browser-agent-server 1.0.0__py3-none-any.whl

browser_agent/__init__.py

@@ -0,0 +1,128 @@
+"""System-wide Browser Agent SDK: control real headful Chrome + X11 mouse from any Python project."""
+from __future__ import annotations
+
+import base64
+import json
+import os
+import urllib.error
+import urllib.request
+from pathlib import Path
+
+from .chrome import RealChromeSession, is_cloudflare_challenge, is_warp_proxy_ready
+from .x11_input import X11Display, bezier_curve
+
+DEFAULT_AGENT_URL = os.environ.get("BROWSER_AGENT_URL", "http://127.0.0.1:8765")
+
+
+class BrowserClient:
+    """Zero-dependency HTTP client for the local Browser Agent daemon (127.0.0.1:8765)."""
+
+    def __init__(self, base_url: str = DEFAULT_AGENT_URL, timeout: float = 75.0):
+        self.base_url = base_url.rstrip("/")
+        self.timeout = timeout
+
+    def _request(self, method: str, path: str, payload: dict | None = None, timeout: float | None = None) -> dict:
+        url = f"{self.base_url}{path}"
+        data = json.dumps(payload).encode("utf-8") if payload is not None else None
+        headers = {"Content-Type": "application/json"} if data is not None else {}
+        req = urllib.request.Request(url, data=data, headers=headers, method=method)
+        with urllib.request.urlopen(req, timeout=timeout or self.timeout) as resp:
+            return json.loads(resp.read().decode("utf-8"))
+
+    def is_available(self) -> bool:
+        try:
+            res = self._request("GET", "/health", timeout=1.2)
+            return bool(res.get("ok"))
+        except Exception:  # noqa: BLE001
+            return False
+
+    def status(self) -> dict:
+        return self._request("GET", "/status", timeout=3.0)
+
+    def fetch(
+        self,
+        url: str,
+        wait_sec: float = 3.5,
+        solve_cloudflare: bool = True,
+        use_proxy: bool | None = None,
+        include_screenshot: bool = False,
+    ) -> dict:
+        return self._request(
+            "POST",
+            "/fetch",
+            {
+                "url": url,
+                "wait_sec": wait_sec,
+                "solve_cloudflare": solve_cloudflare,
+                "use_proxy": use_proxy,
+                "include_screenshot": include_screenshot,
+            },
+        )
+
+    def goto(self, url: str, wait_sec: float = 2.5, solve_cloudflare: bool = True) -> dict:
+        return self._request(
+            "POST",
+            "/action",
+            {"action": "goto", "url": url, "wait_sec": wait_sec, "solve_cloudflare": solve_cloudflare},
+        )
+
+    def move_mouse(self, x: int, y: int) -> dict:
+        return self._request("POST", "/action", {"action": "move", "x": int(x), "y": int(y)})
+
+    def click(self, x: int | None = None, y: int | None = None, button: int = 1, double: bool = False) -> dict:
+        return self._request(
+            "POST",
+            "/action",
+            {"action": "click", "x": x, "y": y, "button": button, "double": double},
+        )
+
+    def type_text(self, text: str, submit: bool = False) -> dict:
+        return self._request("POST", "/action", {"action": "type", "text": text, "submit": submit})
+
+    def press_key(self, key: str) -> dict:
+        return self._request("POST", "/action", {"action": "key", "key": key})
+
+    def scroll(self, clicks: int = 4, direction: str = "down", x: int | None = None, y: int | None = None) -> dict:
+        return self._request(
+            "POST",
+            "/action",
+            {"action": "scroll", "clicks": clicks, "direction": direction, "x": x, "y": y},
+        )
+
+    def screenshot(self, save_path: str | Path | None = None) -> bytes:
+        res = self._request("POST", "/action", {"action": "screenshot"})
+        raw = base64.b64decode(res.get("screenshot_b64") or "")
+        if save_path and raw:
+            Path(save_path).write_bytes(raw)
+        return raw
+
+    def run_agent(self, url: str, goal: str, max_steps: int = 8, use_proxy: bool | None = None) -> dict:
+        return self._request(
+            "POST",
+            "/agent",
+            {"url": url, "goal": goal, "max_steps": max_steps, "use_proxy": use_proxy},
+            timeout=180.0,
+        )
+
+    def stop(self) -> dict:
+        return self._request("POST", "/action", {"action": "stop"}, timeout=10.0)
+
+
+__all__ = [
+    "BrowserClient",
+    "RealChromeSession",
+    "X11Display",
+    "bezier_curve",
+    "is_cloudflare_challenge",
+    "is_warp_proxy_ready",
+]
+
+from importlib.metadata import PackageNotFoundError  # noqa: E402
+from importlib.metadata import version as _v  # noqa: E402
+
+try:
+    __version__ = _v("browser-agent-server")
+except PackageNotFoundError:
+    __version__ = "0.0.0+local"
+
+__all__ = [*__all__, "__version__"]