OneDrive Personal SDK

OneDrive Personal SDK is a Python library for interacting with a personal OneDrive through the Graph API.

Usage

Getting an authentication token

The library is built to support different token providers. To add authentication to the library you need to define a get_access_token method and pass that to the library.

To use msal as a token provider you can create that function like the following:

from msal import PublicClientApplication

app = PublicClientApplication(
    "Client ID",
    authority="https://login.microsoftonline.com/consumers",
)


async def get_access_token(self) -> str:
    result = app.acquire_token_interactive(
        scopes=[
            "Files.ReadWrite.All",
        ]
    )
    return result["access_token"]

Creating a client

To create a client you need to provide a function to retrieve an access token.

from onedrive_personal_sdk import OneDriveClient

client = OneDriveClient(get_access_token)

# can also be created with a custom aiohttp session
client = OneDriveClient(get_access_token, session=session)

Calling the API

The client provides methods to interact with the OneDrive API. The methods are async and return the response from the API. Most methods accept item ids or paths as arguments.

root = await client.get_drive_item("root") # Get the root folder
item = await client.get_drive_item("root:/path/to/item:") # Get an item by path
item = await client.get_drive_item(item.id) # Get an item by id

folder = await client.get_drive_item("root:/path/to/folder:") # Get a folder

# List children of a folder
children = await client.list_children("root:/path/to/folder:")
children = await client.list_children(folder.id)

Uploading files

Smaller files (<250MB) can be uploaded with the upload method. All files need to be provided as AsyncIterators. Such a generator can be created from aiofiles.

import aiofiles
from collections.abc import AsyncIterator

async def file_reader(file_path: str) -> AsyncIterator[bytes]:
    async with aiofiles.open(file_path, "rb") as f:
        while True:
            chunk = await f.read(1024)  # Read in chunks of 1024 bytes
            if not chunk:
                break
            yield chunk

await client.upload_file("root:/TestFolder", "test.txt", file_reader("test.txt"))

Larger files (and small files as well) can be uploaded with a LargeFileUploadClient. The client will handle the upload in chunks.

import os
from onedrive_personal_sdk import LargeFileUploadClient
from onedrive_personal_sdk.models.upload import FileInfo

filename = "testfile.txt"
size = os.path.getsize(filename)

file = FileInfo(
    name=filename,
    size=size,
    folder_path_id="root",
    content_stream=file_reader(filename),
)
file = await LargeFileUploadClient.upload(get_access_token, file)

Adaptive Chunk Size

The LargeFileUploadClient supports adaptive chunk sizing to optimize upload performance based on your connection speed. When enabled, the client dynamically adjusts the chunk size to target approximately 5 seconds per chunk upload, keeping it between 10MB and 60MB by default.

file = await LargeFileUploadClient.upload(
    get_access_token,
    file,
    smart_chunk_size=True,  # Enable adaptive chunk sizing
    upload_chunk_size=320 * 1024,  # Initial chunk size (optional, default: 5.2MB)
    min_chunk_size=32 * 320 * 1024,  # Smallest adapted chunk size (optional, default: 10MB)
    max_chunk_size=60 * 1024 * 1024,  # Largest adapted chunk size (optional, default: 60MB)
    target_chunk_duration=5.0,  # Target seconds per chunk (optional, default: 5.0)
)

All chunk sizes must be multiples of 320kB, and the maximum must not exceed 60MB.