rpa_vision_v3/agent_v0/window_info_crossplatform.py

# window_info_crossplatform.py
"""
Récupération des informations sur la fenêtre active - CROSS-PLATFORM

Supporte:
- Linux (X11 via xdotool)
- Windows (via pywin32)
- macOS (via pyobjc)

Installation des dépendances:
    pip install pywin32  # Windows
    pip install pyobjc-framework-Cocoa  # macOS
    pip install psutil  # Tous OS
"""

from __future__ import annotations

import platform
import subprocess
from typing import Dict, Optional


def _run_cmd(cmd: list[str]) -> Optional[str]:
    """Exécute une commande et renvoie la sortie texte (strippée), ou None en cas d'erreur."""
    try:
        out = subprocess.check_output(cmd, stderr=subprocess.DEVNULL)
        return out.decode("utf-8", errors="ignore").strip()
    except Exception:
        return None


def get_active_window_info() -> Dict[str, str]:
    """
    Renvoie un dict :
    {
      "title": "...",
      "app_name": "..."
    }
    
    Détecte automatiquement l'OS et utilise la méthode appropriée.
    """
    system = platform.system()
    
    if system == "Linux":
        return _get_window_info_linux()
    elif system == "Windows":
        return _get_window_info_windows()
    elif system == "Darwin":  # macOS
        return _get_window_info_macos()
    else:
        return {"title": "unknown_window", "app_name": "unknown_app"}


def _get_window_info_linux() -> Dict[str, str]:
    """
    Linux: utilise xdotool (X11)
    
    Nécessite: sudo apt-get install xdotool
    """
    title = _run_cmd(["xdotool", "getactivewindow", "getwindowname"])
    pid_str = _run_cmd(["xdotool", "getactivewindow", "getwindowpid"])

    app_name: Optional[str] = None
    if pid_str:
        pid_str = pid_str.strip()
        # On récupère le nom du binaire via ps
        app_name = _run_cmd(["ps", "-p", pid_str, "-o", "comm="])

    if not title:
        title = "unknown_window"
    if not app_name:
        app_name = "unknown_app"

    return {
        "title": title,
        "app_name": app_name,
    }


def _get_window_info_windows() -> Dict[str, str]:
    """
    Windows: utilise pywin32 + psutil
    
    Nécessite: pip install pywin32 psutil
    """
    try:
        import win32gui
        import win32process
        import psutil
        
        # Fenêtre au premier plan
        hwnd = win32gui.GetForegroundWindow()
        
        # Titre de la fenêtre
        title = win32gui.GetWindowText(hwnd)
        if not title:
            title = "unknown_window"
        
        # PID du processus
        _, pid = win32process.GetWindowThreadProcessId(hwnd)
        
        # Nom du processus
        try:
            process = psutil.Process(pid)
            app_name = process.name()
        except (psutil.NoSuchProcess, psutil.AccessDenied):
            app_name = "unknown_app"
        
        return {
            "title": title,
            "app_name": app_name,
        }
    
    except ImportError:
        # pywin32 ou psutil non installé
        return {
            "title": "unknown_window (pywin32 missing)",
            "app_name": "unknown_app (pywin32 missing)",
        }
    except Exception as e:
        return {
            "title": f"error: {e}",
            "app_name": "unknown_app",
        }


def _get_window_info_macos() -> Dict[str, str]:
    """
    macOS: utilise pyobjc (AppKit)
    
    Nécessite: pip install pyobjc-framework-Cocoa
    
    Note: Nécessite les permissions "Accessibility" dans System Preferences
    """
    try:
        from AppKit import NSWorkspace
        from Quartz import (
            CGWindowListCopyWindowInfo,
            kCGWindowListOptionOnScreenOnly,
            kCGNullWindowID
        )
        
        # Application active
        active_app = NSWorkspace.sharedWorkspace().activeApplication()
        app_name = active_app.get('NSApplicationName', 'unknown_app')
        
        # Titre de la fenêtre (via Quartz)
        # On cherche la fenêtre de l'app active qui est au premier plan
        window_list = CGWindowListCopyWindowInfo(
            kCGWindowListOptionOnScreenOnly,
            kCGNullWindowID
        )
        
        title = "unknown_window"
        for window in window_list:
            owner_name = window.get('kCGWindowOwnerName', '')
            if owner_name == app_name:
                window_title = window.get('kCGWindowName', '')
                if window_title:
                    title = window_title
                    break
        
        return {
            "title": title,
            "app_name": app_name,
        }
    
    except ImportError:
        # pyobjc non installé
        return {
            "title": "unknown_window (pyobjc missing)",
            "app_name": "unknown_app (pyobjc missing)",
        }
    except Exception as e:
        return {
            "title": f"error: {e}",
            "app_name": "unknown_app",
        }


# Test rapide
if __name__ == "__main__":
    import time
    
    print(f"OS détecté: {platform.system()}")
    print("\nTest de capture fenêtre active (5 secondes)...")
    print("Changez de fenêtre pour tester!\n")
    
    for i in range(5):
        info = get_active_window_info()
        print(f"[{i+1}] App: {info['app_name']:20s} | Title: {info['title']}")
        time.sleep(1)