""" Homelab Automation Dashboard - Backend Optimisé API REST moderne avec FastAPI pour la gestion d'homelab """ from datetime import datetime, timezone, timedelta from pathlib import Path from time import perf_counter, time import os import re import shutil import subprocess import sqlite3 import yaml from abc import ABC, abstractmethod from typing import Literal, Any, List, Dict, Optional from threading import Lock import asyncio import json import uuid # APScheduler imports from apscheduler.schedulers.asyncio import AsyncIOScheduler from apscheduler.triggers.cron import CronTrigger from apscheduler.triggers.date import DateTrigger from apscheduler.jobstores.memory import MemoryJobStore from apscheduler.executors.asyncio import AsyncIOExecutor from croniter import croniter import pytz from fastapi import FastAPI, HTTPException, Depends, Request, Form, WebSocket, WebSocketDisconnect from fastapi.responses import HTMLResponse, JSONResponse, FileResponse from fastapi.security import APIKeyHeader from fastapi.templating import Jinja2Templates from fastapi.middleware.cors import CORSMiddleware from fastapi.staticfiles import StaticFiles from pydantic import BaseModel, Field, field_validator, ConfigDict from sqlalchemy.ext.asyncio import AsyncSession import uvicorn # Import DB layer (async SQLAlchemy) from models.database import get_db # type: ignore from crud.host import HostRepository # type: ignore from crud.bootstrap_status import BootstrapStatusRepository # type: ignore from crud.log import LogRepository # type: ignore from crud.task import TaskRepository # type: ignore from crud.schedule import ScheduleRepository # type: ignore from crud.schedule_run import ScheduleRunRepository # type: ignore from models.database import init_db # type: ignore BASE_DIR = Path(__file__).resolve().parent # Configuration avancée de l'application app = FastAPI( title="Homelab Automation Dashboard API", version="1.0.0", description="API REST moderne pour la gestion automatique d'homelab", docs_url="/api/docs", redoc_url="/api/redoc" ) # Middleware CORS pour le développement app.add_middleware( CORSMiddleware, allow_origins=["*"], # À restreindre en production allow_credentials=True, allow_methods=["*"], allow_headers=["*"], ) app.mount("/static", StaticFiles(directory=BASE_DIR, html=False), name="static") # Configuration des chemins et variables d'environnement LOGS_DIR = Path(os.environ.get("LOGS_DIR", "/logs")) ANSIBLE_DIR = BASE_DIR.parent / "ansible" SSH_KEY_PATH = os.environ.get("SSH_KEY_PATH", str(Path.home() / ".ssh" / "id_rsa")) SSH_USER = os.environ.get("SSH_USER", "automation") SSH_REMOTE_USER = os.environ.get("SSH_REMOTE_USER", "root") DB_PATH = LOGS_DIR / "homelab.db" API_KEY = os.environ.get("API_KEY", "dev-key-12345") # Répertoire pour les logs de tâches en markdown (format YYYY/MM/JJ) DIR_LOGS_TASKS = Path(os.environ.get("DIR_LOGS_TASKS", str(BASE_DIR.parent / "tasks_logs"))) # Fichier JSON pour l'historique des commandes ad-hoc ADHOC_HISTORY_FILE = DIR_LOGS_TASKS / ".adhoc_history.json" # Fichier JSON pour les statuts persistés BOOTSTRAP_STATUS_FILE = DIR_LOGS_TASKS / ".bootstrap_status.json" HOST_STATUS_FILE = ANSIBLE_DIR / ".host_status.json" # Mapping des actions vers les playbooks ACTION_PLAYBOOK_MAP = { 'upgrade': 'vm-upgrade.yml', 'reboot': 'vm-reboot.yml', 'health-check': 'health-check.yml', 'backup': 'backup-config.yml', 'bootstrap': 'bootstrap-host.yml', } # Gestionnaire de clés API api_key_header = APIKeyHeader(name="X-API-Key", auto_error=False) # Modèles Pydantic améliorés class CommandResult(BaseModel): status: str return_code: int stdout: str stderr: Optional[str] = None execution_time: Optional[float] = None timestamp: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) class Host(BaseModel): id: str name: str ip: str status: Literal["online", "offline", "warning"] os: str last_seen: Optional[datetime] = None created_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) groups: List[str] = [] # Groupes Ansible auxquels appartient l'hôte bootstrap_ok: bool = False # Indique si le bootstrap a été effectué avec succès bootstrap_date: Optional[datetime] = None # Date du dernier bootstrap réussi class Config: json_encoders = { datetime: lambda v: v.isoformat() } class Task(BaseModel): id: str name: str host: str status: Literal["pending", "running", "completed", "failed", "cancelled"] progress: int = Field(ge=0, le=100, default=0) start_time: Optional[datetime] = None end_time: Optional[datetime] = None duration: Optional[str] = None output: Optional[str] = None error: Optional[str] = None class Config: json_encoders = { datetime: lambda v: v.isoformat() if v else None } class LogEntry(BaseModel): id: int timestamp: datetime level: Literal["DEBUG", "INFO", "WARN", "ERROR"] message: str source: Optional[str] = None host: Optional[str] = None class Config: json_encoders = { datetime: lambda v: v.isoformat() } class SystemMetrics(BaseModel): online_hosts: int total_tasks: int success_rate: float uptime: float cpu_usage: float memory_usage: float disk_usage: float class HealthCheck(BaseModel): host: str ssh_ok: bool = False ansible_ok: bool = False sudo_ok: bool = False reachable: bool = False error_message: Optional[str] = None response_time: Optional[float] = None cached: bool = False cache_age: int = 0 class AnsibleExecutionRequest(BaseModel): playbook: str = Field(..., description="Nom du playbook à exécuter") target: str = Field(default="all", description="Hôte ou groupe cible") extra_vars: Optional[Dict[str, Any]] = Field(default=None, description="Variables supplémentaires") check_mode: bool = Field(default=False, description="Mode dry-run (--check)") verbose: bool = Field(default=False, description="Mode verbeux") class AnsibleInventoryHost(BaseModel): name: str ansible_host: str group: str groups: List[str] = [] # All groups this host belongs to vars: Dict[str, Any] = {} class TaskRequest(BaseModel): host: Optional[str] = Field(default=None, description="Hôte cible") group: Optional[str] = Field(default=None, description="Groupe cible") action: str = Field(..., description="Action à exécuter") cmd: Optional[str] = Field(default=None, description="Commande personnalisée") extra_vars: Optional[Dict[str, Any]] = Field(default=None, description="Variables Ansible") tags: Optional[List[str]] = Field(default=None, description="Tags Ansible") dry_run: bool = Field(default=False, description="Mode simulation") ssh_user: Optional[str] = Field(default=None, description="Utilisateur SSH") ssh_password: Optional[str] = Field(default=None, description="Mot de passe SSH") @field_validator('action') @classmethod def validate_action(cls, v: str) -> str: valid_actions = ['upgrade', 'reboot', 'health-check', 'backup', 'deploy', 'rollback', 'maintenance', 'bootstrap'] if v not in valid_actions: raise ValueError(f'Action doit être l\'une de: {", ".join(valid_actions)}') return v class HostRequest(BaseModel): name: str = Field(..., min_length=3, max_length=100, description="Hostname (ex: server.domain.home)") # ansible_host peut être soit une IPv4, soit un hostname résolvable → on enlève la contrainte de pattern ip: Optional[str] = Field(default=None, description="Adresse IP ou hostname (optionnel si hostname résolvable)") os: str = Field(default="Linux", min_length=3, max_length=50) ssh_user: Optional[str] = Field(default="root", min_length=1, max_length=50) ssh_port: int = Field(default=22, ge=1, le=65535) description: Optional[str] = Field(default=None, max_length=200) env_group: str = Field(..., description="Groupe d'environnement (ex: env_homelab, env_prod)") role_groups: List[str] = Field(default=[], description="Groupes de rôles (ex: role_proxmox, role_sbc)") class HostUpdateRequest(BaseModel): """Requête de mise à jour d'un hôte""" env_group: Optional[str] = Field(default=None, description="Nouveau groupe d'environnement") role_groups: Optional[List[str]] = Field(default=None, description="Nouveaux groupes de rôles") ansible_host: Optional[str] = Field(default=None, description="Nouvelle adresse ansible_host") class GroupRequest(BaseModel): """Requête pour créer un groupe""" name: str = Field(..., min_length=3, max_length=50, description="Nom du groupe (ex: env_prod, role_web)") type: str = Field(..., description="Type de groupe: 'env' ou 'role'") @field_validator('name') @classmethod def validate_name(cls, v: str) -> str: import re if not re.match(r'^[a-zA-Z0-9_-]+$', v): raise ValueError('Le nom du groupe ne peut contenir que des lettres, chiffres, tirets et underscores') return v @field_validator('type') @classmethod def validate_type(cls, v: str) -> str: if v not in ['env', 'role']: raise ValueError("Le type doit être 'env' ou 'role'") return v class GroupUpdateRequest(BaseModel): """Requête pour modifier un groupe""" new_name: str = Field(..., min_length=3, max_length=50, description="Nouveau nom du groupe") @field_validator('new_name') @classmethod def validate_new_name(cls, v: str) -> str: import re if not re.match(r'^[a-zA-Z0-9_-]+$', v): raise ValueError('Le nom du groupe ne peut contenir que des lettres, chiffres, tirets et underscores') return v class GroupDeleteRequest(BaseModel): """Requête pour supprimer un groupe""" move_hosts_to: Optional[str] = Field(default=None, description="Groupe vers lequel déplacer les hôtes") class AdHocCommandRequest(BaseModel): """Requête pour exécuter une commande ad-hoc Ansible""" target: str = Field(..., description="Hôte ou groupe cible") command: str = Field(..., description="Commande shell à exécuter") module: str = Field(default="shell", description="Module Ansible (shell, command, raw)") become: bool = Field(default=False, description="Exécuter avec sudo") timeout: int = Field(default=60, ge=5, le=600, description="Timeout en secondes") class AdHocCommandResult(BaseModel): """Résultat d'une commande ad-hoc""" target: str command: str success: bool return_code: int stdout: str stderr: Optional[str] = None duration: float hosts_results: Optional[Dict[str, Any]] = None class AdHocHistoryEntry(BaseModel): """Entrée dans l'historique des commandes ad-hoc""" id: str command: str target: str module: str become: bool category: str = "default" description: Optional[str] = None created_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) last_used: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) use_count: int = 1 class AdHocHistoryCategory(BaseModel): """Catégorie pour organiser les commandes ad-hoc""" name: str description: Optional[str] = None color: str = "#7c3aed" icon: str = "fa-folder" class TaskLogFile(BaseModel): """Représentation d'un fichier de log de tâche""" id: str filename: str path: str task_name: str target: str status: str date: str # Format YYYY-MM-DD year: str month: str day: str created_at: datetime size_bytes: int # Nouveaux champs pour affichage enrichi start_time: Optional[str] = None # Format ISO ou HH:MM:SS end_time: Optional[str] = None # Format ISO ou HH:MM:SS duration: Optional[str] = None # Durée formatée duration_seconds: Optional[int] = None # Durée en secondes hosts: List[str] = [] # Liste des hôtes impliqués category: Optional[str] = None # Catégorie (Playbook, Ad-hoc, etc.) subcategory: Optional[str] = None # Sous-catégorie target_type: Optional[str] = None # Type de cible: 'host', 'group', 'role' class TasksFilterParams(BaseModel): """Paramètres de filtrage des tâches""" status: Optional[str] = None # pending, running, completed, failed, all year: Optional[str] = None month: Optional[str] = None day: Optional[str] = None target: Optional[str] = None search: Optional[str] = None # ===== MODÈLES PLANIFICATEUR (SCHEDULER) ===== class ScheduleRecurrence(BaseModel): """Configuration de récurrence pour un schedule""" type: Literal["daily", "weekly", "monthly", "custom"] = "daily" time: str = Field(default="02:00", description="Heure d'exécution HH:MM") days: Optional[List[int]] = Field(default=None, description="Jours de la semaine (1-7, lundi=1) pour weekly") day_of_month: Optional[int] = Field(default=None, ge=1, le=31, description="Jour du mois (1-31) pour monthly") cron_expression: Optional[str] = Field(default=None, description="Expression cron pour custom") class Schedule(BaseModel): """Modèle d'un schedule de playbook""" id: str = Field(default_factory=lambda: f"sched_{uuid.uuid4().hex[:12]}") name: str = Field(..., min_length=3, max_length=100, description="Nom du schedule") description: Optional[str] = Field(default=None, max_length=500) playbook: str = Field(..., description="Nom du playbook à exécuter") target_type: Literal["group", "host"] = Field(default="group", description="Type de cible") target: str = Field(default="all", description="Nom du groupe ou hôte cible") extra_vars: Optional[Dict[str, Any]] = Field(default=None, description="Variables supplémentaires") schedule_type: Literal["once", "recurring"] = Field(default="recurring") recurrence: Optional[ScheduleRecurrence] = Field(default=None) timezone: str = Field(default="America/Montreal", description="Fuseau horaire") start_at: Optional[datetime] = Field(default=None, description="Date de début (optionnel)") end_at: Optional[datetime] = Field(default=None, description="Date de fin (optionnel)") next_run_at: Optional[datetime] = Field(default=None, description="Prochaine exécution calculée") last_run_at: Optional[datetime] = Field(default=None, description="Dernière exécution") last_status: Literal["success", "failed", "running", "never"] = Field(default="never") enabled: bool = Field(default=True, description="Schedule actif ou en pause") retry_on_failure: int = Field(default=0, ge=0, le=3, description="Nombre de tentatives en cas d'échec") timeout: int = Field(default=3600, ge=60, le=86400, description="Timeout en secondes") tags: List[str] = Field(default=[], description="Tags pour catégorisation") run_count: int = Field(default=0, description="Nombre total d'exécutions") success_count: int = Field(default=0, description="Nombre de succès") failure_count: int = Field(default=0, description="Nombre d'échecs") created_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) updated_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) class Config: json_encoders = { datetime: lambda v: v.isoformat() if v else None } @field_validator('recurrence', mode='before') @classmethod def validate_recurrence(cls, v, info): # Si schedule_type est 'once', recurrence n'est pas obligatoire return v class ScheduleRun(BaseModel): """Historique d'une exécution de schedule""" id: str = Field(default_factory=lambda: f"run_{uuid.uuid4().hex[:12]}") schedule_id: str = Field(..., description="ID du schedule parent") task_id: Optional[str] = Field(default=None, description="ID de la tâche créée") started_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc)) finished_at: Optional[datetime] = Field(default=None) status: Literal["running", "success", "failed", "canceled"] = Field(default="running") duration_seconds: Optional[float] = Field(default=None) hosts_impacted: int = Field(default=0) error_message: Optional[str] = Field(default=None) retry_attempt: int = Field(default=0, description="Numéro de la tentative (0 = première)") class Config: json_encoders = { datetime: lambda v: v.isoformat() if v else None } class ScheduleCreateRequest(BaseModel): """Requête de création d'un schedule""" name: str = Field(..., min_length=3, max_length=100) description: Optional[str] = Field(default=None, max_length=500) playbook: str = Field(...) target_type: Literal["group", "host"] = Field(default="group") target: str = Field(default="all") extra_vars: Optional[Dict[str, Any]] = Field(default=None) schedule_type: Literal["once", "recurring"] = Field(default="recurring") recurrence: Optional[ScheduleRecurrence] = Field(default=None) timezone: str = Field(default="America/Montreal") start_at: Optional[datetime] = Field(default=None) end_at: Optional[datetime] = Field(default=None) enabled: bool = Field(default=True) retry_on_failure: int = Field(default=0, ge=0, le=3) timeout: int = Field(default=3600, ge=60, le=86400) tags: List[str] = Field(default=[]) @field_validator('timezone') @classmethod def validate_timezone(cls, v: str) -> str: try: pytz.timezone(v) return v except pytz.exceptions.UnknownTimeZoneError: raise ValueError(f"Fuseau horaire invalide: {v}") class ScheduleUpdateRequest(BaseModel): """Requête de mise à jour d'un schedule""" name: Optional[str] = Field(default=None, min_length=3, max_length=100) description: Optional[str] = Field(default=None, max_length=500) playbook: Optional[str] = Field(default=None) target_type: Optional[Literal["group", "host"]] = Field(default=None) target: Optional[str] = Field(default=None) extra_vars: Optional[Dict[str, Any]] = Field(default=None) schedule_type: Optional[Literal["once", "recurring"]] = Field(default=None) recurrence: Optional[ScheduleRecurrence] = Field(default=None) timezone: Optional[str] = Field(default=None) start_at: Optional[datetime] = Field(default=None) end_at: Optional[datetime] = Field(default=None) enabled: Optional[bool] = Field(default=None) retry_on_failure: Optional[int] = Field(default=None, ge=0, le=3) timeout: Optional[int] = Field(default=None, ge=60, le=86400) tags: Optional[List[str]] = Field(default=None) class ScheduleStats(BaseModel): """Statistiques globales des schedules""" total: int = 0 active: int = 0 paused: int = 0 expired: int = 0 next_execution: Optional[datetime] = None next_schedule_name: Optional[str] = None failures_24h: int = 0 executions_24h: int = 0 success_rate_7d: float = 0.0 # ===== SERVICE DE LOGGING MARKDOWN ===== class TaskLogService: """Service pour gérer les logs de tâches en fichiers markdown""" def __init__(self, base_dir: Path): self.base_dir = base_dir self._ensure_base_dir() def _ensure_base_dir(self): """Crée le répertoire de base s'il n'existe pas""" self.base_dir.mkdir(parents=True, exist_ok=True) def _get_date_path(self, dt: datetime = None) -> Path: """Retourne le chemin du répertoire pour une date donnée (YYYY/MM/JJ)""" if dt is None: dt = datetime.now(timezone.utc) year = dt.strftime("%Y") month = dt.strftime("%m") day = dt.strftime("%d") return self.base_dir / year / month / day def _generate_task_id(self) -> str: """Génère un ID unique pour une tâche""" import uuid return f"task_{datetime.now(timezone.utc).strftime('%H%M%S')}_{uuid.uuid4().hex[:6]}" def save_task_log(self, task: 'Task', output: str = "", error: str = "") -> str: """Sauvegarde un log de tâche en markdown et retourne le chemin""" dt = task.start_time or datetime.now(timezone.utc) date_path = self._get_date_path(dt) date_path.mkdir(parents=True, exist_ok=True) # Générer le nom du fichier task_id = self._generate_task_id() status_emoji = { "completed": "✅", "failed": "❌", "running": "🔄", "pending": "⏳", "cancelled": "🚫" }.get(task.status, "❓") # Sanitize task name and host for filename safe_name = task.name.replace(' ', '_').replace(':', '').replace('/', '-')[:50] safe_host = task.host.replace(' ', '_').replace(':', '').replace('/', '-')[:30] if task.host else 'unknown' filename = f"{task_id}_{safe_host}_{safe_name}_{task.status}.md" filepath = date_path / filename # Créer le contenu markdown md_content = f"""# {status_emoji} {task.name} ## Informations | Propriété | Valeur | |-----------|--------| | **ID** | `{task.id}` | | **Nom** | {task.name} | | **Cible** | `{task.host}` | | **Statut** | {task.status} | | **Progression** | {task.progress}% | | **Début** | {task.start_time.isoformat() if task.start_time else 'N/A'} | | **Fin** | {task.end_time.isoformat() if task.end_time else 'N/A'} | | **Durée** | {task.duration or 'N/A'} | ## Sortie ``` {output or task.output or '(Aucune sortie)'} ``` """ if error or task.error: md_content += f"""## Erreurs ``` {error or task.error} ``` """ md_content += f"""--- *Généré automatiquement par Homelab Automation Dashboard* *Date: {datetime.now(timezone.utc).isoformat()}* """ # Écrire le fichier filepath.write_text(md_content, encoding='utf-8') return str(filepath) def _parse_markdown_metadata(self, content: str) -> Dict[str, Any]: """Parse le contenu markdown pour extraire les métadonnées enrichies""" metadata = { 'start_time': None, 'end_time': None, 'duration': None, 'duration_seconds': None, 'hosts': [], 'category': None, 'subcategory': None, 'target_type': None } # Extraire les heures de début et fin start_match = re.search(r'\|\s*\*\*Début\*\*\s*\|\s*([^|]+)', content) if start_match: start_val = start_match.group(1).strip() if start_val and start_val != 'N/A': metadata['start_time'] = start_val end_match = re.search(r'\|\s*\*\*Fin\*\*\s*\|\s*([^|]+)', content) if end_match: end_val = end_match.group(1).strip() if end_val and end_val != 'N/A': metadata['end_time'] = end_val duration_match = re.search(r'\|\s*\*\*Durée\*\*\s*\|\s*([^|]+)', content) if duration_match: dur_val = duration_match.group(1).strip() if dur_val and dur_val != 'N/A': metadata['duration'] = dur_val # Convertir en secondes si possible metadata['duration_seconds'] = self._parse_duration_to_seconds(dur_val) # Extraire les hôtes depuis la sortie Ansible # Pattern pour les hôtes dans PLAY RECAP ou les résultats de tâches host_patterns = [ r'^([a-zA-Z0-9][a-zA-Z0-9._-]+)\s*:\s*ok=', # PLAY RECAP format r'^\s*([a-zA-Z0-9][a-zA-Z0-9._-]+)\s*\|\s*(SUCCESS|CHANGED|FAILED|UNREACHABLE)', # Ad-hoc format ] hosts_found = set() for pattern in host_patterns: for match in re.finditer(pattern, content, re.MULTILINE): host = match.group(1).strip() if host and len(host) > 2 and '.' in host or len(host) > 5: hosts_found.add(host) metadata['hosts'] = sorted(list(hosts_found)) # Détecter la catégorie task_name_match = re.search(r'^#\s*[✅❌🔄⏳🚫❓]?\s*(.+)$', content, re.MULTILINE) if task_name_match: task_name = task_name_match.group(1).strip().lower() if 'playbook' in task_name: metadata['category'] = 'Playbook' # Extraire sous-catégorie du nom if 'health' in task_name: metadata['subcategory'] = 'Health Check' elif 'backup' in task_name: metadata['subcategory'] = 'Backup' elif 'upgrade' in task_name or 'update' in task_name: metadata['subcategory'] = 'Upgrade' elif 'bootstrap' in task_name: metadata['subcategory'] = 'Bootstrap' elif 'reboot' in task_name: metadata['subcategory'] = 'Reboot' elif 'ad-hoc' in task_name or 'adhoc' in task_name: metadata['category'] = 'Ad-hoc' else: metadata['category'] = 'Autre' # Détecter le type de cible target_match = re.search(r'\|\s*\*\*Cible\*\*\s*\|\s*`([^`]+)`', content) if target_match: target_val = target_match.group(1).strip() if target_val == 'all': metadata['target_type'] = 'group' elif target_val.startswith('env_') or target_val.startswith('role_'): metadata['target_type'] = 'group' elif '.' in target_val: metadata['target_type'] = 'host' else: metadata['target_type'] = 'group' return metadata def _parse_duration_to_seconds(self, duration_str: str) -> Optional[int]: """Convertit une chaîne de durée en secondes""" if not duration_str: return None total_seconds = 0 # Pattern: Xh Xm Xs ou X:XX:XX ou Xs # Format HH:MM:SS hms_match = re.match(r'(\d+):(\d+):(\d+)', duration_str) if hms_match: h, m, s = map(int, hms_match.groups()) return h * 3600 + m * 60 + s # Format avec h, m, s hours = re.search(r'(\d+)\s*h', duration_str) minutes = re.search(r'(\d+)\s*m', duration_str) seconds = re.search(r'(\d+)\s*s', duration_str) if hours: total_seconds += int(hours.group(1)) * 3600 if minutes: total_seconds += int(minutes.group(1)) * 60 if seconds: total_seconds += int(seconds.group(1)) return total_seconds if total_seconds > 0 else None def get_task_logs(self, year: str = None, month: str = None, day: str = None, status: str = None, target: str = None, category: str = None) -> List[TaskLogFile]: """Récupère la liste des logs de tâches avec filtrage""" logs = [] # Déterminer le chemin de recherche if year and month and day: search_paths = [self.base_dir / year / month / day] elif year and month: month_path = self.base_dir / year / month search_paths = list(month_path.glob("*")) if month_path.exists() else [] elif year: year_path = self.base_dir / year search_paths = [] if year_path.exists(): for m in year_path.iterdir(): if m.is_dir(): search_paths.extend(m.glob("*")) else: search_paths = [] if self.base_dir.exists(): for y in self.base_dir.iterdir(): if y.is_dir() and y.name.isdigit(): for m in y.iterdir(): if m.is_dir(): search_paths.extend(m.glob("*")) # Parcourir les répertoires for path in search_paths: if not path.is_dir(): continue for md_file in path.glob("*.md"): try: # Extraire les infos du nom de fichier # Format: task_HHMMSS_XXXXXX_TARGET_TASKNAME_STATUS.md parts = md_file.stem.split("_") if len(parts) >= 4: file_status = parts[-1] # Format nouveau: task_HHMMSS_XXXXXX_target_taskname_status # parts[0] = task, parts[1] = HHMMSS, parts[2] = XXXXXX (id) # parts[3] = target, parts[4:-1] = task_name, parts[-1] = status if len(parts) >= 5: file_target = parts[3] task_name_from_file = "_".join(parts[4:-1]) if len(parts) > 5 else parts[4] if len(parts) > 4 else "unknown" else: file_target = "" task_name_from_file = "_".join(parts[3:-1]) if len(parts) > 4 else parts[3] if len(parts) > 3 else "unknown" # Filtrer par statut si spécifié if status and status != "all" and file_status != status: continue # Extraire la date du chemin rel_path = md_file.relative_to(self.base_dir) path_parts = rel_path.parts if len(path_parts) >= 3: log_year, log_month, log_day = path_parts[0], path_parts[1], path_parts[2] else: continue stat = md_file.stat() # Lire le contenu pour extraire les métadonnées enrichies try: content = md_file.read_text(encoding='utf-8') metadata = self._parse_markdown_metadata(content) # Extraire le nom de tâche et la cible depuis le contenu markdown task_name_match = re.search(r'^#\s*[✅❌🔄⏳🚫❓]?\s*(.+)$', content, re.MULTILINE) if task_name_match: task_name = task_name_match.group(1).strip() else: task_name = task_name_from_file.replace("_", " ") # Extraire la cible depuis le contenu target_match = re.search(r'\|\s*\*\*Cible\*\*\s*\|\s*`([^`]+)`', content) if target_match: file_target = target_match.group(1).strip() except Exception: metadata = {} task_name = task_name_from_file.replace("_", " ") # Filtrer par target si spécifié if target and target != "all" and file_target: if target.lower() not in file_target.lower(): continue # Filtrer par catégorie si spécifié if category and category != "all": file_category = metadata.get('category', '') if file_category and category.lower() not in file_category.lower(): continue logs.append(TaskLogFile( id=parts[0] + "_" + parts[1] + "_" + parts[2] if len(parts) > 2 else parts[0], filename=md_file.name, path=str(md_file), task_name=task_name, target=file_target, status=file_status, date=f"{log_year}-{log_month}-{log_day}", year=log_year, month=log_month, day=log_day, created_at=datetime.fromtimestamp(stat.st_ctime, tz=timezone.utc), size_bytes=stat.st_size, start_time=metadata.get('start_time'), end_time=metadata.get('end_time'), duration=metadata.get('duration'), duration_seconds=metadata.get('duration_seconds'), hosts=metadata.get('hosts', []), category=metadata.get('category'), subcategory=metadata.get('subcategory'), target_type=metadata.get('target_type') )) except Exception: continue # Trier par date décroissante logs.sort(key=lambda x: x.created_at, reverse=True) return logs def get_available_dates(self) -> Dict[str, Any]: """Retourne la structure des dates disponibles pour le filtrage""" dates = {"years": {}} if not self.base_dir.exists(): return dates for year_dir in sorted(self.base_dir.iterdir(), reverse=True): if year_dir.is_dir() and year_dir.name.isdigit(): year = year_dir.name dates["years"][year] = {"months": {}} for month_dir in sorted(year_dir.iterdir(), reverse=True): if month_dir.is_dir() and month_dir.name.isdigit(): month = month_dir.name dates["years"][year]["months"][month] = {"days": []} for day_dir in sorted(month_dir.iterdir(), reverse=True): if day_dir.is_dir() and day_dir.name.isdigit(): day = day_dir.name count = len(list(day_dir.glob("*.md"))) dates["years"][year]["months"][month]["days"].append({ "day": day, "count": count }) return dates def get_stats(self) -> Dict[str, int]: """Retourne les statistiques des tâches""" stats = {"total": 0, "completed": 0, "failed": 0, "running": 0, "pending": 0} for log in self.get_task_logs(): stats["total"] += 1 if log.status in stats: stats[log.status] += 1 return stats # ===== SERVICE HISTORIQUE COMMANDES AD-HOC ===== class AdHocHistoryService: """Service pour gérer l'historique des commandes ad-hoc avec catégories""" def __init__(self, history_file: Path): self.history_file = history_file self._ensure_file() def _ensure_file(self): """Crée le fichier d'historique s'il n'existe pas""" self.history_file.parent.mkdir(parents=True, exist_ok=True) if not self.history_file.exists(): self._save_data({"commands": [], "categories": [ {"name": "default", "description": "Commandes générales", "color": "#7c3aed", "icon": "fa-terminal"}, {"name": "diagnostic", "description": "Commandes de diagnostic", "color": "#10b981", "icon": "fa-stethoscope"}, {"name": "maintenance", "description": "Commandes de maintenance", "color": "#f59e0b", "icon": "fa-wrench"}, {"name": "deployment", "description": "Commandes de déploiement", "color": "#3b82f6", "icon": "fa-rocket"}, ]}) def _load_data(self) -> Dict: """Charge les données depuis le fichier""" try: with open(self.history_file, 'r', encoding='utf-8') as f: return json.load(f) except: return {"commands": [], "categories": []} def _save_data(self, data: Dict): """Sauvegarde les données dans le fichier""" with open(self.history_file, 'w', encoding='utf-8') as f: json.dump(data, f, indent=2, default=str, ensure_ascii=False) def add_command(self, command: str, target: str, module: str, become: bool, category: str = "default", description: str = None) -> AdHocHistoryEntry: """Ajoute ou met à jour une commande dans l'historique""" data = self._load_data() # Chercher si la commande existe déjà existing = None for cmd in data["commands"]: if cmd["command"] == command and cmd["target"] == target: existing = cmd break if existing: existing["last_used"] = datetime.now(timezone.utc).isoformat() existing["use_count"] = existing.get("use_count", 1) + 1 if category != "default": existing["category"] = category if description: existing["description"] = description entry = AdHocHistoryEntry(**existing) else: import uuid entry = AdHocHistoryEntry( id=f"adhoc_{uuid.uuid4().hex[:8]}", command=command, target=target, module=module, become=become, category=category, description=description ) data["commands"].append(entry.dict()) self._save_data(data) return entry def get_commands(self, category: str = None, search: str = None, limit: int = 50) -> List[AdHocHistoryEntry]: """Récupère les commandes de l'historique""" data = self._load_data() commands = [] for cmd in data.get("commands", []): if category and cmd.get("category") != category: continue if search and search.lower() not in cmd.get("command", "").lower(): continue try: # Convertir les dates string en datetime si nécessaire if isinstance(cmd.get("created_at"), str): cmd["created_at"] = datetime.fromisoformat(cmd["created_at"].replace("Z", "+00:00")) if isinstance(cmd.get("last_used"), str): cmd["last_used"] = datetime.fromisoformat(cmd["last_used"].replace("Z", "+00:00")) commands.append(AdHocHistoryEntry(**cmd)) except Exception: continue # Trier par dernière utilisation commands.sort(key=lambda x: x.last_used, reverse=True) return commands[:limit] def get_categories(self) -> List[AdHocHistoryCategory]: """Récupère la liste des catégories""" data = self._load_data() return [AdHocHistoryCategory(**cat) for cat in data.get("categories", [])] def add_category(self, name: str, description: str = None, color: str = "#7c3aed", icon: str = "fa-folder") -> AdHocHistoryCategory: """Ajoute une nouvelle catégorie""" data = self._load_data() # Vérifier si la catégorie existe déjà for cat in data["categories"]: if cat["name"] == name: return AdHocHistoryCategory(**cat) new_cat = AdHocHistoryCategory(name=name, description=description, color=color, icon=icon) data["categories"].append(new_cat.dict()) self._save_data(data) return new_cat def delete_command(self, command_id: str) -> bool: """Supprime une commande de l'historique""" data = self._load_data() original_len = len(data["commands"]) data["commands"] = [c for c in data["commands"] if c.get("id") != command_id] if len(data["commands"]) < original_len: self._save_data(data) return True return False def update_command_category(self, command_id: str, category: str, description: str = None) -> bool: """Met à jour la catégorie d'une commande""" data = self._load_data() for cmd in data["commands"]: if cmd.get("id") == command_id: cmd["category"] = category if description: cmd["description"] = description self._save_data(data) return True return False def update_category(self, category_name: str, new_name: str, description: str, color: str, icon: str) -> bool: """Met à jour une catégorie existante""" data = self._load_data() for cat in data["categories"]: if cat["name"] == category_name: # Mettre à jour les commandes si le nom change if new_name != category_name: for cmd in data["commands"]: if cmd.get("category") == category_name: cmd["category"] = new_name cat["name"] = new_name cat["description"] = description cat["color"] = color cat["icon"] = icon self._save_data(data) return True return False def delete_category(self, category_name: str) -> bool: """Supprime une catégorie et déplace ses commandes vers 'default'""" if category_name == "default": return False data = self._load_data() # Vérifier si la catégorie existe cat_exists = any(cat["name"] == category_name for cat in data["categories"]) if not cat_exists: return False # Déplacer les commandes vers 'default' for cmd in data["commands"]: if cmd.get("category") == category_name: cmd["category"] = "default" # Supprimer la catégorie data["categories"] = [cat for cat in data["categories"] if cat["name"] != category_name] self._save_data(data) return True # ===== SERVICE BOOTSTRAP STATUS ===== class BootstrapStatusService: """Service pour gérer le statut de bootstrap des hôtes""" def __init__(self, status_file: Path): self.status_file = status_file self._ensure_file() def _ensure_file(self): """Crée le fichier de statut s'il n'existe pas""" self.status_file.parent.mkdir(parents=True, exist_ok=True) if not self.status_file.exists(): self._save_data({"hosts": {}}) def _load_data(self) -> Dict: """Charge les données depuis le fichier""" try: with open(self.status_file, 'r', encoding='utf-8') as f: return json.load(f) except: return {"hosts": {}} def _save_data(self, data: Dict): """Sauvegarde les données dans le fichier""" with open(self.status_file, 'w', encoding='utf-8') as f: json.dump(data, f, indent=2, default=str, ensure_ascii=False) def set_bootstrap_status(self, host_name: str, success: bool, details: str = None) -> Dict: """Enregistre le statut de bootstrap d'un hôte""" data = self._load_data() data["hosts"][host_name] = { "bootstrap_ok": success, "bootstrap_date": datetime.now(timezone.utc).isoformat(), "details": details } self._save_data(data) return data["hosts"][host_name] def get_bootstrap_status(self, host_name: str) -> Dict: """Récupère le statut de bootstrap d'un hôte""" data = self._load_data() return data.get("hosts", {}).get(host_name, { "bootstrap_ok": False, "bootstrap_date": None, "details": None }) def get_all_status(self) -> Dict[str, Dict]: """Récupère le statut de tous les hôtes""" data = self._load_data() return data.get("hosts", {}) def remove_host(self, host_name: str) -> bool: """Supprime le statut d'un hôte""" data = self._load_data() if host_name in data.get("hosts", {}): del data["hosts"][host_name] self._save_data(data) return True return False # ===== SERVICE HOST STATUS ===== class HostStatusService: def __init__(self, status_file: Path): self.status_file = status_file self._ensure_file() def _ensure_file(self): self.status_file.parent.mkdir(parents=True, exist_ok=True) if not self.status_file.exists(): self._save_data({"hosts": {}}) def _load_data(self) -> Dict: try: with open(self.status_file, 'r', encoding='utf-8') as f: return json.load(f) except: return {"hosts": {}} def _save_data(self, data: Dict): with open(self.status_file, 'w', encoding='utf-8') as f: json.dump(data, f, indent=2, default=str, ensure_ascii=False) def set_status(self, host_name: str, status: str, last_seen: Optional[datetime], os_info: Optional[str]) -> Dict: data = self._load_data() data.setdefault("hosts", {}) data["hosts"][host_name] = { "status": status, "last_seen": last_seen.isoformat() if isinstance(last_seen, datetime) else last_seen, "os": os_info, } self._save_data(data) return data["hosts"][host_name] def get_status(self, host_name: str) -> Dict: data = self._load_data() hosts = data.get("hosts", {}) return hosts.get(host_name, {"status": "online", "last_seen": None, "os": None}) def get_all_status(self) -> Dict[str, Dict]: data = self._load_data() return data.get("hosts", {}) def remove_host(self, host_name: str) -> bool: data = self._load_data() hosts = data.get("hosts", {}) if host_name in hosts: del hosts[host_name] data["hosts"] = hosts self._save_data(data) return True return False # ===== SERVICE PLANIFICATEUR (SCHEDULER) ===== SCHEDULES_FILE = DIR_LOGS_TASKS / ".schedules.json" SCHEDULE_RUNS_FILE = DIR_LOGS_TASKS / ".schedule_runs.json" class SchedulerService: """Service pour gérer les schedules de playbooks avec APScheduler""" def __init__(self, schedules_file: Path, runs_file: Path): self.schedules_file = schedules_file self.runs_file = runs_file self._ensure_files() # Configurer APScheduler jobstores = {'default': MemoryJobStore()} executors = {'default': AsyncIOExecutor()} job_defaults = {'coalesce': True, 'max_instances': 1, 'misfire_grace_time': 300} self.scheduler = AsyncIOScheduler( jobstores=jobstores, executors=executors, job_defaults=job_defaults, timezone=pytz.UTC ) self._started = False def _ensure_files(self): """Crée les fichiers de données s'ils n'existent pas""" self.schedules_file.parent.mkdir(parents=True, exist_ok=True) if not self.schedules_file.exists(): self._save_schedules([]) if not self.runs_file.exists(): self._save_runs([]) def _load_schedules(self) -> List[Dict]: """Charge les schedules depuis le fichier""" try: with open(self.schedules_file, 'r', encoding='utf-8') as f: data = json.load(f) return data.get("schedules", []) if isinstance(data, dict) else data except: return [] def _save_schedules(self, schedules: List[Dict]): """Sauvegarde les schedules dans le fichier""" with open(self.schedules_file, 'w', encoding='utf-8') as f: json.dump({"schedules": schedules}, f, indent=2, default=str, ensure_ascii=False) def _load_runs(self) -> List[Dict]: """Charge l'historique des exécutions""" try: with open(self.runs_file, 'r', encoding='utf-8') as f: data = json.load(f) return data.get("runs", []) if isinstance(data, dict) else data except: return [] def _save_runs(self, runs: List[Dict]): """Sauvegarde l'historique des exécutions""" # Garder seulement les 1000 dernières exécutions runs = runs[:1000] with open(self.runs_file, 'w', encoding='utf-8') as f: json.dump({"runs": runs}, f, indent=2, default=str, ensure_ascii=False) def start(self): """Démarre le scheduler et charge tous les schedules actifs""" if not self._started: self.scheduler.start() self._started = True # Charger les schedules actifs self._load_active_schedules() print("📅 Scheduler démarré avec succès") def shutdown(self): """Arrête le scheduler proprement""" if self._started: self.scheduler.shutdown(wait=False) self._started = False def _load_active_schedules(self): """Charge tous les schedules actifs dans APScheduler""" schedules = self._load_schedules() for sched_data in schedules: if sched_data.get('enabled', True): try: schedule = Schedule(**sched_data) self._add_job_for_schedule(schedule) except Exception as e: print(f"Erreur chargement schedule {sched_data.get('id')}: {e}") def _build_cron_trigger(self, schedule: Schedule) -> Optional[CronTrigger]: """Construit un trigger cron à partir de la configuration du schedule""" if schedule.schedule_type == "once": return None recurrence = schedule.recurrence if not recurrence: return None tz = pytz.timezone(schedule.timezone) hour, minute = recurrence.time.split(':') if recurrence.time else ("2", "0") try: if recurrence.type == "daily": return CronTrigger(hour=int(hour), minute=int(minute), timezone=tz) elif recurrence.type == "weekly": # Convertir jours (1-7 lundi=1) en format cron (0-6 lundi=0) days = recurrence.days or [1] day_of_week = ','.join(str(d - 1) for d in days) return CronTrigger(day_of_week=day_of_week, hour=int(hour), minute=int(minute), timezone=tz) elif recurrence.type == "monthly": day = recurrence.day_of_month or 1 return CronTrigger(day=day, hour=int(hour), minute=int(minute), timezone=tz) elif recurrence.type == "custom" and recurrence.cron_expression: # Parser l'expression cron parts = recurrence.cron_expression.split() if len(parts) == 5: return CronTrigger.from_crontab(recurrence.cron_expression, timezone=tz) else: # Expression cron étendue (6 champs avec secondes) return CronTrigger( second=parts[0] if len(parts) > 5 else '0', minute=parts[0] if len(parts) == 5 else parts[1], hour=parts[1] if len(parts) == 5 else parts[2], day=parts[2] if len(parts) == 5 else parts[3], month=parts[3] if len(parts) == 5 else parts[4], day_of_week=parts[4] if len(parts) == 5 else parts[5], timezone=tz ) except Exception as e: print(f"Erreur construction trigger cron: {e}") return None return None def _add_job_for_schedule(self, schedule: Schedule): """Ajoute un job APScheduler pour un schedule""" job_id = f"schedule_{schedule.id}" # Supprimer le job existant s'il existe try: self.scheduler.remove_job(job_id) except: pass if schedule.schedule_type == "once": # Exécution unique if schedule.start_at and schedule.start_at > datetime.now(timezone.utc): trigger = DateTrigger(run_date=schedule.start_at, timezone=pytz.UTC) self.scheduler.add_job( self._execute_schedule, trigger, id=job_id, args=[schedule.id], replace_existing=True ) else: # Exécution récurrente trigger = self._build_cron_trigger(schedule) if trigger: self.scheduler.add_job( self._execute_schedule, trigger, id=job_id, args=[schedule.id], replace_existing=True ) # Calculer et mettre à jour next_run_at self._update_next_run(schedule.id) def _update_next_run(self, schedule_id: str): """Met à jour le champ next_run_at d'un schedule""" job_id = f"schedule_{schedule_id}" try: job = self.scheduler.get_job(job_id) if job and job.next_run_time: schedules = self._load_schedules() for s in schedules: if s['id'] == schedule_id: s['next_run_at'] = job.next_run_time.isoformat() break self._save_schedules(schedules) except: pass async def _execute_schedule(self, schedule_id: str): """Exécute un schedule (appelé par APScheduler)""" # Import circulaire évité en utilisant les variables globales global ws_manager, ansible_service, db, task_log_service schedules = self._load_schedules() sched_data = next((s for s in schedules if s['id'] == schedule_id), None) if not sched_data: print(f"Schedule {schedule_id} non trouvé") return schedule = Schedule(**sched_data) # Vérifier si le schedule est encore actif if not schedule.enabled: return # Vérifier la fenêtre temporelle now = datetime.now(timezone.utc) if schedule.end_at and now > schedule.end_at: # Schedule expiré, le désactiver schedule.enabled = False self._update_schedule_in_storage(schedule) return # Créer un ScheduleRun run = ScheduleRun(schedule_id=schedule_id) runs = self._load_runs() runs.insert(0, run.dict()) self._save_runs(runs) # Mettre à jour le schedule schedule.last_run_at = now schedule.last_status = "running" schedule.run_count += 1 self._update_schedule_in_storage(schedule) # Notifier via WebSocket try: await ws_manager.broadcast({ "type": "schedule_run_started", "data": { "schedule_id": schedule_id, "schedule_name": schedule.name, "run": run.dict(), "status": "running" } }) except: pass # Créer une tâche task_id = str(db.get_next_id("tasks")) playbook_name = schedule.playbook.replace('.yml', '').replace('-', ' ').title() task = Task( id=task_id, name=f"[Planifié] {playbook_name}", host=schedule.target, status="running", progress=0, start_time=now ) db.tasks.insert(0, task) # Mettre à jour le run avec le task_id run.task_id = task_id runs = self._load_runs() for r in runs: if r['id'] == run.id: r['task_id'] = task_id break self._save_runs(runs) # Notifier la création de tâche try: await ws_manager.broadcast({ "type": "task_created", "data": task.dict() }) except: pass # Exécuter le playbook start_time = perf_counter() try: result = await ansible_service.execute_playbook( playbook=schedule.playbook, target=schedule.target, extra_vars=schedule.extra_vars, check_mode=False, verbose=True ) execution_time = perf_counter() - start_time success = result.get("success", False) # Mettre à jour la tâche task.status = "completed" if success else "failed" task.progress = 100 task.end_time = datetime.now(timezone.utc) task.duration = f"{execution_time:.1f}s" task.output = result.get("stdout", "") task.error = result.get("stderr", "") if not success else None # Mettre à jour le run run.status = "success" if success else "failed" run.finished_at = datetime.now(timezone.utc) run.duration_seconds = execution_time run.error_message = result.get("stderr", "") if not success else None # Compter les hôtes impactés stdout = result.get("stdout", "") host_count = len(re.findall(r'^[a-zA-Z0-9][a-zA-Z0-9._-]+\s*:\s*ok=', stdout, re.MULTILINE)) run.hosts_impacted = host_count # Mettre à jour le schedule schedule.last_status = "success" if success else "failed" if success: schedule.success_count += 1 else: schedule.failure_count += 1 # Sauvegarder self._update_schedule_in_storage(schedule) runs = self._load_runs() for r in runs: if r['id'] == run.id: r.update(run.dict()) break self._save_runs(runs) # Sauvegarder le log markdown try: task_log_service.save_task_log( task=task, output=result.get("stdout", ""), error=result.get("stderr", "") ) except: pass # Notifier await ws_manager.broadcast({ "type": "schedule_run_finished", "data": { "schedule_id": schedule_id, "schedule_name": schedule.name, "run": run.dict(), "status": run.status, "success": success } }) await ws_manager.broadcast({ "type": "task_completed", "data": { "id": task_id, "status": task.status, "progress": 100, "duration": task.duration, "success": success } }) # Log log_entry = LogEntry( id=db.get_next_id("logs"), timestamp=datetime.now(timezone.utc), level="INFO" if success else "ERROR", message=f"Schedule '{schedule.name}' exécuté: {'succès' if success else 'échec'}", source="scheduler", host=schedule.target ) db.logs.insert(0, log_entry) except Exception as e: # Échec de l'exécution execution_time = perf_counter() - start_time task.status = "failed" task.end_time = datetime.now(timezone.utc) task.error = str(e) run.status = "failed" run.finished_at = datetime.now(timezone.utc) run.duration_seconds = execution_time run.error_message = str(e) schedule.last_status = "failed" schedule.failure_count += 1 self._update_schedule_in_storage(schedule) runs = self._load_runs() for r in runs: if r['id'] == run.id: r.update(run.dict()) break self._save_runs(runs) try: task_log_service.save_task_log(task=task, error=str(e)) except: pass try: await ws_manager.broadcast({ "type": "schedule_run_finished", "data": { "schedule_id": schedule_id, "run": run.dict(), "status": "failed", "error": str(e) } }) await ws_manager.broadcast({ "type": "task_failed", "data": {"id": task_id, "status": "failed", "error": str(e)} }) except: pass log_entry = LogEntry( id=db.get_next_id("logs"), timestamp=datetime.now(timezone.utc), level="ERROR", message=f"Erreur schedule '{schedule.name}': {str(e)}", source="scheduler", host=schedule.target ) db.logs.insert(0, log_entry) # Mettre à jour next_run_at self._update_next_run(schedule_id) def _update_schedule_in_storage(self, schedule: Schedule): """Met à jour un schedule dans le stockage""" schedule.updated_at = datetime.now(timezone.utc) schedules = self._load_schedules() for i, s in enumerate(schedules): if s['id'] == schedule.id: schedules[i] = schedule.dict() break self._save_schedules(schedules) # ===== MÉTHODES PUBLIQUES CRUD ===== def get_all_schedules(self, enabled: Optional[bool] = None, playbook: Optional[str] = None, tag: Optional[str] = None) -> List[Schedule]: """Récupère tous les schedules avec filtrage optionnel""" schedules_data = self._load_schedules() schedules = [] for s in schedules_data: try: schedule = Schedule(**s) # Filtres if enabled is not None and schedule.enabled != enabled: continue if playbook and playbook.lower() not in schedule.playbook.lower(): continue if tag and tag not in schedule.tags: continue schedules.append(schedule) except: continue # Trier par prochaine exécution schedules.sort(key=lambda x: x.next_run_at or datetime.max.replace(tzinfo=timezone.utc)) return schedules def get_schedule(self, schedule_id: str) -> Optional[Schedule]: """Récupère un schedule par ID""" schedules = self._load_schedules() for s in schedules: if s['id'] == schedule_id: return Schedule(**s) return None def create_schedule(self, request: ScheduleCreateRequest) -> Schedule: """Crée un nouveau schedule""" schedule = Schedule( name=request.name, description=request.description, playbook=request.playbook, target_type=request.target_type, target=request.target, extra_vars=request.extra_vars, schedule_type=request.schedule_type, recurrence=request.recurrence, timezone=request.timezone, start_at=request.start_at, end_at=request.end_at, enabled=request.enabled, retry_on_failure=request.retry_on_failure, timeout=request.timeout, tags=request.tags ) # Sauvegarder schedules = self._load_schedules() schedules.append(schedule.dict()) self._save_schedules(schedules) # Ajouter le job si actif if schedule.enabled and self._started: self._add_job_for_schedule(schedule) return schedule def update_schedule(self, schedule_id: str, request: ScheduleUpdateRequest) -> Optional[Schedule]: """Met à jour un schedule existant""" schedule = self.get_schedule(schedule_id) if not schedule: return None # Appliquer les modifications update_data = request.dict(exclude_unset=True, exclude_none=True) for key, value in update_data.items(): # La récurrence arrive du frontend comme un dict, il faut la retransformer # en objet ScheduleRecurrence pour que _build_cron_trigger fonctionne. if key == "recurrence" and isinstance(value, dict): try: value = ScheduleRecurrence(**value) except Exception: # Si la récurrence est invalide, on laisse passer pour que la # validation côté endpoint remonte une erreur explicite. pass if hasattr(schedule, key): setattr(schedule, key, value) schedule.updated_at = datetime.now(timezone.utc) # Sauvegarder self._update_schedule_in_storage(schedule) # Mettre à jour le job if self._started: job_id = f"schedule_{schedule_id}" try: self.scheduler.remove_job(job_id) except: pass if schedule.enabled: self._add_job_for_schedule(schedule) return schedule def delete_schedule(self, schedule_id: str) -> bool: """Supprime un schedule""" schedules = self._load_schedules() original_len = len(schedules) schedules = [s for s in schedules if s['id'] != schedule_id] if len(schedules) < original_len: self._save_schedules(schedules) # Supprimer le job job_id = f"schedule_{schedule_id}" try: self.scheduler.remove_job(job_id) except: pass return True return False def pause_schedule(self, schedule_id: str) -> Optional[Schedule]: """Met en pause un schedule""" schedule = self.get_schedule(schedule_id) if not schedule: return None schedule.enabled = False self._update_schedule_in_storage(schedule) # Supprimer le job job_id = f"schedule_{schedule_id}" try: self.scheduler.remove_job(job_id) except: pass return schedule def resume_schedule(self, schedule_id: str) -> Optional[Schedule]: """Reprend un schedule en pause""" schedule = self.get_schedule(schedule_id) if not schedule: return None schedule.enabled = True self._update_schedule_in_storage(schedule) # Ajouter le job if self._started: self._add_job_for_schedule(schedule) return schedule async def run_now(self, schedule_id: str) -> Optional[ScheduleRun]: """Exécute immédiatement un schedule""" schedule = self.get_schedule(schedule_id) if not schedule: return None # Exécuter de manière asynchrone await self._execute_schedule(schedule_id) # Retourner le dernier run runs = self._load_runs() for r in runs: if r['schedule_id'] == schedule_id: return ScheduleRun(**r) return None def get_schedule_runs(self, schedule_id: str, limit: int = 50) -> List[ScheduleRun]: """Récupère l'historique des exécutions d'un schedule""" runs = self._load_runs() schedule_runs = [] for r in runs: if r['schedule_id'] == schedule_id: try: schedule_runs.append(ScheduleRun(**r)) except: continue return schedule_runs[:limit] def get_stats(self) -> ScheduleStats: """Calcule les statistiques globales des schedules""" schedules = self.get_all_schedules() runs = self._load_runs() now = datetime.now(timezone.utc) yesterday = now - timedelta(days=1) week_ago = now - timedelta(days=7) stats = ScheduleStats() stats.total = len(schedules) stats.active = len([s for s in schedules if s.enabled]) stats.paused = len([s for s in schedules if not s.enabled]) # Schedules expirés stats.expired = len([s for s in schedules if s.end_at and s.end_at < now]) # Prochaine exécution active_schedules = [s for s in schedules if s.enabled and s.next_run_at] if active_schedules: next_schedule = min(active_schedules, key=lambda x: x.next_run_at) stats.next_execution = next_schedule.next_run_at stats.next_schedule_name = next_schedule.name # Stats 24h runs_24h = [] for r in runs: try: started = datetime.fromisoformat(r['started_at'].replace('Z', '+00:00')) if isinstance(r['started_at'], str) else r['started_at'] if started >= yesterday: runs_24h.append(r) except: continue stats.executions_24h = len(runs_24h) stats.failures_24h = len([r for r in runs_24h if r.get('status') == 'failed']) # Taux de succès 7j runs_7d = [] for r in runs: try: started = datetime.fromisoformat(r['started_at'].replace('Z', '+00:00')) if isinstance(r['started_at'], str) else r['started_at'] if started >= week_ago: runs_7d.append(r) except: continue if runs_7d: success_count = len([r for r in runs_7d if r.get('status') == 'success']) stats.success_rate_7d = round((success_count / len(runs_7d)) * 100, 1) return stats def get_upcoming_executions(self, limit: int = 5) -> List[Dict]: """Retourne les prochaines exécutions planifiées""" schedules = self.get_all_schedules(enabled=True) upcoming = [] for s in schedules: if s.next_run_at: upcoming.append({ "schedule_id": s.id, "schedule_name": s.name, "playbook": s.playbook, "target": s.target, "next_run_at": s.next_run_at.isoformat() if s.next_run_at else None, "tags": s.tags }) upcoming.sort(key=lambda x: x['next_run_at'] or '') return upcoming[:limit] def validate_cron_expression(self, expression: str) -> Dict: """Valide une expression cron et retourne les prochaines exécutions""" try: cron = croniter(expression, datetime.now()) next_runs = [cron.get_next(datetime).isoformat() for _ in range(5)] return { "valid": True, "next_runs": next_runs, "expression": expression } except Exception as e: return { "valid": False, "error": str(e), "expression": expression } def get_runs_for_schedule(self, schedule_id: str, limit: int = 50) -> List[Dict]: """Récupère l'historique des exécutions d'un schedule (retourne des dicts)""" runs = self._load_runs() schedule_runs = [r for r in runs if r.get('schedule_id') == schedule_id] return schedule_runs[:limit] def cleanup_old_runs(self, days: int = 90): """Nettoie les exécutions plus anciennes que X jours""" cutoff = datetime.now(timezone.utc) - timedelta(days=days) runs = self._load_runs() new_runs = [] for r in runs: try: started = datetime.fromisoformat(r['started_at'].replace('Z', '+00:00')) if isinstance(r['started_at'], str) else r['started_at'] if started >= cutoff: new_runs.append(r) except: new_runs.append(r) # Garder si on ne peut pas parser la date self._save_runs(new_runs) return len(runs) - len(new_runs) # Instances globales des services task_log_service = TaskLogService(DIR_LOGS_TASKS) adhoc_history_service = AdHocHistoryService(ADHOC_HISTORY_FILE) bootstrap_status_service = BootstrapStatusService(BOOTSTRAP_STATUS_FILE) host_status_service = HostStatusService(HOST_STATUS_FILE) scheduler_service = SchedulerService(SCHEDULES_FILE, SCHEDULE_RUNS_FILE) class WebSocketManager: def __init__(self): self.active_connections: List[WebSocket] = [] self.lock = Lock() async def connect(self, websocket: WebSocket): await websocket.accept() with self.lock: self.active_connections.append(websocket) def disconnect(self, websocket: WebSocket): with self.lock: if websocket in self.active_connections: self.active_connections.remove(websocket) async def broadcast(self, message: dict): with self.lock: disconnected = [] for connection in self.active_connections: try: await connection.send_json(message) except: disconnected.append(connection) # Nettoyer les connexions déconnectées for conn in disconnected: self.active_connections.remove(conn) # Instance globale du gestionnaire WebSocket ws_manager = WebSocketManager() # Service Ansible class AnsibleService: """Service pour exécuter les playbooks Ansible""" def __init__(self, ansible_dir: Path): self.ansible_dir = ansible_dir self.playbooks_dir = ansible_dir / "playbooks" self.inventory_path = ansible_dir / "inventory" / "hosts.yml" self._inventory_cache: Optional[Dict] = None def get_playbooks(self) -> List[Dict[str, Any]]: """Liste les playbooks disponibles avec leurs métadonnées (category/subcategory). Les métadonnées sont lues en priorité dans play['vars'] pour être compatibles avec la syntaxe Ansible (category/subcategory ne sont pas des clés de Play). """ playbooks = [] if self.playbooks_dir.exists(): for pb in self.playbooks_dir.glob("*.yml"): # Récupérer les infos du fichier stat = pb.stat() playbook_info = { "name": pb.stem, "filename": pb.name, "path": str(pb), "category": "general", "subcategory": "other", "size": stat.st_size, "modified": datetime.fromtimestamp(stat.st_mtime, tz=timezone.utc).isoformat() } # Extract category/subcategory from playbook try: with open(pb, 'r', encoding='utf-8') as f: content = yaml.safe_load(f) if content and isinstance(content, list) and len(content) > 0: play = content[0] vars_ = play.get('vars', {}) or {} # Lecture de category avec fallback: play puis vars if 'category' in play: playbook_info['category'] = play['category'] elif 'category' in vars_: playbook_info['category'] = vars_['category'] # Lecture de subcategory avec fallback if 'subcategory' in play: playbook_info['subcategory'] = play['subcategory'] elif 'subcategory' in vars_: playbook_info['subcategory'] = vars_['subcategory'] if 'name' in play: playbook_info['description'] = play['name'] except Exception: # On ignore les erreurs de parsing individuelles pour ne pas # casser l'ensemble de la liste de playbooks. pass playbooks.append(playbook_info) return playbooks def get_playbook_categories(self) -> Dict[str, List[str]]: """Retourne les catégories et sous-catégories des playbooks""" categories = {} for pb in self.get_playbooks(): cat = pb.get('category', 'general') subcat = pb.get('subcategory', 'other') if cat not in categories: categories[cat] = [] if subcat not in categories[cat]: categories[cat].append(subcat) return categories def load_inventory(self) -> Dict: """Charge l'inventaire Ansible depuis le fichier YAML""" if self._inventory_cache: return self._inventory_cache if not self.inventory_path.exists(): return {} with open(self.inventory_path, 'r') as f: self._inventory_cache = yaml.safe_load(f) return self._inventory_cache def get_hosts_from_inventory(self, group_filter: str = None) -> List[AnsibleInventoryHost]: """Extrait la liste des hôtes de l'inventaire sans doublons. Args: group_filter: Si spécifié, filtre les hôtes par ce groupe """ inventory = self.load_inventory() # Use dict to track unique hosts and accumulate their groups hosts_dict: Dict[str, AnsibleInventoryHost] = {} def extract_hosts(data: Dict, current_group: str = ""): if not isinstance(data, dict): return # Extraire les hôtes directs if 'hosts' in data: for host_name, host_data in data['hosts'].items(): host_data = host_data or {} if host_name in hosts_dict: # Host already exists, add group to its groups list if current_group and current_group not in hosts_dict[host_name].groups: hosts_dict[host_name].groups.append(current_group) else: # New host hosts_dict[host_name] = AnsibleInventoryHost( name=host_name, ansible_host=host_data.get('ansible_host', host_name), group=current_group, groups=[current_group] if current_group else [], vars=host_data ) # Parcourir les enfants (sous-groupes) if 'children' in data: for child_name, child_data in data['children'].items(): extract_hosts(child_data, child_name) extract_hosts(inventory.get('all', {})) # Convert to list hosts = list(hosts_dict.values()) # Apply group filter if specified if group_filter and group_filter != 'all': hosts = [h for h in hosts if group_filter in h.groups] return hosts def invalidate_cache(self): """Invalide le cache de l'inventaire pour forcer un rechargement""" self._inventory_cache = None def get_groups(self) -> List[str]: """Extrait la liste des groupes de l'inventaire""" inventory = self.load_inventory() groups = set() def extract_groups(data: Dict, parent: str = ""): if not isinstance(data, dict): return if 'children' in data: for child_name in data['children'].keys(): groups.add(child_name) extract_groups(data['children'][child_name], child_name) extract_groups(inventory.get('all', {})) return sorted(list(groups)) def get_env_groups(self) -> List[str]: """Retourne uniquement les groupes d'environnement (préfixés par env_)""" return [g for g in self.get_groups() if g.startswith('env_')] def get_role_groups(self) -> List[str]: """Retourne uniquement les groupes de rôles (préfixés par role_)""" return [g for g in self.get_groups() if g.startswith('role_')] def _save_inventory(self, inventory: Dict): """Sauvegarde l'inventaire dans le fichier YAML""" # Créer une sauvegarde avant modification backup_path = self.inventory_path.with_suffix('.yml.bak') if self.inventory_path.exists(): import shutil shutil.copy2(self.inventory_path, backup_path) with open(self.inventory_path, 'w', encoding='utf-8') as f: yaml.dump(inventory, f, default_flow_style=False, allow_unicode=True, sort_keys=False) # Invalider le cache self.invalidate_cache() def add_host_to_inventory(self, hostname: str, env_group: str, role_groups: List[str], ansible_host: str = None) -> bool: """Ajoute un hôte à l'inventaire Ansible Args: hostname: Nom de l'hôte (ex: server.domain.home) env_group: Groupe d'environnement (ex: env_homelab) role_groups: Liste des groupes de rôles (ex: ['role_proxmox', 'role_sbc']) ansible_host: Adresse IP ou hostname pour ansible_host (optionnel) Returns: True si l'ajout a réussi """ inventory = self.load_inventory() # S'assurer que la structure existe if 'all' not in inventory: inventory['all'] = {} if 'children' not in inventory['all']: inventory['all']['children'] = {} children = inventory['all']['children'] # Ajouter au groupe d'environnement if env_group not in children: children[env_group] = {'hosts': {}} if 'hosts' not in children[env_group]: children[env_group]['hosts'] = {} # Définir les variables de l'hôte host_vars = None if ansible_host and ansible_host != hostname: host_vars = {'ansible_host': ansible_host} children[env_group]['hosts'][hostname] = host_vars # Ajouter aux groupes de rôles for role_group in role_groups: if role_group not in children: children[role_group] = {'hosts': {}} if 'hosts' not in children[role_group]: children[role_group]['hosts'] = {} children[role_group]['hosts'][hostname] = None self._save_inventory(inventory) return True def remove_host_from_inventory(self, hostname: str) -> bool: """Supprime un hôte de tous les groupes de l'inventaire Args: hostname: Nom de l'hôte à supprimer Returns: True si la suppression a réussi """ inventory = self.load_inventory() if 'all' not in inventory or 'children' not in inventory['all']: return False children = inventory['all']['children'] removed = False # Parcourir tous les groupes et supprimer l'hôte for group_name, group_data in children.items(): if isinstance(group_data, dict) and 'hosts' in group_data: if hostname in group_data['hosts']: del group_data['hosts'][hostname] removed = True if removed: self._save_inventory(inventory) # Supprimer aussi les statuts persistés (bootstrap + health) bootstrap_status_service.remove_host(hostname) try: host_status_service.remove_host(hostname) except Exception: pass return removed def update_host_groups(self, hostname: str, env_group: str = None, role_groups: List[str] = None, ansible_host: str = None) -> bool: """Met à jour les groupes d'un hôte existant Args: hostname: Nom de l'hôte à modifier env_group: Nouveau groupe d'environnement (None = pas de changement) role_groups: Nouvelle liste de groupes de rôles (None = pas de changement) ansible_host: Nouvelle adresse ansible_host (None = pas de changement) Returns: True si la mise à jour a réussi """ inventory = self.load_inventory() if 'all' not in inventory or 'children' not in inventory['all']: return False children = inventory['all']['children'] # Trouver le groupe d'environnement actuel current_env_group = None current_role_groups = [] current_ansible_host = None for group_name, group_data in children.items(): if isinstance(group_data, dict) and 'hosts' in group_data: if hostname in group_data['hosts']: if group_name.startswith('env_'): current_env_group = group_name # Récupérer ansible_host si défini host_vars = group_data['hosts'][hostname] if isinstance(host_vars, dict) and 'ansible_host' in host_vars: current_ansible_host = host_vars['ansible_host'] elif group_name.startswith('role_'): current_role_groups.append(group_name) if not current_env_group: return False # Hôte non trouvé # Appliquer les changements new_env_group = env_group if env_group else current_env_group new_role_groups = role_groups if role_groups is not None else current_role_groups new_ansible_host = ansible_host if ansible_host else current_ansible_host # Supprimer l'hôte de tous les groupes actuels for group_name, group_data in children.items(): if isinstance(group_data, dict) and 'hosts' in group_data: if hostname in group_data['hosts']: del group_data['hosts'][hostname] # Ajouter au nouveau groupe d'environnement if new_env_group not in children: children[new_env_group] = {'hosts': {}} if 'hosts' not in children[new_env_group]: children[new_env_group]['hosts'] = {} host_vars = None if new_ansible_host and new_ansible_host != hostname: host_vars = {'ansible_host': new_ansible_host} children[new_env_group]['hosts'][hostname] = host_vars # Ajouter aux nouveaux groupes de rôles for role_group in new_role_groups: if role_group not in children: children[role_group] = {'hosts': {}} if 'hosts' not in children[role_group]: children[role_group]['hosts'] = {} children[role_group]['hosts'][hostname] = None self._save_inventory(inventory) return True def host_exists(self, hostname: str) -> bool: """Vérifie si un hôte existe dans l'inventaire""" hosts = self.get_hosts_from_inventory() return any(h.name == hostname for h in hosts) def group_exists(self, group_name: str) -> bool: """Vérifie si un groupe existe dans l'inventaire""" return group_name in self.get_groups() def add_group(self, group_name: str) -> bool: """Ajoute un nouveau groupe à l'inventaire Args: group_name: Nom du groupe (doit commencer par env_ ou role_) Returns: True si l'ajout a réussi """ if self.group_exists(group_name): return False # Groupe existe déjà inventory = self.load_inventory() # S'assurer que la structure existe if 'all' not in inventory: inventory['all'] = {} if 'children' not in inventory['all']: inventory['all']['children'] = {} # Ajouter le groupe vide inventory['all']['children'][group_name] = {'hosts': {}} self._save_inventory(inventory) return True def rename_group(self, old_name: str, new_name: str) -> bool: """Renomme un groupe dans l'inventaire Args: old_name: Nom actuel du groupe new_name: Nouveau nom du groupe Returns: True si le renommage a réussi """ if not self.group_exists(old_name): return False # Groupe source n'existe pas if self.group_exists(new_name): return False # Groupe cible existe déjà inventory = self.load_inventory() children = inventory.get('all', {}).get('children', {}) if old_name not in children: return False # Copier les données du groupe vers le nouveau nom children[new_name] = children[old_name] del children[old_name] self._save_inventory(inventory) return True def delete_group(self, group_name: str, move_hosts_to: str = None) -> Dict[str, Any]: """Supprime un groupe de l'inventaire Args: group_name: Nom du groupe à supprimer move_hosts_to: Groupe vers lequel déplacer les hôtes (optionnel) Returns: Dict avec le résultat de l'opération """ if not self.group_exists(group_name): return {"success": False, "error": "Groupe non trouvé"} inventory = self.load_inventory() children = inventory.get('all', {}).get('children', {}) if group_name not in children: return {"success": False, "error": "Groupe non trouvé dans children"} group_data = children[group_name] hosts_in_group = list(group_data.get('hosts', {}).keys()) if group_data else [] # Si des hôtes sont dans le groupe et qu'on veut les déplacer if hosts_in_group and move_hosts_to: if not self.group_exists(move_hosts_to) and move_hosts_to != group_name: # Créer le groupe cible s'il n'existe pas children[move_hosts_to] = {'hosts': {}} if move_hosts_to in children: if 'hosts' not in children[move_hosts_to]: children[move_hosts_to]['hosts'] = {} # Déplacer les hôtes for hostname in hosts_in_group: host_vars = group_data['hosts'].get(hostname) children[move_hosts_to]['hosts'][hostname] = host_vars # Supprimer le groupe del children[group_name] self._save_inventory(inventory) return { "success": True, "hosts_affected": hosts_in_group, "hosts_moved_to": move_hosts_to if hosts_in_group and move_hosts_to else None } def get_group_hosts(self, group_name: str) -> List[str]: """Retourne la liste des hôtes dans un groupe Args: group_name: Nom du groupe Returns: Liste des noms d'hôtes """ inventory = self.load_inventory() children = inventory.get('all', {}).get('children', {}) if group_name not in children: return [] group_data = children[group_name] if not group_data or 'hosts' not in group_data: return [] return list(group_data['hosts'].keys()) async def execute_playbook( self, playbook: str, target: str = "all", extra_vars: Optional[Dict[str, Any]] = None, check_mode: bool = False, verbose: bool = False ) -> Dict[str, Any]: """Exécute un playbook Ansible""" # Résoudre le chemin du playbook # On accepte soit un nom avec extension, soit un nom sans extension (ex: "health-check") playbook_path = self.playbooks_dir / playbook # Si le fichier n'existe pas tel quel, essayer avec des extensions courantes if not playbook_path.exists(): from pathlib import Path pb_name = Path(playbook).name # enlever d'éventuels chemins # Si aucune extension n'est fournie, tester .yml puis .yaml if not Path(pb_name).suffix: for ext in (".yml", ".yaml"): candidate = self.playbooks_dir / f"{pb_name}{ext}" if candidate.exists(): playbook_path = candidate break if not playbook_path.exists(): # À ce stade, on n'a trouvé aucun fichier correspondant raise FileNotFoundError(f"Playbook introuvable: {playbook}") # Construire la commande ansible-playbook cmd = [ "ansible-playbook", str(playbook_path), "-i", str(self.inventory_path), "--limit", target ] if check_mode: cmd.append("--check") if verbose: cmd.append("-v") if extra_vars: cmd.extend(["--extra-vars", json.dumps(extra_vars)]) private_key = find_ssh_private_key() if private_key: cmd.extend(["--private-key", private_key]) if SSH_USER: cmd.extend(["-u", SSH_USER]) start_time = perf_counter() try: # Exécuter la commande process = await asyncio.create_subprocess_exec( *cmd, stdout=asyncio.subprocess.PIPE, stderr=asyncio.subprocess.PIPE, cwd=str(self.ansible_dir) ) stdout, stderr = await process.communicate() execution_time = perf_counter() - start_time return { "success": process.returncode == 0, "return_code": process.returncode, "stdout": stdout.decode('utf-8', errors='replace'), "stderr": stderr.decode('utf-8', errors='replace'), "execution_time": round(execution_time, 2), "command": " ".join(cmd) } except FileNotFoundError: return { "success": False, "return_code": -1, "stdout": "", "stderr": "ansible-playbook non trouvé. Vérifiez que Ansible est installé.", "execution_time": 0, "command": " ".join(cmd) } except Exception as e: return { "success": False, "return_code": -1, "stdout": "", "stderr": str(e), "execution_time": perf_counter() - start_time, "command": " ".join(cmd) } # Instance globale du service Ansible ansible_service = AnsibleService(ANSIBLE_DIR) # ===== SERVICE BOOTSTRAP SSH ===== class BootstrapRequest(BaseModel): """Requête de bootstrap pour un hôte""" host: str = Field(..., description="Adresse IP ou hostname de l'hôte") root_password: str = Field(..., description="Mot de passe root pour la connexion initiale") automation_user: str = Field(default="automation", description="Nom de l'utilisateur d'automatisation à créer") class CommandResult(BaseModel): """Résultat d'une commande SSH""" status: str return_code: int stdout: str stderr: Optional[str] = None def find_ssh_private_key() -> Optional[str]: """Trouve une clé privée SSH disponible en inspectant plusieurs répertoires.""" candidate_dirs = [] env_path = Path(SSH_KEY_PATH) candidate_dirs.append(env_path.parent) candidate_dirs.append(Path("/app/ssh_keys")) candidate_dirs.append(Path.home() / ".ssh") seen = set() key_paths: List[str] = [] for directory in candidate_dirs: if not directory or not directory.exists(): continue for name in [ env_path.name, "id_automation_ansible", "id_rsa", "id_ed25519", "id_ecdsa", ]: path = directory / name if str(path) not in seen: seen.add(str(path)) key_paths.append(str(path)) # Ajouter dynamiquement toutes les clés sans extension .pub for file in directory.iterdir(): if file.is_file() and not file.suffix and not file.name.startswith("known_hosts"): if str(file) not in seen: seen.add(str(file)) key_paths.append(str(file)) for key_path in key_paths: if key_path and Path(key_path).exists(): return key_path return None def run_ssh_command( host: str, command: str, ssh_user: str = "root", ssh_password: Optional[str] = None, timeout: int = 60 ) -> tuple: """Exécute une commande SSH sur un hôte distant. Returns: tuple: (return_code, stdout, stderr) """ ssh_cmd = ["ssh"] # Options SSH communes ssh_opts = [ "-o", "StrictHostKeyChecking=no", "-o", "UserKnownHostsFile=/dev/null", "-o", "ConnectTimeout=10", "-o", "BatchMode=no" if ssh_password else "BatchMode=yes", ] # Si pas de mot de passe, utiliser la clé SSH if not ssh_password: private_key = find_ssh_private_key() if private_key: ssh_opts.extend(["-i", private_key]) ssh_cmd.extend(ssh_opts) ssh_cmd.append(f"{ssh_user}@{host}") ssh_cmd.append(command) try: if ssh_password: # Utiliser sshpass pour l'authentification par mot de passe full_cmd = ["sshpass", "-p", ssh_password] + ssh_cmd else: full_cmd = ssh_cmd result = subprocess.run( full_cmd, capture_output=True, text=True, timeout=timeout ) return result.returncode, result.stdout, result.stderr except subprocess.TimeoutExpired: return -1, "", f"Timeout après {timeout} secondes" except FileNotFoundError as e: if "sshpass" in str(e): return -1, "", "sshpass n'est pas installé. Installez-le avec: apt install sshpass" return -1, "", str(e) except Exception as e: return -1, "", str(e) def bootstrap_host(host: str, root_password: str, automation_user: str = "automation") -> CommandResult: """Prépare un hôte pour Ansible (création user, clé SSH, sudo, python3) pour Debian/Alpine/FreeBSD. Utilise un script shell complet uploadé via heredoc pour éviter les problèmes de quoting. """ import logging logger = logging.getLogger("bootstrap") # Chercher la clé publique dans plusieurs emplacements possibles primary_dirs = [ Path(SSH_KEY_PATH).parent, Path("/app/ssh_keys"), Path.home() / ".ssh", ] ssh_dir = primary_dirs[0] pub_paths = [ SSH_KEY_PATH + ".pub", "/app/ssh_keys/id_rsa.pub", "/app/ssh_keys/id_ed25519.pub", "/app/ssh_keys/id_ecdsa.pub", "/app/ssh_keys/id_automation_ansible.pub", ] # Ajouter dynamiquement toutes les clés .pub trouvées dans le répertoire SSH for directory in primary_dirs: if not directory.exists(): continue for f in directory.iterdir(): if f.is_file() and f.suffix == ".pub" and str(f) not in pub_paths: pub_paths.append(str(f)) logger.info(f"SSH_KEY_PATH = {SSH_KEY_PATH}") logger.info(f"Recherche de clé publique dans: {pub_paths}") pub_key = None pub_path_used = None for pub_path in pub_paths: try: if Path(pub_path).exists(): pub_key = Path(pub_path).read_text(encoding="utf-8").strip() if pub_key: pub_path_used = pub_path logger.info(f"Clé publique trouvée: {pub_path}") break except Exception as e: logger.warning(f"Erreur lecture {pub_path}: {e}") continue if not pub_key: # Lister les fichiers disponibles pour le debug ssh_dir = Path(SSH_KEY_PATH).parent available_files = [] if ssh_dir.exists(): available_files = [f.name for f in ssh_dir.iterdir()] raise HTTPException( status_code=500, detail=f"Clé publique SSH non trouvée. Chemins testés: {pub_paths}. Fichiers disponibles dans {ssh_dir}: {available_files}", ) # Script shell complet, robuste, avec logs détaillés bootstrap_script = f"""#!/bin/sh set -e AUT_USER="{automation_user}" echo "=== Bootstrap Ansible Host ===" echo "User: $AUT_USER" echo "" # 1) Détection OS if command -v apk >/dev/null 2>&1; then OS_TYPE="alpine" echo "[1/7] OS détecté: Alpine Linux" elif [ "$(uname -s 2>/dev/null)" = "FreeBSD" ] || \ command -v pkg >/dev/null 2>&1 || \ ( [ -f /etc/os-release ] && grep -qi 'ID=freebsd' /etc/os-release ); then OS_TYPE="freebsd" echo "[1/7] OS détecté: FreeBSD" else OS_TYPE="debian" echo "[1/7] OS détecté: Debian-like" fi # 2) Vérification / préparation utilisateur echo "[2/7] Vérification utilisateur/groupe..." if id "$AUT_USER" >/dev/null 2>&1; then echo " - Utilisateur déjà existant: $AUT_USER (aucune suppression)" else echo " - Utilisateur inexistant, il sera créé" fi # 3) Création utilisateur (idempotent) echo "[3/7] Création utilisateur $AUT_USER..." if id "$AUT_USER" >/dev/null 2>&1; then echo " - Utilisateur déjà présent, réutilisation" elif [ "$OS_TYPE" = "alpine" ]; then adduser -D "$AUT_USER" echo " - Utilisateur créé (Alpine: adduser -D)" elif [ "$OS_TYPE" = "freebsd" ]; then pw useradd "$AUT_USER" -m -s /bin/sh echo " - Utilisateur créé (FreeBSD: pw useradd)" else useradd -m -s /bin/bash "$AUT_USER" || useradd -m -s /bin/sh "$AUT_USER" echo " - Utilisateur créé (Debian: useradd -m)" fi # 3b) S'assurer que le compte n'est pas verrouillé echo " - Vérification du verrouillage du compte..." if command -v passwd >/dev/null 2>&1; then passwd -u "$AUT_USER" 2>/dev/null || true fi if command -v usermod >/dev/null 2>&1; then usermod -U "$AUT_USER" 2>/dev/null || true fi # 4) Configuration clé SSH echo "[4/7] Configuration clé SSH..." HOME_DIR=$(getent passwd "$AUT_USER" | cut -d: -f6) if [ -z "$HOME_DIR" ]; then HOME_DIR="/home/$AUT_USER" fi echo " - HOME_DIR: $HOME_DIR" mkdir -p "$HOME_DIR/.ssh" chown "$AUT_USER":"$AUT_USER" "$HOME_DIR/.ssh" chmod 700 "$HOME_DIR/.ssh" echo " - Répertoire .ssh créé et configuré" cat > "$HOME_DIR/.ssh/authorized_keys" << 'SSHKEY_EOF' {pub_key} SSHKEY_EOF chown "$AUT_USER":"$AUT_USER" "$HOME_DIR/.ssh/authorized_keys" chmod 600 "$HOME_DIR/.ssh/authorized_keys" echo " - Clé publique installée dans authorized_keys" if [ -s "$HOME_DIR/.ssh/authorized_keys" ]; then KEY_COUNT=$(wc -l < "$HOME_DIR/.ssh/authorized_keys") echo " - Vérification: $KEY_COUNT clé(s) dans authorized_keys" else echo " - ERREUR: authorized_keys vide ou absent!" exit 1 fi # 5) Installation sudo echo "[5/7] Installation sudo..." if command -v sudo >/dev/null 2>&1; then echo " - sudo déjà installé" else if [ "$OS_TYPE" = "alpine" ]; then apk add --no-cache sudo echo " - sudo installé (apk)" elif [ "$OS_TYPE" = "freebsd" ]; then pkg install -y sudo echo " - sudo installé (pkg)" else apt-get update -qq && apt-get install -y sudo echo " - sudo installé (apt)" fi fi # 6) Configuration sudoers echo "[6/7] Configuration sudoers..." if [ ! -d /etc/sudoers.d ]; then mkdir -p /etc/sudoers.d chmod 750 /etc/sudoers.d 2>/dev/null || true echo " - Répertoire /etc/sudoers.d créé" fi echo "$AUT_USER ALL=(ALL) NOPASSWD: ALL" > /etc/sudoers.d/automation chmod 440 /etc/sudoers.d/automation echo " - Sudoers configuré: /etc/sudoers.d/automation" # 7) Installation Python3 echo "[7/7] Installation Python3..." if command -v python3 >/dev/null 2>&1; then PYTHON_VERSION=$(python3 --version 2>&1) echo " - Python3 déjà installé: $PYTHON_VERSION" else if [ "$OS_TYPE" = "alpine" ]; then apk add --no-cache python3 echo " - Python3 installé (apk)" elif [ "$OS_TYPE" = "freebsd" ]; then pkg install -y python3 echo " - Python3 installé (pkg)" else apt-get update -qq && apt-get install -y python3 echo " - Python3 installé (apt)" fi fi echo "" echo "=== Bootstrap terminé avec succès ===" echo "Utilisateur: $AUT_USER" echo "HOME: $HOME_DIR" echo "SSH: $HOME_DIR/.ssh/authorized_keys" echo "Sudo: /etc/sudoers.d/automation" """ # Envoyer le script de manière compatible avec tous les shells lines = bootstrap_script.splitlines() def _sh_single_quote(s: str) -> str: """Protège une chaîne pour un shell POSIX en simple quotes.""" return "'" + s.replace("'", "'\"'\"'") + "'" quoted_lines = " ".join(_sh_single_quote(line) for line in lines) remote_cmd = f"printf '%s\\n' {quoted_lines} | sh" rc, out, err = run_ssh_command( host, remote_cmd, ssh_user="root", ssh_password=root_password, ) if rc != 0: raise HTTPException( status_code=500, detail={ "status": "error", "return_code": rc, "stdout": out, "stderr": err, }, ) # Vérification: tester la connexion SSH par clé avec l'utilisateur d'automatisation verify_rc, verify_out, verify_err = run_ssh_command( host, "echo 'ssh_key_ok'", ssh_user=automation_user, ssh_password=None, ) if verify_rc != 0: combined_stdout = (out or "") + f"\n\n[SSH VERIFY] Échec de la connexion par clé pour {automation_user}@{host}\n" + (verify_out or "") combined_stderr = (err or "") + f"\n\n[SSH VERIFY] " + (verify_err or "Aucune erreur détaillée") raise HTTPException( status_code=500, detail={ "status": "error", "return_code": verify_rc, "stdout": combined_stdout, "stderr": combined_stderr, }, ) # Succès complet final_stdout = (out or "") + f"\n\n[SSH VERIFY] Connexion par clé OK pour {automation_user}@{host}" return CommandResult( status="ok", return_code=0, stdout=final_stdout, stderr=err, ) # Base de données hybride : hôtes depuis Ansible, tâches/logs en mémoire class HybridDB: """Base de données qui charge les hôtes depuis l'inventaire Ansible""" def __init__(self, ansible_svc: AnsibleService): self.ansible_service = ansible_svc self._hosts_cache: Optional[List[Host]] = None self._hosts_cache_time: float = 0 self._cache_ttl: float = 60 # Cache de 60 secondes # Statuts runtime des hôtes (en mémoire) rechargés depuis le fichier JSON persistant self._host_runtime_status: Dict[str, Dict[str, Any]] = {} try: persisted_hosts = host_status_service.get_all_status() for host_name, info in persisted_hosts.items(): last_seen_raw = info.get("last_seen") last_seen_dt: Optional[datetime] = None if isinstance(last_seen_raw, str): try: last_seen_dt = datetime.fromisoformat(last_seen_raw.replace("Z", "+00:00")) except Exception: last_seen_dt = None elif isinstance(last_seen_raw, datetime): last_seen_dt = last_seen_raw self._host_runtime_status[host_name] = { "status": info.get("status", "online"), "last_seen": last_seen_dt, "os": info.get("os"), } except Exception: # En cas de problème de lecture, on repartira d'un état en mémoire vierge self._host_runtime_status = {} # Tâches et logs en mémoire (persistés pendant l'exécution) self.tasks: List[Task] = [] self.logs: List[LogEntry] = [ LogEntry(id=1, timestamp=datetime.now(timezone.utc), level="INFO", message="Dashboard démarré - Inventaire Ansible chargé") ] self._id_counters = {"hosts": 100, "tasks": 1, "logs": 2} @property def hosts(self) -> List[Host]: """Charge les hôtes depuis l'inventaire Ansible avec cache""" current_time = time() # Retourner le cache si valide if self._hosts_cache and (current_time - self._hosts_cache_time) < self._cache_ttl: return self._hosts_cache # Recharger depuis Ansible self._hosts_cache = self._load_hosts_from_ansible() self._hosts_cache_time = current_time return self._hosts_cache def _load_hosts_from_ansible(self) -> List[Host]: """Convertit l'inventaire Ansible en liste d'hôtes (sans doublons)""" hosts = [] ansible_hosts = self.ansible_service.get_hosts_from_inventory() # Charger tous les statuts de bootstrap all_bootstrap_status = bootstrap_status_service.get_all_status() for idx, ah in enumerate(ansible_hosts, start=1): # Extraire le groupe principal depuis les groupes primary_group = ah.groups[0] if ah.groups else "unknown" # Récupérer le statut bootstrap pour cet hôte bootstrap_info = all_bootstrap_status.get(ah.name, {}) bootstrap_ok = bootstrap_info.get("bootstrap_ok", False) bootstrap_date_str = bootstrap_info.get("bootstrap_date") bootstrap_date = None if bootstrap_date_str: try: bootstrap_date = datetime.fromisoformat(bootstrap_date_str.replace("Z", "+00:00")) except: pass runtime_status = self._host_runtime_status.get(ah.name, {}) status = runtime_status.get("status", "online") last_seen = runtime_status.get("last_seen") os_label = runtime_status.get("os", f"Linux ({primary_group})") host = Host( id=str(idx), name=ah.name, ip=ah.ansible_host, status=status, os=os_label, last_seen=last_seen, groups=ah.groups, # Tous les groupes de l'hôte bootstrap_ok=bootstrap_ok, bootstrap_date=bootstrap_date ) hosts.append(host) return hosts def refresh_hosts(self): """Force le rechargement des hôtes depuis Ansible""" self._hosts_cache = None return self.hosts def update_host_status(self, host_name: str, status: str, os_info: str = None): """Met à jour le statut d'un hôte après un health-check""" for host in self.hosts: if host.name == host_name: host.status = status host.last_seen = datetime.now(timezone.utc) if os_info: host.os = os_info self._host_runtime_status[host_name] = { "status": host.status, "last_seen": host.last_seen, "os": host.os, } # Persister dans le fichier JSON partagé avec Ansible try: host_status_service.set_status(host_name, host.status, host.last_seen, host.os) except Exception: # Ne pas casser l'exécution si la persistance échoue pass break @property def metrics(self) -> SystemMetrics: """Calcule les métriques en temps réel basées sur les logs de tâches""" hosts = self.hosts # Utiliser les statistiques des fichiers de logs de tâches task_stats = task_log_service.get_stats() total_tasks = task_stats.get("total", 0) completed_tasks = task_stats.get("completed", 0) failed_tasks = task_stats.get("failed", 0) total_finished = completed_tasks + failed_tasks return SystemMetrics( online_hosts=len([h for h in hosts if h.status == "online"]), total_tasks=total_tasks, success_rate=round((completed_tasks / total_finished * 100) if total_finished > 0 else 100, 1), uptime=99.9, cpu_usage=0, memory_usage=0, disk_usage=0 ) def get_next_id(self, collection: str) -> int: self._id_counters[collection] += 1 return self._id_counters[collection] - 1 # Instance globale de la base de données hybride db = HybridDB(ansible_service) # Dépendances FastAPI async def verify_api_key(api_key: str = Depends(api_key_header)) -> bool: """Vérifie la clé API fournie""" if not api_key or api_key != API_KEY: raise HTTPException(status_code=401, detail="Clé API invalide ou manquante") return True # Routes API @app.get("/", response_class=HTMLResponse) async def root(request: Request): """Page principale du dashboard""" return FileResponse(BASE_DIR / "index.html") @app.get("/api", response_class=HTMLResponse) async def api_home(request: Request): """Page d'accueil de l'API Homelab Dashboard""" return """
API REST moderne pour la gestion automatique d'homelab
Explorez les endpoints disponibles et testez les fonctionnalités
/api/hosts
- Liste des hôtes
/api/tasks
- Créer une tâche
/api/metrics
- Métriques système
/ws
- WebSocket temps réel
Version 1.0.0 | Développé avec FastAPI et technologies modernes