descarga_adjuntos_correos_cpe/outlook_utilities.py

import msal
import requests
import warnings
import base64
import os
from datetime import datetime, timedelta
from config_loader import CONFIG

warnings.filterwarnings("ignore")

########################################################
def get_access_token():
    """
    Obtiene un token de acceso de Microsoft Graph usando las credenciales
    configuradas en config.json (Client ID, Client Secret y Tenant ID).
    """
    if not CONFIG or not CONFIG.outlook:
        raise ValueError("La configuración de Outlook no está cargada correctamente.")

    authority = f"https://login.microsoftonline.com/{CONFIG.outlook.tenant_id}"
    
    app = msal.ConfidentialClientApplication(
        CONFIG.outlook.client_id,
        authority=authority,
        client_credential=CONFIG.outlook.client_secret,
        verify=CONFIG.outlook.verify_ssl
    )
    
    scopes = ["https://graph.microsoft.com/.default"]
    result = app.acquire_token_for_client(scopes=scopes)
    
    if "access_token" in result:
        return result["access_token"]
    else:
        error = result.get("error")
        error_description = result.get("error_description")
        correlation_id = result.get("correlation_id")
        raise Exception(f"Error al obtener token: {error} - {error_description} (ID: {correlation_id})")

########################################################
def get_emails(user_email, token, fecha_inicio, fecha_fin):
    """
    Obtiene los correos de una cuenta específica en un rango de fechas.
    Retorna una lista de JSON con: id, fecha (ajustada -5h), sender, subject y has_attach.
    """
    headers = {
        "Authorization": f"Bearer {token}",
        "Content-Type": "application/json"
    }

    dt_inicio_utc = datetime.strptime(fecha_inicio, "%Y-%m-%d") + timedelta(hours=5)
    dt_fin_utc = datetime.strptime(fecha_fin, "%Y-%m-%d") + timedelta(days=1, hours=5)

    filter_query = (
        f"receivedDateTime ge {dt_inicio_utc.strftime('%Y-%m-%dT%H:%M:%SZ')} and "
        f"receivedDateTime lt {dt_fin_utc.strftime('%Y-%m-%dT%H:%M:%SZ')}"
    )

    params = {
        "$filter": filter_query,
        "$select": "id,receivedDateTime,hasAttachments,subject,from",
        "$top": 50
    }

    url = f"https://graph.microsoft.com/v1.0/users/{user_email}/messages"
    all_mails = []

    while url:
        response = requests.get(
            url, 
            headers=headers, 
            params=params,
            verify=CONFIG.outlook.verify_ssl
        )

        if response.status_code != 200:
            raise Exception(f"Error Graph API ({user_email}): {response.status_code} - {response.text}")

        data = response.json()
        messages = data.get("value", [])

        for msg in messages:
            utc_str = msg.get("receivedDateTime")
            dt_utc = datetime.fromisoformat(utc_str.replace("Z", "+00:00"))
            dt_local = dt_utc - timedelta(hours=5)

            from_obj = msg.get("from", {})
            sender = from_obj.get("emailAddress", {}).get("address", "Desconocido")

            all_mails.append({
                "id": msg.get("id"),
                "fecha": dt_local.strftime("%Y-%m-%d %H:%M:%S"),
                "sender": sender,
                "subject": msg.get("subject", ""),
                "has_attach": msg.get("hasAttachments")
            })

        url = data.get("@odata.nextLink")
        params = None 

    return all_mails

########################################################
def download_attachments(user_email, message_id, token, download_path):
    """
    Descarga los adjuntos de un mensaje específico y los guarda en la ruta indicada.
    Retorna una lista con los nombres de los archivos descargados.
    """
    if not os.path.exists(download_path):
        os.makedirs(download_path, exist_ok=True)

    headers = {
        "Authorization": f"Bearer {token}",
        "Content-Type": "application/json"
    }

    url = f"https://graph.microsoft.com/v1.0/users/{user_email}/messages/{message_id}/attachments"
    
    response = requests.get(url, headers=headers, verify=CONFIG.outlook.verify_ssl)
    
    if response.status_code != 200:
        raise Exception(f"Error al obtener adjuntos ({message_id}): {response.status_code} - {response.text}")

    attachments = response.json().get("value", [])
    downloaded_files = []

    for attachment in attachments:
        if attachment.get("@odata.type") == "#microsoft.graph.fileAttachment":
            file_name = attachment.get("name")
            content_bytes = attachment.get("contentBytes")
            
            if content_bytes:
                file_path = os.path.join(download_path, file_name)
                with open(file_path, "wb") as f:
                    f.write(base64.b64decode(content_bytes))
                downloaded_files.append(file_name)
    
    return downloaded_files

########################################################
if __name__ == "__main__":
    try:
        print("Intentando obtener token de acceso...")
        token = get_access_token()
        print("✅ Token obtenido exitosamente.")
        
        if CONFIG.mails:
            test_email = CONFIG.mails[0].email
            hoy = datetime.now()
            ayer = hoy - timedelta(days=1)
            f_inicio = ayer.strftime("%Y-%m-%d")
            f_fin = hoy.strftime("%Y-%m-%d")
            
            print(f"\nBuscando correos para {test_email}...")
            emails = get_emails(test_email, token, f_inicio, f_fin)
            print(f"✅ Se encontraron {len(emails)} correos.")
            
            for e in emails:
                if e['has_attach']:
                    print(f"Probando descarga para correo ID: {e['id'][:10]}...")
                    archivos = download_attachments(test_email, e['id'], token, "./test_download")
                    print(f"Descargados: {archivos}")
                    break
    except Exception as e:
        print(f"❌ Error: {e}")