")
-        await self.env.browser().start()
-        cdp = await self.env.browser().describe()
-
-        # Connect to browser
-        self._browser = await self._playwright.chromium.connect_over_cdp(
-            cdp.cdp_browser_url
-        )
-        self._page = self._browser.contexts[0].pages[0]
-        await self._page.set_viewport_size(
-            {"width": self.display_width, "height": self.display_height}
-        )
-
-        self._started = True
-        print(f"Track agent: {cdp.cdp_devtools_url}")
-
-    async def close(self):
-        """Close the browser connection."""
-        if self._playwright:
-            await self._playwright.stop()
-            self._playwright = None
-            self._browser = None
-            self._page = None
-            self._started = False
-
-    def _ensure_started(self):
-        """Ensure browser is started before operations."""
-        if not self._started:
-            raise RuntimeError("Browser not started. Call await browser.start() first.")
-
-    @property
-    def openai_cua_tool(self) -> Dict[str, Any]:
-        """
-        Tool definition for OpenAI computer use API.
-
-        Returns:
-            Tool definition dict for use with OpenAI responses API
-        """
-        return {
-            "type": "computer_use_preview",
-            "display_width": self.display_width,
-            "display_height": self.display_height,
-            "environment": "browser",
-        }
-
-    async def screenshot(self) -> str:
-        """
-        Take a screenshot and return base64 encoded string.
-
-        Returns:
-            Base64 encoded PNG screenshot
-        """
-        self._ensure_started()
-
-        png_bytes = await self._page.screenshot(full_page=False)
-        return base64.b64encode(png_bytes).decode("utf-8")
-
-    def get_current_url(self) -> str:
-        """Get the current page URL."""
-        self._ensure_started()
-        return self._page.url
-
-    async def execute_computer_action(self, action: Dict[str, Any]) -> Dict[str, Any]:
-        """
-        Execute a computer action and return the result for OpenAI API.
-
-        Args:
-            action: Computer action dict from OpenAI response
-
-        Returns:
-            Result dict for computer_call_output
-        """
-        self._ensure_started()
-
-        action_type = action["type"]
-        action_args = {k: v for k, v in action.items() if k != "type"}
-
-        print(f"Executing: {action_type}({action_args})")
-
-        # Execute the action
-        if hasattr(self, f"_{action_type}"):
-            method = getattr(self, f"_{action_type}")
-            await method(**action_args)
-        else:
-            raise ValueError(f"Unsupported action type: {action_type}")
-
-        # Take screenshot after action
-        screenshot_base64 = await self.screenshot()
-
-        return {
-            "type": "input_image",
-            "image_url": f"data:image/png;base64,{screenshot_base64}",
-            "current_url": self.get_current_url(),
-        }
-
-    # Computer action implementations
-    async def _click(self, x: int, y: int, button: str = "left") -> None:
-        """Click at coordinates."""
-        self._ensure_started()
-        await self._page.mouse.click(x, y, button=button)
-
-    async def _double_click(self, x: int, y: int) -> None:
-        """Double-click at coordinates."""
-        self._ensure_started()
-        await self._page.mouse.dblclick(x, y)
-
-    async def _scroll(self, x: int, y: int, scroll_x: int, scroll_y: int) -> None:
-        """Scroll from coordinates."""
-        self._ensure_started()
-        await self._page.mouse.move(x, y)
-        await self._page.evaluate(f"window.scrollBy({scroll_x}, {scroll_y})")
-
-    async def _type(self, text: str) -> None:
-        """Type text."""
-        self._ensure_started()
-        await self._page.keyboard.type(text)
-
-    async def _keypress(self, keys: List[str]) -> None:
-        """Press key combination."""
-        self._ensure_started()
-        mapped_keys = [CUA_KEY_TO_PLAYWRIGHT_KEY.get(key.lower(), key) for key in keys]
-        for key in mapped_keys:
-            await self._page.keyboard.down(key)
-        for key in reversed(mapped_keys):
-            await self._page.keyboard.up(key)
-
-    async def _move(self, x: int, y: int) -> None:
-        """Move mouse to coordinates."""
-        self._ensure_started()
-        await self._page.mouse.move(x, y)
-
-    async def _drag(self, path: List[Dict[str, int]]) -> None:
-        """Drag mouse along path."""
-        self._ensure_started()
-        if not path:
-            return
-        await self._page.mouse.move(path[0]["x"], path[0]["y"])
-        await self._page.mouse.down()
-        for point in path[1:]:
-            await self._page.mouse.move(point["x"], point["y"])
-        await self._page.mouse.up()
-
-    async def _wait(self, ms: int = 1000) -> None:
-        """Wait for specified milliseconds."""
-        import asyncio
-
-        await asyncio.sleep(ms / 1000)
-
-    # Browser-specific actions
-    async def _goto(self, url: str) -> None:
-        """Navigate to URL."""
-        self._ensure_started()
-        try:
-            await self._page.goto(url)
-        except Exception as e:
-            print(f"Error navigating to {url}: {e}")
-
-    async def _back(self) -> None:
-        """Go back in browser history."""
-        self._ensure_started()
-        await self._page.go_back()
-
-    async def _forward(self) -> None:
-        """Go forward in browser history."""
-        self._ensure_started()
-        await self._page.go_forward()
-
-    async def _refresh(self) -> None:
-        """Refresh the page."""
-        self._ensure_started()
-        await self._page.reload()
-
-    # ------------------------------------------------------------------
-    # Public aliases (no leading underscore) expected by the Agent &
-    # OpenAI computer-use API. They forward directly to the underscored
-    # implementations above so the external interface matches the older
-    # BasePlaywrightComputer class.
-    # ------------------------------------------------------------------
-
-    # Mouse / keyboard actions
-    click = _click
-    double_click = _double_click
-    scroll = _scroll
-    type = _type  # noqa: A003 – shadowing built-in for API compatibility
-    keypress = _keypress
-    move = _move
-    drag = _drag
-    wait = _wait
-
-    # Browser navigation actions
-    goto = _goto
-    back = _back
-    forward = _forward
-    refresh = _refresh

fleet_python-0.2.17/fleet/playwright.py

@@ -1,290 +0,0 @@
-import base64
-from typing import List, Dict, Any
-from playwright.sync_api import sync_playwright, Browser, Page
-from .client import Environment
-
-
-# Key mapping for computer use actions
-CUA_KEY_TO_PLAYWRIGHT_KEY = {
-    "/": "Divide",
-    "\\": "Backslash",
-    "alt": "Alt",
-    "arrowdown": "ArrowDown",
-    "arrowleft": "ArrowLeft",
-    "arrowright": "ArrowRight",
-    "arrowup": "ArrowUp",
-    "backspace": "Backspace",
-    "capslock": "CapsLock",
-    "cmd": "Meta",
-    "ctrl": "Control",
-    "delete": "Delete",
-    "end": "End",
-    "enter": "Enter",
-    "esc": "Escape",
-    "home": "Home",
-    "insert": "Insert",
-    "option": "Alt",
-    "pagedown": "PageDown",
-    "pageup": "PageUp",
-    "shift": "Shift",
-    "space": " ",
-    "super": "Meta",
-    "tab": "Tab",
-    "win": "Meta",
-}
-
-
-class FleetPlaywrightWrapper:
-    """
-    A wrapper that adds Playwright browser automation to Fleet environment instances.
-
-    This class handles:
-    - Browser connection via CDP
-    - Computer actions (click, scroll, type, etc.)
-    - Screenshot capture
-    - Integration with OpenAI computer use API
-
-    Usage:
-        instance = fleet.env.make(env_key="hubspot", version="v1.2.7")
-        browser = FleetPlaywrightWrapper(instance)
-        browser.start()
-
-        # Use browser methods
-        screenshot = browser.screenshot()
-        tools = [browser.openai_cua_tool]
-
-        # Clean up when done
-        browser.close()
-    """
-
-    def get_environment(self):
-        return "browser"
-
-    def get_dimensions(self):
-        return (1920, 1080)
-
-    def __init__(
-        self,
-        env: Environment,
-        display_width: int = 1920,
-        display_height: int = 1080,
-    ):
-        """
-        Initialize the Fleet Playwright wrapper.
-
-        Args:
-            env: Fleet environment instance
-            display_width: Browser viewport width
-            display_height: Browser viewport height
-        """
-        self.env = env
-        self.display_width = display_width
-        self.display_height = display_height
-
-        self._playwright = None
-        self._browser: Browser | None = None
-        self._page: Page | None = None
-        self._started = False
-
-    def start(self):
-        """Start the browser and establish connection."""
-        if self._started:
-            return
-
-        # Start Playwright
-        self._playwright = sync_playwright().start()
-
-        # Start browser on the Fleet instance
-        print("Starting browser...")
-        self.env.browser().start()
-        cdp = self.env.browser().describe()
-
-        # Connect to browser
-        self._browser = self._playwright.chromium.connect_over_cdp(
-            cdp.cdp_browser_url
-        )
-        self._page = self._browser.contexts[0].pages[0]
-        self._page.set_viewport_size(
-            {"width": self.display_width, "height": self.display_height}
-        )
-
-        self._started = True
-        print(f"Track agent: {cdp.cdp_devtools_url}")
-
-    def close(self):
-        """Close the browser connection."""
-        if self._playwright:
-            self._playwright.stop()
-            self._playwright = None
-            self._browser = None
-            self._page = None
-            self._started = False
-
-    def _ensure_started(self):
-        """Ensure browser is started before operations."""
-        if not self._started:
-            raise RuntimeError("Browser not started. Call browser.start() first.")
-
-    @property
-    def openai_cua_tool(self) -> Dict[str, Any]:
-        """
-        Tool definition for OpenAI computer use API.
-
-        Returns:
-            Tool definition dict for use with OpenAI responses API
-        """
-        return {
-            "type": "computer_use_preview",
-            "display_width": self.display_width,
-            "display_height": self.display_height,
-            "environment": "browser",
-        }
-
-    def screenshot(self) -> str:
-        """
-        Take a screenshot and return base64 encoded string.
-
-        Returns:
-            Base64 encoded PNG screenshot
-        """
-        self._ensure_started()
-
-        png_bytes = self._page.screenshot(full_page=False)
-        return base64.b64encode(png_bytes).decode("utf-8")
-
-    def get_current_url(self) -> str:
-        """Get the current page URL."""
-        self._ensure_started()
-        return self._page.url
-
-    def execute_computer_action(self, action: Dict[str, Any]) -> Dict[str, Any]:
-        """
-        Execute a computer action and return the result for OpenAI API.
-
-        Args:
-            action: Computer action dict from OpenAI response
-
-        Returns:
-            Result dict for computer_call_output
-        """
-        self._ensure_started()
-
-        action_type = action["type"]
-        action_args = {k: v for k, v in action.items() if k != "type"}
-
-        print(f"Executing: {action_type}({action_args})")
-
-        # Execute the action
-        if hasattr(self, f"_{action_type}"):
-            method = getattr(self, f"_{action_type}")
-            method(**action_args)
-        else:
-            raise ValueError(f"Unsupported action type: {action_type}")
-
-        # Take screenshot after action
-        screenshot_base64 = self.screenshot()
-
-        return {
-            "type": "input_image",
-            "image_url": f"data:image/png;base64,{screenshot_base64}",
-            "current_url": self.get_current_url(),
-        }
-
-    # Computer action implementations
-    def _click(self, x: int, y: int, button: str = "left") -> None:
-        """Click at coordinates."""
-        self._ensure_started()
-        self._page.mouse.click(x, y, button=button)
-
-    def _double_click(self, x: int, y: int) -> None:
-        """Double-click at coordinates."""
-        self._ensure_started()
-        self._page.mouse.dblclick(x, y)
-
-    def _scroll(self, x: int, y: int, scroll_x: int, scroll_y: int) -> None:
-        """Scroll from coordinates."""
-        self._ensure_started()
-        self._page.mouse.move(x, y)
-        self._page.evaluate(f"window.scrollBy({scroll_x}, {scroll_y})")
-
-    def _type(self, text: str) -> None:
-        """Type text."""
-        self._ensure_started()
-        self._page.keyboard.type(text)
-
-    def _keypress(self, keys: List[str]) -> None:
-        """Press key combination."""
-        self._ensure_started()
-        mapped_keys = [CUA_KEY_TO_PLAYWRIGHT_KEY.get(key.lower(), key) for key in keys]
-        for key in mapped_keys:
-            self._page.keyboard.down(key)
-        for key in reversed(mapped_keys):
-            self._page.keyboard.up(key)
-
-    def _move(self, x: int, y: int) -> None:
-        """Move mouse to coordinates."""
-        self._ensure_started()
-        self._page.mouse.move(x, y)
-
-    def _drag(self, path: List[Dict[str, int]]) -> None:
-        """Drag mouse along path."""
-        self._ensure_started()
-        if not path:
-            return
-        self._page.mouse.move(path[0]["x"], path[0]["y"])
-        self._page.mouse.down()
-        for point in path[1:]:
-            self._page.mouse.move(point["x"], point["y"])
-        self._page.mouse.up()
-
-    def _wait(self, ms: int = 1000) -> None:
-        """Wait for specified milliseconds."""
-
-        time.sleep(ms / 1000)
-
-    # Browser-specific actions
-    def _goto(self, url: str) -> None:
-        """Navigate to URL."""
-        self._ensure_started()
-        try:
-            self._page.goto(url)
-        except Exception as e:
-            print(f"Error navigating to {url}: {e}")
-
-    def _back(self) -> None:
-        """Go back in browser history."""
-        self._ensure_started()
-        self._page.go_back()
-
-    def _forward(self) -> None:
-        """Go forward in browser history."""
-        self._ensure_started()
-        self._page.go_forward()
-
-    def _refresh(self) -> None:
-        """Refresh the page."""
-        self._ensure_started()
-        self._page.reload()
-
-    # ------------------------------------------------------------------
-    # Public aliases (no leading underscore) expected by the Agent &
-    # OpenAI computer-use API. They forward directly to the underscored
-    # implementations above so the external interface matches the older
-    # BasePlaywrightComputer class.
-    # ------------------------------------------------------------------
-
-    # Mouse / keyboard actions
-    click = _click
-    double_click = _double_click
-    scroll = _scroll
-    type = _type  # noqa: A003 – shadowing built-in for API compatibility
-    keypress = _keypress
-    move = _move
-    drag = _drag
-    wait = _wait
-
-    # Browser navigation actions
-    goto = _goto
-    back = _back
-    forward = _forward
-    refresh = _refresh

fleet_python-0.2.17/fleet/verifiers/code.py

@@ -1 +0,0 @@
-TASK_SUCCESSFUL_SCORE = 1