#!/usr/bin/env python3
"""
Session Cleaner -- Outil leger de nettoyage de sessions avant replay.

Petit serveur Flask standalone qui permet de :
- Lister les sessions enregistrees recentes
- Visualiser chaque session avec ses screenshots (crop + full)
- Marquer les clics parasites a supprimer (auto-detection des toasts,
  clics droit, fenetres Lea/systray, derniers 3 evenements)
- Re-construire un replay nettoye et l'injecter dans la queue

Option A du rapport audit VWB.
Port : 5006
"""

import json
import logging
import os
import uuid
from datetime import datetime
from pathlib import Path
from typing import Any, Dict, List, Optional, Tuple

from flask import (
    Flask,
    redirect,
    render_template_string,
    request,
    send_from_directory,
    url_for,
)

# ---------------------------------------------------------------------------
# Configuration
# ---------------------------------------------------------------------------

STREAMING_SERVER = os.environ.get("RPA_STREAMING_SERVER", "http://localhost:5005")
LIVE_SESSIONS_DIR = os.environ.get(
    "RPA_LIVE_SESSIONS_DIR",
    os.path.join(os.path.dirname(__file__), "..", "data", "training", "live_sessions"),
)
PORT = int(os.environ.get("SESSION_CLEANER_PORT", "5006"))

# Charger le token API depuis l'environnement ou .env.local
API_TOKEN = os.environ.get("RPA_API_TOKEN", "")
if not API_TOKEN:
    env_local = os.path.join(os.path.dirname(__file__), "..", ".env.local")
    if os.path.isfile(env_local):
        try:
            with open(env_local, encoding="utf-8") as f:
                for line in f:
                    line = line.strip()
                    if line.startswith("RPA_API_TOKEN="):
                        API_TOKEN = line.split("=", 1)[1].strip().strip('"').strip("'")
                        break
        except OSError:
            pass

# ---------------------------------------------------------------------------
# Import optionnel de build_replay_from_raw_events
# ---------------------------------------------------------------------------

_build_replay_fn = None
try:
    from agent_v0.server_v1.stream_processor import build_replay_from_raw_events
    _build_replay_fn = build_replay_from_raw_events
except ImportError:
    pass

# ---------------------------------------------------------------------------
# Logging
# ---------------------------------------------------------------------------

logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(levelname)s] %(message)s")
logger = logging.getLogger("session_cleaner")

# ---------------------------------------------------------------------------
# Application Flask
# ---------------------------------------------------------------------------

app = Flask(__name__)


# ---------------------------------------------------------------------------
# Utilitaires
# ---------------------------------------------------------------------------

# Fenetres considerees comme parasites
_PARASITIC_WINDOW_PATTERNS = [
    "program manager",
    "fenetre de depassement",
    "fenêtre de dépassement",
    "léa",
    "lea",
    "assistant",
    "activer windows",
]

# Types d'evenements exploitables (affiches a l'utilisateur)
_ACTIONABLE_TYPES = frozenset({"mouse_click", "text_input", "key_combo", "key_press", "type"})


def _resolve_sessions_dir() -> Path:
    """Resoudre le repertoire racine des live_sessions."""
    return Path(LIVE_SESSIONS_DIR).resolve()


def _discover_sessions(limit: int = 50) -> List[Dict[str, Any]]:
    """Decouvrir les sessions recentes.

    Parcourt deux niveaux :
    - <base>/<machine_id>/sess_* (format actuel)
    - <base>/sess_* (ancien format, sessions au niveau racine)
    """
    base = _resolve_sessions_dir()
    if not base.is_dir():
        logger.warning("Repertoire live_sessions introuvable : %s", base)
        return []

    sessions: List[Dict[str, Any]] = []

    for item in base.iterdir():
        if not item.is_dir():
            continue

        # Sessions directement a la racine (ancien format)
        if item.name.startswith("sess_"):
            jsonl = item / "live_events.jsonl"
            if jsonl.is_file():
                sessions.append(_build_session_info("(racine)", item.name, item, jsonl))
            continue

        # Ignorer les dossiers systeme
        if item.name.startswith(".") or item.name in ("embeddings", "streaming_sessions", "workflows", "test_gpu"):
            continue

        # Sous-dossiers machine_id
        for sub in item.iterdir():
            if sub.is_dir() and sub.name.startswith("sess_"):
                jsonl = sub / "live_events.jsonl"
                if jsonl.is_file():
                    sessions.append(_build_session_info(item.name, sub.name, sub, jsonl))

    # Tri par date decroissante (mtime du JSONL)
    sessions.sort(key=lambda s: s["mtime"], reverse=True)
    return sessions[:limit]


def _build_session_info(machine_id: str, session_id: str, session_dir: Path, jsonl_path: Path) -> Dict[str, Any]:
    """Construire les metadonnees d'une session."""
    mtime = jsonl_path.stat().st_mtime
    event_count = 0
    try:
        with open(jsonl_path, encoding="utf-8") as f:
            for line in f:
                if line.strip():
                    event_count += 1
    except OSError:
        pass

    # Extraire la date depuis le nom de session (sess_YYYYMMDDTHHMMSS_...)
    date_str = ""
    try:
        parts = session_id.split("_")
        if len(parts) >= 2:
            raw = parts[1]  # 20260410T222352
            dt = datetime.strptime(raw, "%Y%m%dT%H%M%S")
            date_str = dt.strftime("%d/%m/%Y %H:%M:%S")
    except (ValueError, IndexError):
        date_str = datetime.fromtimestamp(mtime).strftime("%d/%m/%Y %H:%M:%S")

    return {
        "machine_id": machine_id,
        "session_id": session_id,
        "session_dir": str(session_dir),
        "date_str": date_str,
        "event_count": event_count,
        "mtime": mtime,
    }


def _load_events(session_dir: Path) -> List[Dict[str, Any]]:
    """Charger les evenements depuis live_events.jsonl."""
    jsonl = session_dir / "live_events.jsonl"
    events: List[Dict[str, Any]] = []
    if not jsonl.is_file():
        return events
    try:
        with open(jsonl, encoding="utf-8") as f:
            for line in f:
                line = line.strip()
                if line:
                    try:
                        events.append(json.loads(line))
                    except json.JSONDecodeError:
                        continue
    except OSError as e:
        logger.error("Erreur lecture %s : %s", jsonl, e)
    return events


def _get_window_title(event: Dict[str, Any]) -> str:
    """Extraire le titre de fenetre d'un evenement.

    Les evenements plus recents stockent la fenetre dans event.window.title,
    les anciens dans event.active_window_title.
    """
    inner = event.get("event", {})
    # Format actuel : inner.window.title
    window = inner.get("window") or {}
    if isinstance(window, dict) and window.get("title"):
        return window["title"]
    # Ancien format
    return inner.get("active_window_title", "")


def _get_shot_filename(click_index: int, session_dir: Path) -> Optional[str]:
    """Trouver le fichier screenshot pour un clic donne.

    Essaie dans l'ordre :
    1. shot_XXXX_crop.png (ancien format)
    2. shot_XXXX_full.png (ancien format)
    3. res_shot_XXXX.png (format recent — resultat post-action)

    ``click_index`` est 1-based (premier clic = 1).
    """
    shots_dir = session_dir / "shots"
    if not shots_dir.is_dir():
        return None

    shot_id = f"shot_{click_index:04d}"

    # Priorite au crop (plus informatif en thumbnail)
    for pattern in [f"{shot_id}_crop.png", f"{shot_id}_full.png", f"res_{shot_id}.png"]:
        if (shots_dir / pattern).is_file():
            return pattern

    return None


def _is_parasitic(event: Dict[str, Any], index: int, total: int) -> bool:
    """Determiner si un evenement est probablement parasite.

    Criteres :
    - Fenetre contenant un pattern parasite (systray, Program Manager, Lea, etc.)
    - Clic droit
    - Types non-exploitables (heartbeat, focus_change, action_result)
    - Parmi les 3 derniers evenements (souvent = arret enregistrement)
    """
    inner = event.get("event", {})
    etype = inner.get("type", "")

    # Types toujours parasites
    if etype in ("heartbeat", "focus_change", "window_focus_change", "action_result",
                 "screenshot", "status", "ping", "pong"):
        return True

    # Clics droit
    if etype == "mouse_click" and inner.get("button") == "right":
        return True

    # Fenetre parasite
    win_title = _get_window_title(event).lower()
    if win_title:
        for pattern in _PARASITIC_WINDOW_PATTERNS:
            if pattern in win_title:
                return True

    # Derniers 3 evenements exploitables de la session
    # (on les marque UNIQUEMENT si c'est un evenement exploitable, pas un heartbeat)
    if etype in _ACTIONABLE_TYPES and index >= total - 3:
        return True

    return False


def _parse_actions(events: List[Dict[str, Any]], session_dir: Path) -> List[Dict[str, Any]]:
    """Convertir les evenements bruts en liste d'actions affichables.

    Retourne une liste de dicts avec : index_global, type, position, fenetre,
    texte, touches, shot_file, is_parasitic, etc.
    """
    actions: List[Dict[str, Any]] = []
    click_count = 0
    total_events = len(events)

    # Pre-calculer les 3 derniers indices d'evenements exploitables
    actionable_indices = [
        i for i, ev in enumerate(events)
        if ev.get("event", {}).get("type", "") in _ACTIONABLE_TYPES
    ]
    last_3_actionable = set(actionable_indices[-3:]) if len(actionable_indices) >= 3 else set(actionable_indices)

    for i, event in enumerate(events):
        inner = event.get("event", {})
        etype = inner.get("type", "")

        # Ne montrer que les evenements exploitables
        if etype not in _ACTIONABLE_TYPES:
            continue

        action: Dict[str, Any] = {
            "global_index": i,
            "type": etype,
            "position": "",
            "window_title": _get_window_title(event),
            "text": "",
            "keys": "",
            "shot_file": None,
            "is_parasitic": False,
        }

        # Position (pour les clics)
        pos = inner.get("pos")
        if pos and isinstance(pos, (list, tuple)) and len(pos) >= 2:
            action["position"] = f"({pos[0]}, {pos[1]})"

        # Bouton de clic
        if etype == "mouse_click":
            action["button"] = inner.get("button", "left")
            click_count += 1
            action["shot_file"] = _get_shot_filename(click_count, session_dir)
            action["click_number"] = click_count

        # Texte tape
        if etype in ("text_input", "type"):
            action["text"] = inner.get("text", "")

        # Touches pour key_combo / key_press
        if etype in ("key_combo", "key_press"):
            keys = inner.get("keys", [])
            if isinstance(keys, list):
                action["keys"] = " + ".join(str(k) for k in keys)
            else:
                action["keys"] = str(inner.get("key", keys))

        # Detection parasite
        # Utiliser les 3 derniers indices exploitables (pas les indices globaux)
        parasitic = False
        inner_type = etype

        # Clic droit
        if inner_type == "mouse_click" and inner.get("button") == "right":
            parasitic = True

        # Fenetre parasite
        win_lower = action["window_title"].lower()
        if win_lower:
            for pattern in _PARASITIC_WINDOW_PATTERNS:
                if pattern in win_lower:
                    parasitic = True
                    break

        # Derniers 3 evenements exploitables
        if i in last_3_actionable:
            parasitic = True

        action["is_parasitic"] = parasitic
        actions.append(action)

    return actions


# ---------------------------------------------------------------------------
# Templates HTML
# ---------------------------------------------------------------------------

_BASE_CSS = """
body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
       margin: 0; padding: 20px; background: #f5f5f5; color: #333; }
h1 { color: #2c3e50; border-bottom: 2px solid #3498db; padding-bottom: 10px; }
h2 { color: #34495e; }
a { color: #2980b9; text-decoration: none; }
a:hover { text-decoration: underline; }
table { border-collapse: collapse; width: 100%; background: white; border-radius: 6px;
        overflow: hidden; box-shadow: 0 1px 3px rgba(0,0,0,0.12); }
th { background: #2c3e50; color: white; padding: 12px 15px; text-align: left; }
td { padding: 10px 15px; border-bottom: 1px solid #eee; }
tr:hover { background: #f0f7ff; }
.btn { display: inline-block; padding: 10px 20px; background: #e74c3c; color: white;
       border: none; border-radius: 4px; cursor: pointer; font-size: 14px; }
.btn:hover { background: #c0392b; }
.btn-secondary { background: #3498db; }
.btn-secondary:hover { background: #2980b9; }
.info-box { background: #eaf4ff; border: 1px solid #b8d4f0; border-radius: 6px;
            padding: 15px; margin: 15px 0; }
.warning-box { background: #fff3cd; border: 1px solid #ffc107; border-radius: 6px;
               padding: 15px; margin: 15px 0; }
.success-box { background: #d4edda; border: 1px solid #28a745; border-radius: 6px;
               padding: 15px; margin: 15px 0; }
.error-box { background: #f8d7da; border: 1px solid #dc3545; border-radius: 6px;
             padding: 15px; margin: 15px 0; }
.parasitic { background: #ffe0e0; }
.normal { background: #e0ffe0; }
.counter { font-size: 18px; font-weight: bold; margin: 15px 0; }
.counter .remove { color: #e74c3c; }
.counter .total { color: #2c3e50; }
img.thumb { max-height: 80px; border: 1px solid #ccc; border-radius: 4px; cursor: pointer; }
img.thumb:hover { box-shadow: 0 2px 8px rgba(0,0,0,0.3); }
.nav { margin-bottom: 20px; }
.mono { font-family: 'Fira Code', 'Consolas', monospace; font-size: 13px; }
label { cursor: pointer; }
"""

_INDEX_TEMPLATE = """<!DOCTYPE html>
<html lang="fr">
<head>
    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1">
    <title>Session Cleaner -- Lea</title>
    <style>{{ css }}</style>
</head>
<body>
    <h1>Session Cleaner</h1>
    <p>Outil de nettoyage des sessions avant replay. Selectionnez une session pour voir ses actions.</p>

    {% if sessions %}
    <table>
        <thead>
            <tr>
                <th>Date</th>
                <th>Machine</th>
                <th>Session ID</th>
                <th>Evenements</th>
                <th>Action</th>
            </tr>
        </thead>
        <tbody>
        {% for s in sessions %}
            <tr>
                <td>{{ s.date_str }}</td>
                <td class="mono">{{ s.machine_id }}</td>
                <td class="mono">{{ s.session_id }}</td>
                <td>{{ s.event_count }}</td>
                <td><a href="{{ url_for('view_session', machine_id=s.machine_id, session_id=s.session_id) }}"
                       class="btn btn-secondary">Voir</a></td>
            </tr>
        {% endfor %}
        </tbody>
    </table>
    {% else %}
    <div class="warning-box">
        <p>Aucune session trouvee dans <code>{{ sessions_dir }}</code>.</p>
        <p>Lancez un enregistrement depuis l'Agent V1 pour creer des sessions.</p>
    </div>
    {% endif %}
</body>
</html>"""

_SESSION_TEMPLATE = """<!DOCTYPE html>
<html lang="fr">
<head>
    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1">
    <title>Session {{ session_id }} -- Session Cleaner</title>
    <style>{{ css }}</style>
    <style>
        .overlay { display: none; position: fixed; top: 0; left: 0; width: 100%; height: 100%;
                   background: rgba(0,0,0,0.8); z-index: 1000; justify-content: center;
                   align-items: center; }
        .overlay.active { display: flex; }
        .overlay img { max-width: 90%; max-height: 90%; border-radius: 8px; }
    </style>
</head>
<body>
    <div class="nav">
        <a href="{{ url_for('index') }}">&larr; Retour a la liste</a>
    </div>

    <h1>Session : <span class="mono">{{ session_id }}</span></h1>
    <div class="info-box">
        <strong>Machine :</strong> {{ machine_id }} |
        <strong>Date :</strong> {{ date_str }} |
        <strong>Evenements bruts :</strong> {{ total_events }}
    </div>

    {% if actions %}
    <div class="counter" id="counter">
        <span class="remove" id="remove-count">{{ parasitic_count }}</span> actions a supprimer /
        <span class="total">{{ actions|length }}</span> total
    </div>

    <form method="POST" action="{{ url_for('clean_and_replay') }}" id="clean-form">
        <input type="hidden" name="session_id" value="{{ session_id }}">
        <input type="hidden" name="machine_id" value="{{ machine_id }}">

        <table>
            <thead>
                <tr>
                    <th>Supprimer</th>
                    <th>#</th>
                    <th>Type</th>
                    <th>Position</th>
                    <th>Fenetre</th>
                    <th>Texte / Touches</th>
                    <th>Screenshot</th>
                </tr>
            </thead>
            <tbody>
            {% for a in actions %}
                <tr class="{{ 'parasitic' if a.is_parasitic else 'normal' }}">
                    <td>
                        <label>
                            <input type="checkbox" name="remove_indices"
                                   value="{{ a.global_index }}"
                                   {{ 'checked' if a.is_parasitic else '' }}
                                   onchange="updateCounter()">
                        </label>
                    </td>
                    <td>{{ loop.index }}</td>
                    <td class="mono">
                        {{ a.type }}
                        {% if a.button is defined and a.button == 'right' %}
                            <span style="color:#e74c3c">(droit)</span>
                        {% endif %}
                    </td>
                    <td class="mono">{{ a.position }}</td>
                    <td>{{ a.window_title|truncate(40) }}</td>
                    <td class="mono">
                        {% if a.text %}{{ a.text|truncate(60) }}{% endif %}
                        {% if a.keys %}{{ a.keys }}{% endif %}
                    </td>
                    <td>
                        {% if a.shot_file %}
                            <img src="{{ url_for('serve_shot', machine_id=machine_id, session_id=session_id, filename=a.shot_file) }}"
                                 class="thumb"
                                 alt="Screenshot action {{ loop.index }}"
                                 onclick="showOverlay(this.src)"
                                 loading="lazy">
                        {% else %}
                            <span style="color:#aaa">--</span>
                        {% endif %}
                    </td>
                </tr>
            {% endfor %}
            </tbody>
        </table>

        <div style="margin-top: 20px; display: flex; gap: 10px; align-items: center;">
            <button type="submit" class="btn" id="btn-replay">
                Nettoyer et relancer le replay
            </button>
            <button type="button" class="btn btn-secondary" onclick="selectAll(true)">
                Tout cocher
            </button>
            <button type="button" class="btn btn-secondary" onclick="selectAll(false)">
                Tout decocher
            </button>
            <button type="button" class="btn btn-secondary" onclick="resetParasitic()">
                Reinitialiser (auto-detection)
            </button>
        </div>
    </form>
    {% else %}
    <div class="warning-box">
        Aucune action exploitable dans cette session.
    </div>
    {% endif %}

    <!-- Overlay pour zoom screenshot -->
    <div class="overlay" id="overlay" onclick="this.classList.remove('active')">
        <img id="overlay-img" src="" alt="Screenshot agrandi">
    </div>

    <script>
    // Indices des evenements auto-detectes comme parasites (pour reset)
    var parasiticIndices = {{ parasitic_indices|tojson }};

    function updateCounter() {
        var checked = document.querySelectorAll('input[name="remove_indices"]:checked').length;
        document.getElementById('remove-count').textContent = checked;
        // Mettre a jour les couleurs des lignes
        document.querySelectorAll('input[name="remove_indices"]').forEach(function(cb) {
            var tr = cb.closest('tr');
            tr.className = cb.checked ? 'parasitic' : 'normal';
        });
    }

    function selectAll(state) {
        document.querySelectorAll('input[name="remove_indices"]').forEach(function(cb) {
            cb.checked = state;
        });
        updateCounter();
    }

    function resetParasitic() {
        document.querySelectorAll('input[name="remove_indices"]').forEach(function(cb) {
            cb.checked = parasiticIndices.indexOf(parseInt(cb.value)) !== -1;
        });
        updateCounter();
    }

    function showOverlay(src) {
        document.getElementById('overlay-img').src = src;
        document.getElementById('overlay').classList.add('active');
    }
    </script>
</body>
</html>"""

_RESULT_TEMPLATE = """<!DOCTYPE html>
<html lang="fr">
<head>
    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1">
    <title>Replay lance -- Session Cleaner</title>
    <style>{{ css }}</style>
</head>
<body>
    <div class="nav">
        <a href="{{ url_for('index') }}">&larr; Retour a la liste</a>
    </div>

    <h1>Replay lance</h1>

    {% if success %}
    <div class="success-box">
        <p><strong>Replay demarre avec succes.</strong></p>
        <p><strong>Replay ID :</strong> <code>{{ replay_id }}</code></p>
        <p><strong>Session :</strong> <code>{{ session_id }}</code></p>
        <p><strong>Machine cible :</strong> <code>{{ machine_id }}</code></p>
        <p><strong>Actions injectees :</strong> {{ action_count }}</p>
        <p><strong>Actions supprimees :</strong> {{ removed_count }}</p>
    </div>
    {% else %}
    <div class="error-box">
        <p><strong>Erreur lors du lancement du replay.</strong></p>
        <p>{{ error_message }}</p>
    </div>
    {% endif %}
</body>
</html>"""


# ---------------------------------------------------------------------------
# Routes
# ---------------------------------------------------------------------------

@app.route("/")
def index():
    """Page d'accueil : liste des sessions recentes."""
    sessions = _discover_sessions(limit=50)
    return render_template_string(
        _INDEX_TEMPLATE,
        sessions=sessions,
        sessions_dir=str(_resolve_sessions_dir()),
        css=_BASE_CSS,
    )


@app.route("/session/<machine_id>/<session_id>")
def view_session(machine_id: str, session_id: str):
    """Vue detaillee d'une session avec ses actions."""
    session_dir = _find_session_dir(machine_id, session_id)
    if session_dir is None:
        return render_template_string(
            """<!DOCTYPE html><html lang="fr"><head><meta charset="utf-8">
            <title>Session introuvable</title><style>{{ css }}</style></head>
            <body><div class="error-box"><p>Session <code>{{ sid }}</code>
            introuvable pour la machine <code>{{ mid }}</code>.</p></div>
            <a href="{{ url_for('index') }}">Retour</a></body></html>""",
            sid=session_id, mid=machine_id, css=_BASE_CSS,
        ), 404

    events = _load_events(session_dir)
    actions = _parse_actions(events, session_dir)

    # Compter les parasites et collecter leurs indices globaux
    parasitic_count = sum(1 for a in actions if a["is_parasitic"])
    parasitic_indices = [a["global_index"] for a in actions if a["is_parasitic"]]

    # Date depuis le nom de session
    date_str = ""
    try:
        parts = session_id.split("_")
        if len(parts) >= 2:
            dt = datetime.strptime(parts[1], "%Y%m%dT%H%M%S")
            date_str = dt.strftime("%d/%m/%Y %H:%M:%S")
    except (ValueError, IndexError):
        date_str = "?"

    return render_template_string(
        _SESSION_TEMPLATE,
        session_id=session_id,
        machine_id=machine_id,
        date_str=date_str,
        total_events=len(events),
        actions=actions,
        parasitic_count=parasitic_count,
        parasitic_indices=parasitic_indices,
        css=_BASE_CSS,
    )


@app.route("/shots/<machine_id>/<session_id>/<filename>")
def serve_shot(machine_id: str, session_id: str, filename: str):
    """Servir un fichier screenshot."""
    session_dir = _find_session_dir(machine_id, session_id)
    if session_dir is None:
        return "Session introuvable", 404

    shots_dir = session_dir / "shots"
    if not shots_dir.is_dir():
        return "Repertoire shots introuvable", 404

    # Securite : empecher la traversee de repertoire
    safe_name = Path(filename).name
    if safe_name != filename:
        return "Nom de fichier invalide", 400

    target = shots_dir / safe_name
    if not target.is_file():
        return "Fichier introuvable", 404

    return send_from_directory(str(shots_dir), safe_name, mimetype="image/png")


@app.route("/clean-and-replay", methods=["POST"])
def clean_and_replay():
    """Nettoyer les evenements et lancer un replay."""
    session_id = request.form.get("session_id", "")
    machine_id = request.form.get("machine_id", "")
    remove_indices_raw = request.form.getlist("remove_indices")

    # Convertir les indices en entiers
    remove_indices = set()
    for idx_str in remove_indices_raw:
        try:
            remove_indices.add(int(idx_str))
        except ValueError:
            continue

    # Trouver le repertoire de session
    session_dir = _find_session_dir(machine_id, session_id)
    if session_dir is None:
        return render_template_string(
            _RESULT_TEMPLATE,
            success=False,
            error_message=f"Session {session_id} introuvable pour la machine {machine_id}.",
            replay_id="", session_id=session_id, machine_id=machine_id,
            action_count=0, removed_count=0, css=_BASE_CSS,
        )

    # Charger les evenements et filtrer
    all_events = _load_events(session_dir)
    cleaned_events = [
        ev for i, ev in enumerate(all_events)
        if i not in remove_indices
    ]
    removed_count = len(all_events) - len(cleaned_events)

    logger.info(
        "Nettoyage session %s : %d evenements -> %d (suppression de %d)",
        session_id, len(all_events), len(cleaned_events), removed_count,
    )

    # Construire les actions de replay
    replay_actions = None
    error_message = ""

    if _build_replay_fn is not None:
        # Chemin principal : utiliser build_replay_from_raw_events
        try:
            replay_actions = _build_replay_fn(
                cleaned_events,
                session_id=session_id,
                session_dir=str(session_dir),
            )
            logger.info("build_replay_from_raw_events a produit %d actions", len(replay_actions))
        except Exception as e:
            logger.error("Erreur build_replay_from_raw_events : %s", e)
            error_message = f"Erreur lors de la construction du replay : {e}"

    if not replay_actions:
        # Fallback : filtrage simple et conversion directe.
        # Se declenche si build_replay_from_raw_events a crashe OU
        # retourne une liste vide OU n'est pas disponible.
        try:
            replay_actions = _simple_build_replay(cleaned_events, session_dir)
            logger.info("Fallback simple_build_replay a produit %d actions", len(replay_actions))
            error_message = ""  # le fallback a reussi, on efface l'erreur precedente
        except Exception as e:
            logger.error("Erreur fallback simple_build_replay : %s", e)
            error_message = f"Erreur lors de la construction du replay (fallback) : {e}"

    if not replay_actions:
        if not error_message:
            error_message = "Aucune action exploitable apres nettoyage."
        return render_template_string(
            _RESULT_TEMPLATE,
            success=False, error_message=error_message,
            replay_id="", session_id=session_id, machine_id=machine_id,
            action_count=0, removed_count=removed_count, css=_BASE_CSS,
        )

    # Envoyer au streaming server
    replay_id = f"replay_clean_{uuid.uuid4().hex[:8]}"
    try:
        import requests as _requests

        headers = {"Content-Type": "application/json"}
        if API_TOKEN:
            headers["Authorization"] = f"Bearer {API_TOKEN}"

        payload = {
            "session_id": session_id,
            "actions": replay_actions,
            "machine_id": machine_id if machine_id != "(racine)" else "",
            "task_description": f"Replay nettoye de {session_id} ({removed_count} actions supprimees)",
        }

        resp = _requests.post(
            f"{STREAMING_SERVER}/api/v1/traces/stream/replay/raw",
            json=payload,
            headers=headers,
            timeout=30,
        )

        if resp.status_code == 200:
            data = resp.json()
            replay_id = data.get("replay_id", replay_id)
            logger.info("Replay lance : %s (%d actions)", replay_id, len(replay_actions))
            return render_template_string(
                _RESULT_TEMPLATE,
                success=True, replay_id=replay_id,
                session_id=session_id, machine_id=machine_id,
                action_count=len(replay_actions), removed_count=removed_count,
                error_message="", css=_BASE_CSS,
            )
        else:
            error_message = f"Serveur streaming a repondu {resp.status_code} : {resp.text[:300]}"
            logger.error("Erreur POST replay : %s", error_message)

    except ImportError:
        error_message = (
            "Module 'requests' non disponible. "
            "Installez-le avec : pip install requests"
        )
    except Exception as e:
        error_message = f"Erreur de connexion au serveur streaming ({STREAMING_SERVER}) : {e}"
        logger.error("Erreur connexion streaming : %s", e)

    return render_template_string(
        _RESULT_TEMPLATE,
        success=False, error_message=error_message,
        replay_id="", session_id=session_id, machine_id=machine_id,
        action_count=0, removed_count=removed_count, css=_BASE_CSS,
    )


# ---------------------------------------------------------------------------
# Helpers internes
# ---------------------------------------------------------------------------

def _find_session_dir(machine_id: str, session_id: str) -> Optional[Path]:
    """Trouver le repertoire d'une session.

    Cherche dans :
    1. <base>/<machine_id>/<session_id>/
    2. <base>/<session_id>/ (ancien format, racine)
    """
    base = _resolve_sessions_dir()

    # Sous machine_id
    if machine_id and machine_id != "(racine)":
        candidate = base / machine_id / session_id
        if candidate.is_dir():
            return candidate

    # Directement a la racine
    candidate = base / session_id
    if candidate.is_dir():
        return candidate

    # Recherche exhaustive (au cas ou le machine_id a change)
    for item in base.iterdir():
        if item.is_dir() and not item.name.startswith("."):
            candidate = item / session_id
            if candidate.is_dir():
                return candidate

    return None


def _simple_build_replay(events: List[Dict[str, Any]], session_dir: Path) -> List[Dict[str, Any]]:
    """Construire un replay simplifie sans dependre de stream_processor.

    Convertit les evenements bruts en actions normalisees simples :
    - mouse_click -> action click (coordonnees en pixels)
    - text_input / type -> action type
    - key_combo / key_press -> action key_combo

    C'est un fallback pour quand build_replay_from_raw_events n'est pas disponible.
    Les coordonnees ne sont PAS converties en pourcentages (le serveur les accepte
    aussi en pixels).
    """
    actions: List[Dict[str, Any]] = []
    click_count = 0

    # Essayer d'extraire la resolution d'ecran
    screen_w, screen_h = 1920, 1080
    for ev in events:
        inner = ev.get("event", {})
        meta = inner.get("screen_metadata", {})
        res = meta.get("screen_resolution")
        if res and isinstance(res, (list, tuple)) and len(res) >= 2:
            screen_w, screen_h = int(res[0]), int(res[1])
            break

    for ev in events:
        inner = ev.get("event", {})
        etype = inner.get("type", "")

        if etype not in _ACTIONABLE_TYPES:
            continue

        action_id = f"act_clean_{uuid.uuid4().hex[:6]}"

        if etype == "mouse_click":
            pos = inner.get("pos", [0, 0])
            click_count += 1

            action = {
                "action_id": action_id,
                "type": "click",
                "x_pct": round(pos[0] / screen_w, 6) if screen_w else 0.0,
                "y_pct": round(pos[1] / screen_h, 6) if screen_h else 0.0,
                "button": inner.get("button", "left"),
                "visual_mode": False,  # pas d'enrichissement → coords brutes
                "wait_before": 0.5,
            }
            actions.append(action)

        elif etype in ("text_input", "type"):
            text = inner.get("text", "")
            if text:
                action = {
                    "action_id": action_id,
                    "type": "type",
                    "text": text,
                    "wait_before": 0.3,
                }
                actions.append(action)

        elif etype in ("key_combo", "key_press"):
            keys = inner.get("keys", [])
            if isinstance(keys, str):
                keys = [keys]
            key_single = inner.get("key", "")
            if not keys and key_single:
                keys = [key_single]
            if keys:
                action = {
                    "action_id": action_id,
                    "type": "key_combo",
                    "keys": keys,
                    "wait_before": 0.3,
                }
                actions.append(action)

    return actions


# ---------------------------------------------------------------------------
# Point d'entree
# ---------------------------------------------------------------------------

def main():
    """Demarrer le serveur Session Cleaner."""
    import argparse

    parser = argparse.ArgumentParser(
        description="Session Cleaner -- Nettoyage de sessions avant replay",
    )
    parser.add_argument(
        "--port", type=int, default=PORT,
        help=f"Port du serveur (defaut: {PORT})",
    )
    parser.add_argument(
        "--host", default="0.0.0.0",
        help="Adresse d'ecoute (defaut: 0.0.0.0)",
    )
    parser.add_argument(
        "--debug", action="store_true",
        help="Mode debug Flask",
    )
    args = parser.parse_args()

    logger.info("Session Cleaner demarre sur http://%s:%d", args.host, args.port)
    logger.info("Repertoire sessions : %s", _resolve_sessions_dir())
    logger.info("Serveur streaming : %s", STREAMING_SERVER)
    logger.info("Token API : %s", "configure" if API_TOKEN else "non configure")

    app.run(host=args.host, port=args.port, debug=args.debug)


if __name__ == "__main__":
    main()