Function build_dataframe

plotly/express/_core.py:1480–1870

Constructs a dataframe and modifies `args` in-place. The argument values in `args` can be either strings corresponding to existing columns of a dataframe, or data arrays (lists, numpy arrays, pandas columns, series). Parameters ---------- args : OrderedDict arg

(args, constructor)

def build_dataframe(args, constructor):
    """
    Constructs a dataframe and modifies `args` in-place.

    The argument values in `args` can be either strings corresponding to
    existing columns of a dataframe, or data arrays (lists, numpy arrays,
    pandas columns, series).

    Parameters
    ----------
    args : OrderedDict
        arguments passed to the px function and subsequently modified
    constructor : graph_object trace class
        the trace type selected for this figure
    """

    # make copies of all the fields via dict() and list()
    for field in args:
        if field in array_attrables and args[field] is not None:
            if isinstance(args[field], dict):
                args[field] = dict(args[field])
            elif field in ["custom_data", "hover_data"] and isinstance(
                args[field], str
            ):
                args[field] = [args[field]]
            else:
                args[field] = list(args[field])

    # Cast data_frame argument to DataFrame (it could be a numpy array, dict etc.)
    df_provided = args["data_frame"] is not None

    # Flag that indicates if the resulting data_frame after parsing is pandas-like
    # (in terms of resulting Narwhals DataFrame).
    # True if pandas, modin.pandas or cudf DataFrame/Series instance, or converted from
    # PySpark to pandas.
    is_pd_like = False

    # Flag that indicates if data_frame needs to be converted to PyArrow.
    # True if Ibis, DuckDB, Vaex, or implements __dataframe__
    needs_interchanging = False

    # If data_frame is provided, we parse it into a narwhals DataFrame, while accounting
    # for compatibility with pandas specific paths (e.g. Index/MultiIndex case).
    if df_provided:
        # data_frame is pandas-like DataFrame (pandas, modin.pandas, cudf)
        if nw.dependencies.is_pandas_like_dataframe(args["data_frame"]):
            columns = args["data_frame"].columns  # This can be multi index
            args["data_frame"] = nw.from_native(args["data_frame"], eager_only=True)
            is_pd_like = True

        # data_frame is pandas-like Series (pandas, modin.pandas, cudf)
        elif nw.dependencies.is_pandas_like_series(args["data_frame"]):
            args["data_frame"] = nw.from_native(
                args["data_frame"], series_only=True
            ).to_frame()
            columns = args["data_frame"].columns
            is_pd_like = True