bbi.files

Read your uploaded files from inside a notebook via bbi.files.

Files are fetched through short-lived presigned URLs minted by the backend after an access check, so no cloud credentials ever reach the notebook. read_csv is a convenience over pandas and read_json parses JSON into Python objects; get_url and download are the universal primitives that work with any format and any reader.

list() → Sequence[UserFile]

List the files you can access: your own plus files shared with your org.

show() → None

Print the files you can access as a table: name, size, visibility, owned.

get_url(name_or_id: str) → str

Return a short-lived presigned download URL for a file (by name or id).

download(name_or_id: str, dest: str | None = None) → str

Download a file to a local path and return that path. Works for any format.

Parameters:

• name_or_id – The file’s name (e.g. "sales.csv") or its id.

• dest – Optional destination path. Defaults to a temp file in the kernel.

read_bytes(name_or_id: str) → bytes

Read a file’s raw content into memory as bytes.

open(name_or_id: str) → BytesIO

Open a file as an in-memory binary stream.

Works with any IO[bytes]-aware reader, e.g. PIL.Image.open(files.open("photo.png")).

read_csv(name_or_id: str, **kwargs: Any) → pandas.DataFrame

Read a CSV file into a pandas DataFrame. kwargs pass to pandas.read_csv.

read_json(name_or_id: str, **kwargs: Any) → Any

Parse a JSON file into Python objects (dict or list).

kwargs pass to json.loads. For a DataFrame, use pandas.read_json(files.get_url(...)).