feat(storage): add SharePoint integration for document storage

Add Microsoft SharePoint Online as a storage provider using the
Microsoft Graph API. Includes OAuth2 authentication via MSAL,
site/drive resolution, chunked upload sessions, and metadata sync.

Co-authored-by: christianlouis <361235+christianlouis@users.noreply.github.com>

.env.demo

@@ -351,6 +351,15 @@ ONEDRIVE_TENANT_ID=common
 ONEDRIVE_REFRESH_TOKEN=your-refresh-token
 ONEDRIVE_FOLDER_PATH=Documents/Uploads
 
+# SharePoint
+SHAREPOINT_CLIENT_ID=your-client-id
+SHAREPOINT_CLIENT_SECRET=your-client-secret
+SHAREPOINT_TENANT_ID=common
+SHAREPOINT_REFRESH_TOKEN=your-refresh-token
+SHAREPOINT_SITE_URL=https://tenant.sharepoint.com/sites/sitename
+SHAREPOINT_DOCUMENT_LIBRARY=Documents
+SHAREPOINT_FOLDER_PATH=Uploads
+
 # WebDAV
 WEBDAV_URL=https://webdav.example.com/path
 WEBDAV_USERNAME=webdav_user

app/celery_worker.py

@@ -50,6 +50,7 @@ from app.tasks.upload_to_onedrive import upload_to_onedrive  # noqa: F401
 from app.tasks.upload_to_paperless import upload_to_paperless  # noqa: F401
 from app.tasks.upload_to_s3 import upload_to_s3  # noqa: F401
 from app.tasks.upload_to_sftp import upload_to_sftp  # noqa: F401
+from app.tasks.upload_to_sharepoint import upload_to_sharepoint  # noqa: F401
 from app.tasks.upload_to_user_integration import upload_to_user_integration  # noqa: F401
 from app.tasks.upload_to_webdav import upload_to_webdav  # noqa: F401
 from app.tasks.upload_with_rclone import send_to_all_rclone_destinations, upload_with_rclone  # noqa: F401

app/config.py

@@ -457,6 +457,15 @@ class Settings(BaseSettings):
     onedrive_refresh_token: Optional[str] = None  # Required for personal accounts
     onedrive_folder_path: Optional[str] = None
 
+    # SharePoint settings
+    sharepoint_client_id: Optional[str] = None
+    sharepoint_client_secret: Optional[str] = None
+    sharepoint_tenant_id: Optional[str] = "common"
+    sharepoint_refresh_token: Optional[str] = None
+    sharepoint_site_url: Optional[str] = None  # e.g. https://tenant.sharepoint.com/sites/sitename
+    sharepoint_document_library: Optional[str] = "Documents"  # Document library name
+    sharepoint_folder_path: Optional[str] = None  # Subfolder inside the library
+
     # AWS S3 settings
     aws_access_key_id: Optional[str] = None
     aws_secret_access_key: Optional[str] = None

app/models.py

@@ -505,6 +505,7 @@ class IntegrationType:
     EMAIL = "EMAIL"
     PAPERLESS = "PAPERLESS"
     RCLONE = "RCLONE"
+    SHAREPOINT = "SHAREPOINT"
 
     ALL = {
         IMAP,
@@ -521,6 +522,7 @@ class IntegrationType:
         EMAIL,
         PAPERLESS,
         RCLONE,
+        SHAREPOINT,
     }
 
 

app/tasks/send_to_all.py

@@ -17,6 +17,7 @@ from app.tasks.upload_to_onedrive import upload_to_onedrive
 from app.tasks.upload_to_paperless import upload_to_paperless
 from app.tasks.upload_to_s3 import upload_to_s3
 from app.tasks.upload_to_sftp import upload_to_sftp
+from app.tasks.upload_to_sharepoint import upload_to_sharepoint
 from app.tasks.upload_to_webdav import upload_to_webdav
 from app.utils.config_validator import get_provider_status
 from app.utils.logging import log_task_progress
@@ -79,6 +80,18 @@ def _should_upload_to_s3():
     return bool(settings.s3_bucket_name and settings.aws_access_key_id and settings.aws_secret_access_key)
 
 
+def _should_upload_to_sharepoint():
+    return bool(
+        settings.sharepoint_client_id
+        and settings.sharepoint_client_secret
+        and settings.sharepoint_site_url
+        and (
+            settings.sharepoint_refresh_token
+            or (settings.sharepoint_tenant_id and settings.sharepoint_tenant_id != "common")
+        )
+    )
+
+
 def get_configured_services_from_validator():
     """
     Use the config validator to determine which services are configured properly.
@@ -98,6 +111,7 @@ def get_configured_services_from_validator():
         "Email": "email",
         "OneDrive": "onedrive",
         "S3 Storage": "s3",
+        "SharePoint": "sharepoint",
     }
 
     result = {}
@@ -206,6 +220,11 @@ def send_to_all_destinations(self, file_path: str, use_validator=True, file_id:
             "should_upload": _should_upload_to_s3,
             "upload_func": upload_to_s3,
         },
+        {
+            "name": "sharepoint",
+            "should_upload": _should_upload_to_sharepoint,
+            "upload_func": upload_to_sharepoint,
+        },
     ]
 
     # Optionally get configuration status from validator

app/tasks/upload_to_sharepoint.py

@@ -0,0 +1,338 @@
+#!/usr/bin/env python3
+"""Upload documents to Microsoft SharePoint via the Microsoft Graph API.
+
+This module authenticates using MSAL (same OAuth2 flow as OneDrive) and
+uploads files to a configurable SharePoint Online document library using
+the chunked upload session approach for reliability with large files.
+
+Key differences from the OneDrive provider:
+- Uses ``/sites/{siteId}/drives/{driveId}`` instead of ``/me/drive``
+- Requires a SharePoint site URL to resolve the site and drive IDs
+- Targets a named document library (default: ``Documents``)
+"""
+
+import logging
+import os
+import time
+import urllib.parse
+
+import msal
+import requests
+
+from app.celery_app import celery
+from app.config import settings
+from app.tasks.retry_config import UploadTaskWithRetry
+from app.utils import log_task_progress
+
+logger = logging.getLogger(__name__)
+
+
+def get_sharepoint_token() -> str:
+    """Acquire a Microsoft Graph API access token for SharePoint.
+
+    Uses MSAL ``ConfidentialClientApplication`` with the refresh-token flow
+    (delegated permissions) or the client-credentials flow (application
+    permissions) depending on configuration.
+
+    Returns:
+        A valid access token string.
+
+    Raises:
+        ValueError: When required settings are missing or token acquisition fails.
+    """
+    if not settings.sharepoint_client_id or not settings.sharepoint_client_secret:
+        raise ValueError("SharePoint client ID and client secret must be configured")
+
+    tenant = settings.sharepoint_tenant_id or "common"
+    logger.info("Using SharePoint tenant: %s", tenant)
+
+    scopes = ["https://graph.microsoft.com/.default"]
+
+    if settings.sharepoint_refresh_token:
+        app = msal.ConfidentialClientApplication(
+            client_id=settings.sharepoint_client_id,
+            client_credential=settings.sharepoint_client_secret,
+            authority=f"https://login.microsoftonline.com/{tenant}",
+        )
+
+        logger.info("Attempting to acquire SharePoint token using refresh token")
+        token_response = app.acquire_token_by_refresh_token(
+            refresh_token=settings.sharepoint_refresh_token, scopes=scopes
+        )
+
+        if "access_token" not in token_response:
+            error = token_response.get("error", "")
+            error_desc = token_response.get("error_description", "Unknown error")
+            logger.error("Failed to get SharePoint access token: %s - %s", error, error_desc)
+            raise ValueError(f"Failed to get SharePoint access token: {error} - {error_desc}")
+
+        if "refresh_token" in token_response:
+            settings.sharepoint_refresh_token = token_response["refresh_token"]
+            logger.info("Updated SharePoint refresh token in memory")
+
+        return token_response["access_token"]
+
+    elif settings.sharepoint_tenant_id and settings.sharepoint_tenant_id != "common":
+        authority = f"https://login.microsoftonline.com/{settings.sharepoint_tenant_id}"
+        app = msal.ConfidentialClientApplication(
+            client_id=settings.sharepoint_client_id,
+            client_credential=settings.sharepoint_client_secret,
+            authority=authority,
+        )
+
+        token_response = app.acquire_token_for_client(scopes=scopes)
+
+        if "access_token" not in token_response:
+            error = token_response.get("error", "")
+            error_desc = token_response.get("error_description", "Unknown error")
+            raise ValueError(f"Failed to get SharePoint access token: {error} - {error_desc}")
+
+        return token_response["access_token"]
+
+    else:
+        raise ValueError("For SharePoint, either a refresh token or a non-'common' tenant ID is required")
+
+
+def resolve_sharepoint_drive(access_token: str, site_url: str, library_name: str) -> tuple[str, str]:
+    """Resolve the Graph API site ID and drive ID for a SharePoint site.
+
+    Args:
+        access_token: Valid Microsoft Graph API token.
+        site_url: Full SharePoint site URL, e.g.
+            ``https://tenant.sharepoint.com/sites/sitename``.
+        library_name: Display name of the document library (e.g. ``Documents``).
+
+    Returns:
+        A ``(site_id, drive_id)`` tuple.
+
+    Raises:
+        ValueError: When the site URL cannot be parsed.
+        RuntimeError: When the Graph API call fails.
+    """
+    parsed = urllib.parse.urlparse(site_url)
+    hostname = parsed.hostname
+    site_path = parsed.path.rstrip("/")
+
+    if not hostname or not site_path:
+        raise ValueError(
+            f"Invalid SharePoint site URL '{site_url}'. Expected format: https://tenant.sharepoint.com/sites/sitename"
+        )
+
+    headers = {"Authorization": f"Bearer {access_token}"}
+
+    # Resolve site ID
+    site_api_url = f"https://graph.microsoft.com/v1.0/sites/{hostname}:{site_path}"
+    logger.info("Resolving SharePoint site: %s", site_api_url)
+    resp = requests.get(site_api_url, headers=headers, timeout=settings.http_request_timeout)
+
+    if resp.status_code != 200:
+        raise RuntimeError(f"Failed to resolve SharePoint site: {resp.status_code} - {resp.text}")
+
+    site_id = resp.json()["id"]
+    logger.info("Resolved SharePoint site ID: %s", site_id)
+
+    # Resolve drive ID from the document library name
+    drives_url = f"https://graph.microsoft.com/v1.0/sites/{site_id}/drives"
+    resp = requests.get(drives_url, headers=headers, timeout=settings.http_request_timeout)
+
+    if resp.status_code != 200:
+        raise RuntimeError(f"Failed to list SharePoint drives: {resp.status_code} - {resp.text}")
+
+    drives = resp.json().get("value", [])
+    drive_id = None
+    for drive in drives:
+        if drive.get("name", "").lower() == library_name.lower():
+            drive_id = drive["id"]
+            break
+
+    if not drive_id:
+        available = [d.get("name") for d in drives]
+        raise RuntimeError(f"Document library '{library_name}' not found on site. Available libraries: {available}")
+
+    logger.info("Resolved SharePoint drive ID: %s (library: %s)", drive_id, library_name)
+    return site_id, drive_id
+
+
+def create_sharepoint_upload_session(
+    filename: str, folder_path: str | None, drive_id: str, site_id: str, access_token: str
+) -> str:
+    """Create a resumable upload session on a SharePoint document library.
+
+    Args:
+        filename: Name of the file to upload.
+        folder_path: Optional subfolder path inside the library.
+        drive_id: Graph API drive ID of the document library.
+        site_id: Graph API site ID.
+        access_token: Valid access token.
+
+    Returns:
+        The upload session URL for chunked PUT requests.
+
+    Raises:
+        RuntimeError: When session creation fails.
+    """
+    base_url = f"https://graph.microsoft.com/v1.0/sites/{site_id}/drives/{drive_id}"
+
+    if folder_path:
+        folder_path = folder_path.strip("/")
+        path_components = folder_path.split("/")
+        encoded_path = "/".join(urllib.parse.quote(component) for component in path_components)
+        encoded_filename = urllib.parse.quote(filename)
+        item_path = f"/root:/{encoded_path}/{encoded_filename}:/createUploadSession"
+    else:
+        encoded_filename = urllib.parse.quote(filename)
+        item_path = f"/root:/{encoded_filename}:/createUploadSession"
+
+    url = f"{base_url}{item_path}"
+    request_body = {"item": {"@microsoft.graph.conflictBehavior": "replace"}}
+    headers = {"Authorization": f"Bearer {access_token}", "Content-Type": "application/json"}
+
+    logger.info("Creating SharePoint upload session for %s at path %s", filename, folder_path)
+    response = requests.post(url, headers=headers, json=request_body, timeout=settings.http_request_timeout)
+
+    if response.status_code == 200:
+        upload_url = response.json().get("uploadUrl")
+        logger.info("SharePoint upload session created for %s", filename)
+        return upload_url
+    else:
+        raise RuntimeError(f"Failed to create SharePoint upload session: {response.status_code} - {response.text}")
+
+
+def upload_large_file_sharepoint(file_path: str, upload_url: str) -> dict:
+    """Upload a file to SharePoint using a chunked upload session.
+
+    Args:
+        file_path: Local path to the file.
+        upload_url: The upload session URL from ``create_sharepoint_upload_session``.
+
+    Returns:
+        The Graph API response dict containing file metadata.
+
+    Raises:
+        RuntimeError: When a chunk upload fails after retries.
+    """
+    file_size = os.path.getsize(file_path)
+    chunk_size = 10 * 1024 * 1024  # 10 MB
+
+    response = None
+    with open(file_path, "rb") as f:
+        chunk_number = 0
+        while True:
+            chunk = f.read(chunk_size)
+            if not chunk:
+                break
+
+            chunk_start = chunk_number * chunk_size
+            chunk_end = chunk_start + len(chunk) - 1
+            content_range = f"bytes {chunk_start}-{chunk_end}/{file_size}"
+
+            headers = {"Content-Length": str(len(chunk)), "Content-Range": content_range}
+
+            max_retries = 3
+            retry_delay = 2
+
+            for attempt in range(max_retries):
+                try:
+                    response = requests.put(
+                        upload_url, headers=headers, data=chunk, timeout=settings.http_request_timeout
+                    )
+                    if response.status_code in (201, 202):
+                        break
+                    else:
+                        logger.warning(
+                            "SharePoint chunk upload failed (attempt %d): %d", attempt + 1, response.status_code
+                        )
+                        if attempt < max_retries - 1:
+                            time.sleep(retry_delay * (attempt + 1))
+                except Exception as e:
+                    logger.warning("SharePoint chunk upload error (attempt %d): %s", attempt + 1, str(e))
+                    if attempt < max_retries - 1:
+                        time.sleep(retry_delay * (attempt + 1))
+
+            if response is None or response.status_code not in (201, 202):
+                status = response.status_code if response else "no response"
+                text = response.text if response else ""
+                raise RuntimeError(f"Failed to upload chunk after {max_retries} attempts: {status} - {text}")
+
+            chunk_number += 1
+
+    return response.json() if response else {}
+
+
+@celery.task(base=UploadTaskWithRetry, bind=True)
+def upload_to_sharepoint(self, file_path: str, file_id: int = None, folder_override: str = None):
+    """Upload a file to SharePoint Online.
+
+    Args:
+        file_path: Path to the file to upload.
+        file_id: Optional file ID to associate with logs.
+        folder_override: Optional folder path override.
+
+    Returns:
+        A dict with upload status and file details.
+
+    Raises:
+        FileNotFoundError: When the file does not exist.
+        ValueError: When SharePoint is not configured.
+        RuntimeError: When the upload fails.
+    """
+    task_id = self.request.id
+    logger.info("[%s] Starting SharePoint upload: %s", task_id, file_path)
+    log_task_progress(
+        task_id,
+        "upload_to_sharepoint",
+        "in_progress",
+        f"Uploading to SharePoint: {os.path.basename(file_path)}",
+        file_id=file_id,
+    )
+
+    if not os.path.exists(file_path):
+        error_msg = f"File not found: {file_path}"
+        logger.error("[%s] %s", task_id, error_msg)
+        log_task_progress(task_id, "upload_to_sharepoint", "failure", error_msg, file_id=file_id)
+        raise FileNotFoundError(error_msg)
+
+    filename = os.path.basename(file_path)
+
+    if not settings.sharepoint_client_id:
+        error_msg = "SharePoint client ID is not configured"
+        logger.error("[%s] %s", task_id, error_msg)
+        log_task_progress(task_id, "upload_to_sharepoint", "failure", error_msg, file_id=file_id)
+        raise ValueError(error_msg)
+
+    if not settings.sharepoint_site_url:
+        error_msg = "SharePoint site URL is not configured"
+        logger.error("[%s] %s", task_id, error_msg)
+        log_task_progress(task_id, "upload_to_sharepoint", "failure", error_msg, file_id=file_id)
+        raise ValueError(error_msg)
+
+    try:
+        access_token = get_sharepoint_token()
+
+        library_name = settings.sharepoint_document_library or "Documents"
+        site_id, drive_id = resolve_sharepoint_drive(access_token, settings.sharepoint_site_url, library_name)
+
+        folder_path = folder_override if folder_override is not None else settings.sharepoint_folder_path
+
+        upload_url = create_sharepoint_upload_session(filename, folder_path, drive_id, site_id, access_token)
+        result = upload_large_file_sharepoint(file_path, upload_url)
+
+        web_url = result.get("webUrl", "Not available")
+        logger.info("[%s] Successfully uploaded %s to SharePoint", task_id, filename)
+        logger.info("[%s] File accessible at: %s", task_id, web_url)
+        log_task_progress(
+            task_id, "upload_to_sharepoint", "success", f"Uploaded to SharePoint: {filename}", file_id=file_id
+        )
+
+        return {
+            "status": "Completed",
+            "file_path": file_path,
+            "sharepoint_path": f"{folder_path or ''}/{filename}",
+            "web_url": web_url,
+        }
+
+    except Exception as e:
+        error_msg = f"Failed to upload {filename} to SharePoint: {str(e)}"
+        logger.error("[%s] %s", task_id, error_msg)
+        log_task_progress(task_id, "upload_to_sharepoint", "failure", error_msg, file_id=file_id)
+        raise Exception(error_msg)

app/tasks/upload_to_user_integration.py

@@ -571,6 +571,113 @@ def _upload_rclone(file_path: str, cfg: dict[str, Any], creds: dict[str, Any], t
     return {"status": "Completed", "rclone_dest": dest}
 
 
+def _upload_sharepoint(file_path: str, cfg: dict[str, Any], creds: dict[str, Any], task_id: str) -> dict[str, Any]:
+    """Upload *file_path* to SharePoint using per-user MSAL credentials."""
+    import urllib.parse
+
+    import msal
+    import requests as _requests
+
+    client_id = creds.get("client_id") or ""
+    client_secret = creds.get("client_secret") or ""
+    refresh_token = creds.get("refresh_token") or ""
+    tenant = cfg.get("tenant_id") or "common"
+    site_url = cfg.get("site_url") or ""
+    library_name = cfg.get("document_library") or "Documents"
+    folder_path = cfg.get("folder_path") or ""
+
+    if not (client_id and client_secret):
+        raise ValueError("SharePoint integration is missing client_id or client_secret in credentials")
+    if not site_url:
+        raise ValueError("SharePoint integration is missing site_url in config")
+
+    scopes = ["https://graph.microsoft.com/.default"]
+    msal_app = msal.ConfidentialClientApplication(
+        client_id=client_id,
+        client_credential=client_secret,
+        authority=f"https://login.microsoftonline.com/{tenant}",
+    )
+
+    if refresh_token:
+        token_resp = msal_app.acquire_token_by_refresh_token(refresh_token=refresh_token, scopes=scopes)
+    else:
+        token_resp = msal_app.acquire_token_for_client(scopes=scopes)
+
+    if "access_token" not in token_resp:
+        raise ValueError(f"SharePoint token acquisition failed: {token_resp.get('error_description', 'unknown')}")
+
+    access_token = token_resp["access_token"]
+    headers = {"Authorization": f"Bearer {access_token}"}
+
+    # Resolve site ID
+    parsed = urllib.parse.urlparse(site_url)
+    hostname = parsed.hostname
+    site_path = parsed.path.rstrip("/")
+    if not hostname or not site_path:
+        raise ValueError(f"Invalid SharePoint site URL: {site_url}")
+
+    resp = _requests.get(f"https://graph.microsoft.com/v1.0/sites/{hostname}:{site_path}", headers=headers, timeout=30)
+    resp.raise_for_status()
+    site_id = resp.json()["id"]
+
+    # Resolve drive ID
+    resp = _requests.get(f"https://graph.microsoft.com/v1.0/sites/{site_id}/drives", headers=headers, timeout=30)
+    resp.raise_for_status()
+    drive_id = None
+    for drive in resp.json().get("value", []):
+        if drive.get("name", "").lower() == library_name.lower():
+            drive_id = drive["id"]
+            break
+    if not drive_id:
+        raise RuntimeError(f"Document library '{library_name}' not found on SharePoint site")
+
+    filename = os.path.basename(file_path)
+
+    # Build upload-session URL
+    base_url = f"https://graph.microsoft.com/v1.0/sites/{site_id}/drives/{drive_id}"
+    if folder_path:
+        folder_path = folder_path.strip("/")
+        encoded_path = "/".join(urllib.parse.quote(p) for p in folder_path.split("/"))
+        encoded_file = urllib.parse.quote(filename)
+        item_path = f"/root:/{encoded_path}/{encoded_file}:/createUploadSession"
+    else:
+        encoded_file = urllib.parse.quote(filename)
+        item_path = f"/root:/{encoded_file}:/createUploadSession"
+
+    session_url = f"{base_url}{item_path}"
+    session_headers = {"Authorization": f"Bearer {access_token}", "Content-Type": "application/json"}
+    resp = _requests.post(
+        session_url,
+        headers=session_headers,
+        json={"item": {"@microsoft.graph.conflictBehavior": "replace"}},
+        timeout=30,
+    )
+    resp.raise_for_status()
+    upload_url = resp.json()["uploadUrl"]
+
+    file_size = os.path.getsize(file_path)
+    chunk_size = 10 * 1024 * 1024
+    with open(file_path, "rb") as fh:
+        chunk_num = 0
+        while True:
+            chunk = fh.read(chunk_size)
+            if not chunk:
+                break
+            start = chunk_num * chunk_size
+            end = start + len(chunk) - 1
+            upload_headers = {
+                "Content-Length": str(len(chunk)),
+                "Content-Range": f"bytes {start}-{end}/{file_size}",
+            }
+            upload_resp = _requests.put(upload_url, headers=upload_headers, data=chunk, timeout=120)
+            if upload_resp.status_code not in (201, 202):
+                raise RuntimeError(f"SharePoint chunk upload failed: {upload_resp.status_code}")
+            chunk_num += 1
+
+    logger.info("[%s] SharePoint upload complete: %s/%s", task_id, folder_path, filename)
+    return {"status": "Completed", "sharepoint_folder": folder_path, "filename": filename}
+
+
 # Map IntegrationType → upload helper
 _UPLOAD_HANDLERS = {
     IntegrationType.DROPBOX: _upload_dropbox,
@@ -584,6 +691,7 @@ _UPLOAD_HANDLERS = {
     IntegrationType.PAPERLESS: _upload_paperless,
     IntegrationType.EMAIL: _upload_email,
     IntegrationType.RCLONE: _upload_rclone,
+    IntegrationType.SHAREPOINT: _upload_sharepoint,
 }
 
 

app/utils/config_validator/providers.py

@@ -296,6 +296,28 @@ def get_provider_status() -> dict[str, dict[str, object]]:
         },
     }
 
+    # Check SharePoint configuration
+    providers["SharePoint"] = {
+        "name": "SharePoint",
+        "icon": "fa-brands fa-microsoft",
+        "configured": bool(
+            getattr(settings, "sharepoint_client_id", None)
+            and getattr(settings, "sharepoint_client_secret", None)
+            and getattr(settings, "sharepoint_site_url", None)
+        ),
+        "enabled": True,
+        "description": "Store documents in Microsoft SharePoint Online",
+        "details": {
+            "client_id": getattr(settings, "sharepoint_client_id", "Not set"),
+            "client_secret": mask_sensitive_value(getattr(settings, "sharepoint_client_secret", None)),
+            "tenant_id": getattr(settings, "sharepoint_tenant_id", "Not set"),
+            "refresh_token": mask_sensitive_value(getattr(settings, "sharepoint_refresh_token", None)),
+            "site_url": getattr(settings, "sharepoint_site_url", "Not set"),
+            "document_library": getattr(settings, "sharepoint_document_library", "Not set"),
+            "folder_path": getattr(settings, "sharepoint_folder_path", "Not set"),
+        },
+    }
+
     # Check S3 configuration
     providers["S3 Storage"] = {
         "name": "S3 Storage",

app/utils/settings_service.py

@@ -675,6 +675,63 @@ SETTING_METADATA = {
         "required": False,
         "restart_required": False,
     },
+    # Storage Providers - SharePoint
+    "sharepoint_client_id": {
+        "category": "Storage Providers",
+        "description": "SharePoint Azure AD application (client) ID",
+        "type": "string",
+        "sensitive": False,
+        "required": False,
+        "restart_required": False,
+    },
+    "sharepoint_client_secret": {
+        "category": "Storage Providers",
+        "description": "SharePoint Azure AD client secret",
+        "type": "string",
+        "sensitive": True,
+        "required": False,
+        "restart_required": False,
+    },
+    "sharepoint_tenant_id": {
+        "category": "Storage Providers",
+        "description": "SharePoint Azure AD tenant ID (use 'common' for multi-tenant apps)",
+        "type": "string",
+        "sensitive": False,
+        "required": False,
+        "restart_required": False,
+    },
+    "sharepoint_refresh_token": {
+        "category": "Storage Providers",
+        "description": "SharePoint OAuth refresh token",
+        "type": "string",
+        "sensitive": True,
+        "required": False,
+        "restart_required": False,
+    },
+    "sharepoint_site_url": {
+        "category": "Storage Providers",
+        "description": "SharePoint site URL (e.g. https://tenant.sharepoint.com/sites/sitename)",
+        "type": "string",
+        "sensitive": False,
+        "required": False,
+        "restart_required": False,
+    },
+    "sharepoint_document_library": {
+        "category": "Storage Providers",
+        "description": "SharePoint document library name (default: 'Documents')",
+        "type": "string",
+        "sensitive": False,
+        "required": False,
+        "restart_required": False,
+    },
+    "sharepoint_folder_path": {
+        "category": "Storage Providers",
+        "description": "Subfolder path inside the SharePoint document library",
+        "type": "string",
+        "sensitive": False,
+        "required": False,
+        "restart_required": False,
+    },
     # Storage Providers - WebDAV
     "webdav_url": {
         "category": "Storage Providers",
@@ -1618,14 +1675,26 @@ SETTING_METADATA = {
         "category": "Backup",
         "description": (
             "Storage provider for remote backup copies. "
-            "Accepted values: s3, dropbox, google_drive, onedrive, nextcloud, webdav, ftp, sftp, email. "
+            "Accepted values: s3, dropbox, google_drive, onedrive, sharepoint, nextcloud, webdav, ftp, sftp, email. "
             "Leave empty to keep backups local only."
         ),
         "type": "string",
         "sensitive": False,
         "required": False,
         "restart_required": False,
-        "options": ["", "s3", "dropbox", "google_drive", "onedrive", "nextcloud", "webdav", "ftp", "sftp", "email"],
+        "options": [
+            "",
+            "s3",
+            "dropbox",
+            "google_drive",
+            "onedrive",
+            "sharepoint",
+            "nextcloud",
+            "webdav",
+            "ftp",
+            "sftp",
+            "email",
+        ],
     },
     "backup_remote_folder": {
         "category": "Backup",

frontend/templates/files.html

@@ -525,6 +525,7 @@
           <option value="webdav" {% if storage_provider == "webdav" %}selected{% endif %}>WebDAV</option>
           <option value="ftp" {% if storage_provider == "ftp" %}selected{% endif %}>FTP</option>
           <option value="sftp" {% if storage_provider == "sftp" %}selected{% endif %}>SFTP</option>
+          <option value="sharepoint" {% if storage_provider == "sharepoint" %}selected{% endif %}>SharePoint</option>
         </select>
       </div>
 

frontend/templates/help.html

@@ -33,7 +33,7 @@
       "name": "Which cloud storage providers does DocuElevate support?",
       "acceptedAnswer": {
         "@type": "Answer",
-        "text": "DocuElevate integrates with Dropbox, Google Drive, OneDrive, Amazon S3, Nextcloud, WebDAV, SFTP, FTP, and Paperless-ngx."
+        "text": "DocuElevate integrates with Dropbox, Google Drive, OneDrive, SharePoint, Amazon S3, Nextcloud, WebDAV, SFTP, FTP, and Paperless-ngx."
       }
     },
     {
@@ -196,6 +196,14 @@
         </div>
       </div>
 
+      <div class="flex items-start bg-white shadow rounded-lg p-5">
+        <i class="fab fa-microsoft text-purple-600 text-xl mr-3 mt-0.5" aria-hidden="true"></i>
+        <div>
+          <h3 class="font-semibold text-gray-800 text-sm">SharePoint</h3>
+          <p class="text-gray-500 text-xs">Upload to SharePoint Online document libraries via Graph API.</p>
+        </div>
+      </div>
+
       <div class="flex items-start bg-white shadow rounded-lg p-5">
         <i class="fab fa-aws text-yellow-600 text-xl mr-3 mt-0.5" aria-hidden="true"></i>
         <div>

tests/test_upload_to_sharepoint.py

@@ -0,0 +1,554 @@
+"""
+Tests for app/tasks/upload_to_sharepoint.py module.
+
+Covers get_sharepoint_token, resolve_sharepoint_drive,
+create_sharepoint_upload_session, upload_large_file_sharepoint,
+and upload_to_sharepoint Celery task.
+"""
+
+from unittest.mock import Mock, patch
+
+import pytest
+
+
+@pytest.mark.unit
+class TestGetSharepointToken:
+    """Tests for get_sharepoint_token function."""
+
+    @patch("app.tasks.upload_to_sharepoint.msal.ConfidentialClientApplication")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_refresh_token_flow(self, mock_settings, mock_msal):
+        """Test token acquisition using refresh token."""
+        from app.tasks.upload_to_sharepoint import get_sharepoint_token
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "client-secret"
+        mock_settings.sharepoint_refresh_token = "refresh-token"
+        mock_settings.sharepoint_tenant_id = "common"
+
+        mock_app = Mock()
+        mock_app.acquire_token_by_refresh_token.return_value = {
+            "access_token": "new-access-token",
+        }
+        mock_msal.return_value = mock_app
+
+        token = get_sharepoint_token()
+        assert token == "new-access-token"
+
+    @patch("app.tasks.upload_to_sharepoint.msal.ConfidentialClientApplication")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_refresh_token_updates_new_token(self, mock_settings, mock_msal):
+        """Test that a new refresh token updates settings."""
+        from app.tasks.upload_to_sharepoint import get_sharepoint_token
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "client-secret"
+        mock_settings.sharepoint_refresh_token = "old-refresh-token"
+        mock_settings.sharepoint_tenant_id = "common"
+
+        mock_app = Mock()
+        mock_app.acquire_token_by_refresh_token.return_value = {
+            "access_token": "access-token",
+            "refresh_token": "new-refresh-token",
+        }
+        mock_msal.return_value = mock_app
+
+        get_sharepoint_token()
+        assert mock_settings.sharepoint_refresh_token == "new-refresh-token"
+
+    @patch("app.tasks.upload_to_sharepoint.msal.ConfidentialClientApplication")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_refresh_token_failure(self, mock_settings, mock_msal):
+        """Test error handling when refresh token fails."""
+        from app.tasks.upload_to_sharepoint import get_sharepoint_token
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "client-secret"
+        mock_settings.sharepoint_refresh_token = "expired-token"
+        mock_settings.sharepoint_tenant_id = "common"
+
+        mock_app = Mock()
+        mock_app.acquire_token_by_refresh_token.return_value = {
+            "error": "invalid_grant",
+            "error_description": "Token expired",
+        }
+        mock_msal.return_value = mock_app
+
+        with pytest.raises(ValueError, match="Failed to get SharePoint access token"):
+            get_sharepoint_token()
+
+    @patch("app.tasks.upload_to_sharepoint.msal.ConfidentialClientApplication")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_client_credentials_flow(self, mock_settings, mock_msal):
+        """Test token acquisition using client credentials (org accounts)."""
+        from app.tasks.upload_to_sharepoint import get_sharepoint_token
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "client-secret"
+        mock_settings.sharepoint_refresh_token = ""
+        mock_settings.sharepoint_tenant_id = "org-tenant-id"
+
+        mock_app = Mock()
+        mock_app.acquire_token_for_client.return_value = {
+            "access_token": "client-cred-token",
+        }
+        mock_msal.return_value = mock_app
+
+        token = get_sharepoint_token()
+        assert token == "client-cred-token"
+
+    @patch("app.tasks.upload_to_sharepoint.msal.ConfidentialClientApplication")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_client_credentials_failure(self, mock_settings, mock_msal):
+        """Test error handling when client credentials flow fails."""
+        from app.tasks.upload_to_sharepoint import get_sharepoint_token
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "client-secret"
+        mock_settings.sharepoint_refresh_token = ""
+        mock_settings.sharepoint_tenant_id = "org-tenant-id"
+
+        mock_app = Mock()
+        mock_app.acquire_token_for_client.return_value = {
+            "error": "unauthorized_client",
+            "error_description": "Not authorized",
+        }
+        mock_msal.return_value = mock_app
+
+        with pytest.raises(ValueError, match="Failed to get SharePoint access token"):
+            get_sharepoint_token()
+
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_missing_client_id(self, mock_settings):
+        """Test error when client ID is missing."""
+        from app.tasks.upload_to_sharepoint import get_sharepoint_token
+
+        mock_settings.sharepoint_client_id = ""
+        mock_settings.sharepoint_client_secret = "secret"
+
+        with pytest.raises(ValueError, match="client ID and client secret"):
+            get_sharepoint_token()
+
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_no_refresh_token_common_tenant(self, mock_settings):
+        """Test error for common tenant without refresh token."""
+        from app.tasks.upload_to_sharepoint import get_sharepoint_token
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "secret"
+        mock_settings.sharepoint_refresh_token = ""
+        mock_settings.sharepoint_tenant_id = "common"
+
+        with pytest.raises(ValueError, match="either a refresh token or a non-'common' tenant ID"):
+            get_sharepoint_token()
+
+
+@pytest.mark.unit
+class TestResolveSharepointDrive:
+    """Tests for resolve_sharepoint_drive function."""
+
+    @patch("app.tasks.upload_to_sharepoint.requests.get")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_successful_resolution(self, mock_settings, mock_get):
+        """Test successful site and drive resolution."""
+        from app.tasks.upload_to_sharepoint import resolve_sharepoint_drive
+
+        mock_settings.http_request_timeout = 30
+
+        site_resp = Mock()
+        site_resp.status_code = 200
+        site_resp.json.return_value = {"id": "site-id-123"}
+
+        drives_resp = Mock()
+        drives_resp.status_code = 200
+        drives_resp.json.return_value = {
+            "value": [
+                {"id": "drive-1", "name": "Documents"},
+                {"id": "drive-2", "name": "Site Assets"},
+            ]
+        }
+
+        mock_get.side_effect = [site_resp, drives_resp]
+
+        site_id, drive_id = resolve_sharepoint_drive(
+            "access-token", "https://tenant.sharepoint.com/sites/mysite", "Documents"
+        )
+
+        assert site_id == "site-id-123"
+        assert drive_id == "drive-1"
+
+    @patch("app.tasks.upload_to_sharepoint.requests.get")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_library_not_found(self, mock_settings, mock_get):
+        """Test error when document library is not found."""
+        from app.tasks.upload_to_sharepoint import resolve_sharepoint_drive
+
+        mock_settings.http_request_timeout = 30
+
+        site_resp = Mock()
+        site_resp.status_code = 200
+        site_resp.json.return_value = {"id": "site-id-123"}
+
+        drives_resp = Mock()
+        drives_resp.status_code = 200
+        drives_resp.json.return_value = {
+            "value": [
+                {"id": "drive-1", "name": "Documents"},
+            ]
+        }
+
+        mock_get.side_effect = [site_resp, drives_resp]
+
+        with pytest.raises(RuntimeError, match="not found on site"):
+            resolve_sharepoint_drive("access-token", "https://tenant.sharepoint.com/sites/mysite", "NonExistentLibrary")
+
+    @patch("app.tasks.upload_to_sharepoint.requests.get")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_site_resolution_failure(self, mock_settings, mock_get):
+        """Test error when site resolution fails."""
+        from app.tasks.upload_to_sharepoint import resolve_sharepoint_drive
+
+        mock_settings.http_request_timeout = 30
+
+        site_resp = Mock()
+        site_resp.status_code = 404
+        site_resp.text = "Site not found"
+
+        mock_get.return_value = site_resp
+
+        with pytest.raises(RuntimeError, match="Failed to resolve SharePoint site"):
+            resolve_sharepoint_drive("access-token", "https://tenant.sharepoint.com/sites/nonexistent", "Documents")
+
+    def test_invalid_site_url(self):
+        """Test error with invalid site URL."""
+        from app.tasks.upload_to_sharepoint import resolve_sharepoint_drive
+
+        with pytest.raises(ValueError, match="Invalid SharePoint site URL"):
+            resolve_sharepoint_drive("access-token", "not-a-url", "Documents")
+
+    @patch("app.tasks.upload_to_sharepoint.requests.get")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_case_insensitive_library_match(self, mock_settings, mock_get):
+        """Test that library name matching is case-insensitive."""
+        from app.tasks.upload_to_sharepoint import resolve_sharepoint_drive
+
+        mock_settings.http_request_timeout = 30
+
+        site_resp = Mock()
+        site_resp.status_code = 200
+        site_resp.json.return_value = {"id": "site-id"}
+
+        drives_resp = Mock()
+        drives_resp.status_code = 200
+        drives_resp.json.return_value = {
+            "value": [
+                {"id": "drive-1", "name": "Shared Documents"},
+            ]
+        }
+
+        mock_get.side_effect = [site_resp, drives_resp]
+
+        site_id, drive_id = resolve_sharepoint_drive(
+            "access-token", "https://tenant.sharepoint.com/sites/mysite", "shared documents"
+        )
+
+        assert drive_id == "drive-1"
+
+
+@pytest.mark.unit
+class TestCreateSharepointUploadSession:
+    """Tests for create_sharepoint_upload_session function."""
+
+    @patch("app.tasks.upload_to_sharepoint.requests.post")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_successful_session_creation(self, mock_settings, mock_post):
+        """Test successful upload session creation."""
+        from app.tasks.upload_to_sharepoint import create_sharepoint_upload_session
+
+        mock_settings.http_request_timeout = 30
+
+        mock_response = Mock()
+        mock_response.status_code = 200
+        mock_response.json.return_value = {"uploadUrl": "https://upload.url/session123"}
+        mock_post.return_value = mock_response
+
+        url = create_sharepoint_upload_session("test.pdf", "Uploads", "drive-id", "site-id", "access-token")
+
+        assert url == "https://upload.url/session123"
+
+    @patch("app.tasks.upload_to_sharepoint.requests.post")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_session_without_folder(self, mock_settings, mock_post):
+        """Test upload session creation without folder path."""
+        from app.tasks.upload_to_sharepoint import create_sharepoint_upload_session
+
+        mock_settings.http_request_timeout = 30
+
+        mock_response = Mock()
+        mock_response.status_code = 200
+        mock_response.json.return_value = {"uploadUrl": "https://upload.url/session456"}
+        mock_post.return_value = mock_response
+
+        url = create_sharepoint_upload_session("test.pdf", None, "drive-id", "site-id", "access-token")
+
+        assert url == "https://upload.url/session456"
+
+    @patch("app.tasks.upload_to_sharepoint.requests.post")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_session_creation_failure(self, mock_settings, mock_post):
+        """Test error handling when session creation fails."""
+        from app.tasks.upload_to_sharepoint import create_sharepoint_upload_session
+
+        mock_settings.http_request_timeout = 30
+
+        mock_response = Mock()
+        mock_response.status_code = 403
+        mock_response.text = "Access denied"
+        mock_post.return_value = mock_response
+
+        with pytest.raises(RuntimeError, match="Failed to create SharePoint upload session"):
+            create_sharepoint_upload_session("test.pdf", "Uploads", "drive-id", "site-id", "access-token")
+
+    @patch("app.tasks.upload_to_sharepoint.requests.post")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_url_encoding_special_characters(self, mock_settings, mock_post):
+        """Test that special characters in folder path are URL-encoded."""
+        from app.tasks.upload_to_sharepoint import create_sharepoint_upload_session
+
+        mock_settings.http_request_timeout = 30
+
+        mock_response = Mock()
+        mock_response.status_code = 200
+        mock_response.json.return_value = {"uploadUrl": "https://upload.url/session"}
+        mock_post.return_value = mock_response
+
+        create_sharepoint_upload_session("file with spaces.pdf", "My Documents/Uploads", "drive-id", "site-id", "token")
+
+        call_url = mock_post.call_args[0][0]
+        assert "My%20Documents" in call_url
+        assert "file%20with%20spaces.pdf" in call_url
+
+
+@pytest.mark.unit
+class TestUploadLargeFileSharepoint:
+    """Tests for upload_large_file_sharepoint function."""
+
+    @patch("app.tasks.upload_to_sharepoint.requests.put")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_small_single_chunk_upload(self, mock_settings, mock_put, tmp_path):
+        """Test uploading a file that fits in a single chunk."""
+        from app.tasks.upload_to_sharepoint import upload_large_file_sharepoint
+
+        mock_settings.http_request_timeout = 30
+
+        test_file = tmp_path / "small.pdf"
+        test_file.write_bytes(b"small content")
+
+        mock_response = Mock()
+        mock_response.status_code = 201
+        mock_response.json.return_value = {"id": "file123", "name": "small.pdf"}
+        mock_put.return_value = mock_response
+
+        result = upload_large_file_sharepoint(str(test_file), "https://upload.url/session")
+
+        assert result["id"] == "file123"
+
+    @patch("app.tasks.upload_to_sharepoint.time.sleep")
+    @patch("app.tasks.upload_to_sharepoint.requests.put")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_chunk_upload_retry_on_failure(self, mock_settings, mock_put, mock_sleep, tmp_path):
+        """Test retry logic when a chunk upload fails."""
+        from app.tasks.upload_to_sharepoint import upload_large_file_sharepoint
+
+        mock_settings.http_request_timeout = 30
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        mock_fail = Mock()
+        mock_fail.status_code = 500
+
+        mock_success = Mock()
+        mock_success.status_code = 201
+        mock_success.json.return_value = {"id": "file123"}
+
+        mock_put.side_effect = [mock_fail, mock_success]
+
+        result = upload_large_file_sharepoint(str(test_file), "https://upload.url/session")
+
+        assert result["id"] == "file123"
+
+    @patch("app.tasks.upload_to_sharepoint.time.sleep")
+    @patch("app.tasks.upload_to_sharepoint.requests.put")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_chunk_upload_retry_on_exception(self, mock_settings, mock_put, mock_sleep, tmp_path):
+        """Test retry logic when an exception occurs during upload."""
+        from app.tasks.upload_to_sharepoint import upload_large_file_sharepoint
+
+        mock_settings.http_request_timeout = 30
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        mock_success = Mock()
+        mock_success.status_code = 201
+        mock_success.json.return_value = {"id": "file123"}
+
+        mock_put.side_effect = [Exception("Network error"), mock_success]
+
+        result = upload_large_file_sharepoint(str(test_file), "https://upload.url/session")
+
+        assert result["id"] == "file123"
+
+    @patch("app.tasks.upload_to_sharepoint.time.sleep")
+    @patch("app.tasks.upload_to_sharepoint.requests.put")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_all_retries_exhausted(self, mock_settings, mock_put, mock_sleep, tmp_path):
+        """Test that exhausting all retries raises an exception."""
+        from app.tasks.upload_to_sharepoint import upload_large_file_sharepoint
+
+        mock_settings.http_request_timeout = 30
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        mock_fail = Mock()
+        mock_fail.status_code = 500
+        mock_fail.text = "Server Error"
+        mock_put.return_value = mock_fail
+
+        with pytest.raises(RuntimeError, match="Failed to upload chunk"):
+            upload_large_file_sharepoint(str(test_file), "https://upload.url/session")
+
+
+@pytest.mark.unit
+class TestUploadToSharepoint:
+    """Tests for upload_to_sharepoint Celery task."""
+
+    @patch("app.tasks.upload_to_sharepoint.log_task_progress")
+    def test_file_not_found(self, mock_log):
+        """Test that missing file raises FileNotFoundError."""
+        from app.tasks.upload_to_sharepoint import upload_to_sharepoint
+
+        with pytest.raises(FileNotFoundError):
+            upload_to_sharepoint.__wrapped__("/nonexistent/file.pdf", file_id=1)
+
+    @patch("app.tasks.upload_to_sharepoint.log_task_progress")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_missing_client_id(self, mock_settings, mock_log, tmp_path):
+        """Test error when SharePoint client ID is not configured."""
+        from app.tasks.upload_to_sharepoint import upload_to_sharepoint
+
+        mock_settings.sharepoint_client_id = ""
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        with pytest.raises(ValueError, match="client ID is not configured"):
+            upload_to_sharepoint.__wrapped__(str(test_file), file_id=1)
+
+    @patch("app.tasks.upload_to_sharepoint.log_task_progress")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_missing_site_url(self, mock_settings, mock_log, tmp_path):
+        """Test error when SharePoint site URL is not configured."""
+        from app.tasks.upload_to_sharepoint import upload_to_sharepoint
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_site_url = ""
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        with pytest.raises(ValueError, match="site URL is not configured"):
+            upload_to_sharepoint.__wrapped__(str(test_file), file_id=1)
+
+    @patch("app.tasks.upload_to_sharepoint.upload_large_file_sharepoint")
+    @patch("app.tasks.upload_to_sharepoint.create_sharepoint_upload_session")
+    @patch("app.tasks.upload_to_sharepoint.resolve_sharepoint_drive")
+    @patch("app.tasks.upload_to_sharepoint.get_sharepoint_token")
+    @patch("app.tasks.upload_to_sharepoint.log_task_progress")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_successful_upload(
+        self, mock_settings, mock_log, mock_token, mock_resolve, mock_session, mock_upload, tmp_path
+    ):
+        """Test successful SharePoint upload."""
+        from app.tasks.upload_to_sharepoint import upload_to_sharepoint
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "secret"
+        mock_settings.sharepoint_refresh_token = "token"
+        mock_settings.sharepoint_site_url = "https://tenant.sharepoint.com/sites/mysite"
+        mock_settings.sharepoint_document_library = "Documents"
+        mock_settings.sharepoint_folder_path = "Uploads"
+        mock_settings.sharepoint_tenant_id = "common"
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        mock_token.return_value = "access-token"
+        mock_resolve.return_value = ("site-id", "drive-id")
+        mock_session.return_value = "https://upload.url/session"
+        mock_upload.return_value = {"webUrl": "https://tenant.sharepoint.com/sites/mysite/test.pdf"}
+
+        result = upload_to_sharepoint.apply(args=[str(test_file)], kwargs={"file_id": 1}).get()
+
+        assert result["status"] == "Completed"
+        assert "Uploads" in result["sharepoint_path"]
+        assert result["web_url"] == "https://tenant.sharepoint.com/sites/mysite/test.pdf"
+
+    @patch("app.tasks.upload_to_sharepoint.get_sharepoint_token")
+    @patch("app.tasks.upload_to_sharepoint.log_task_progress")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_upload_exception_handling(self, mock_settings, mock_log, mock_token, tmp_path):
+        """Test that upload errors are properly handled."""
+        from app.tasks.upload_to_sharepoint import upload_to_sharepoint
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "secret"
+        mock_settings.sharepoint_site_url = "https://tenant.sharepoint.com/sites/mysite"
+        mock_settings.sharepoint_folder_path = "Uploads"
+        mock_settings.sharepoint_document_library = "Documents"
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        mock_token.side_effect = ValueError("Token error")
+
+        with pytest.raises(Exception, match="Failed to upload"):
+            upload_to_sharepoint.apply(args=[str(test_file)], kwargs={"file_id": 1}).get()
+
+    @patch("app.tasks.upload_to_sharepoint.upload_large_file_sharepoint")
+    @patch("app.tasks.upload_to_sharepoint.create_sharepoint_upload_session")
+    @patch("app.tasks.upload_to_sharepoint.resolve_sharepoint_drive")
+    @patch("app.tasks.upload_to_sharepoint.get_sharepoint_token")
+    @patch("app.tasks.upload_to_sharepoint.log_task_progress")
+    @patch("app.tasks.upload_to_sharepoint.settings")
+    def test_folder_override(
+        self, mock_settings, mock_log, mock_token, mock_resolve, mock_session, mock_upload, tmp_path
+    ):
+        """Test that folder_override is used instead of settings."""
+        from app.tasks.upload_to_sharepoint import upload_to_sharepoint
+
+        mock_settings.sharepoint_client_id = "client-id"
+        mock_settings.sharepoint_client_secret = "secret"
+        mock_settings.sharepoint_refresh_token = "token"
+        mock_settings.sharepoint_site_url = "https://tenant.sharepoint.com/sites/mysite"
+        mock_settings.sharepoint_document_library = "Documents"
+        mock_settings.sharepoint_folder_path = "DefaultFolder"
+        mock_settings.sharepoint_tenant_id = "common"
+
+        test_file = tmp_path / "test.pdf"
+        test_file.write_bytes(b"test content")
+
+        mock_token.return_value = "access-token"
+        mock_resolve.return_value = ("site-id", "drive-id")
+        mock_session.return_value = "https://upload.url/session"
+        mock_upload.return_value = {"webUrl": "https://example.com/test.pdf"}
+
+        result = upload_to_sharepoint.apply(
+            args=[str(test_file)], kwargs={"file_id": 1, "folder_override": "CustomFolder"}
+        ).get()
+
+        # Verify the session was created with the override folder
+        mock_session.assert_called_once_with("test.pdf", "CustomFolder", "drive-id", "site-id", "access-token")
+        assert result["status"] == "Completed"