Skip to content

NP_2_DF

Download Flojoy Studio to try this app
Infer the type of an array-like DataContainer, then convert it to a DataFrame DataContainer'. Params: default : DataContainer The input numpy array which we apply the conversion to. Returns: out : DataFrame The dataframe result from the conversion of the input.
Python Code
import pandas as pd
import numpy as np
from flojoy import (
    flojoy,
    DataContainer,
    DataFrame,
    OrderedPair,
    OrderedTriple,
    Matrix,
    Grayscale,
    Image,
    ParametricDataFrame,
    ParametricOrderedPair,
    ParametricOrderedTriple,
    ParametricImage,
    ParametricGrayscale,
    ParametricMatrix,
)


@flojoy
def NP_2_DF(default: DataContainer) -> DataFrame:
    """Infer the type of an array-like DataContainer, then convert it to a DataFrame DataContainer'.

    Parameters
    ----------
    default : DataContainer
        The input numpy array which we apply the conversion to.

    Returns
    -------
    DataFrame
        The dataframe result from the conversion of the input.
    """

    match default:
        case DataFrame() | ParametricDataFrame():
            return default

        case OrderedPair() | ParametricOrderedPair():
            df = pd.DataFrame(default.y)
            return DataFrame(df=df)

        case OrderedTriple() | ParametricOrderedTriple():
            df = pd.DataFrame(default.z)
            return DataFrame(df=df)

        case Matrix() | ParametricMatrix():
            np_array = np.asarray(default.m)
            df = pd.DataFrame(np_array)
            return DataFrame(df=df)
        case Grayscale() | ParametricGrayscale():
            np_array = np.asarray(default.m)
            df = pd.DataFrame(np_array)
            return DataFrame(df=df)

        case Image() | ParametricImage():
            red = default.r
            green = default.g
            blue = default.b

            if default.a is None:
                merge = np.stack((red, green, blue), axis=2)
                merge = merge.reshape(-1, merge.shape[-1])
                df = pd.DataFrame(merge)
                return DataFrame(df=df)
            else:
                alpha = default.a
                merge = np.stack((red, green, blue, alpha), axis=2)
                merge = merge.reshape(-1, merge.shape[-1])
                df = pd.DataFrame(merge)
                return DataFrame(df=df)
        case _:
            raise ValueError("unsupported DataContainer type passed for NP_2_DF")

Find this Flojoy Block on GitHub

Example App

Having problems with this example app? Join our Discord community and we will help you out!
React Flow mini map

In this example we use the BASIC_OSCILLATOR node to generate a numpy array. The parameters of the node are set at:

sample_rate: 100 time: 10 waveform: sine amplitude: 1 frequency: 1 offset: 0 phase: 0

With the left TABLE node we can see the x values representing the time when the value was taken and the y represent the values themselves. Then in the left LINE node we can see what the data looks like from time 0 to time 10.

Then we use the NP_2_DF node to convert the input data from the numpy array type to the dataframe type.

With the right TABLE node we can see that we now have only 1 column which is normal since with a dataframe type of data we only use the values since the time is represent by the number of the row at which the value is read in the column. We can observe this also in the right LINE node where we can see the same data but here instead of being from time 0 to time 10 itโ€™s from 0 to 1000. This is due to the fact that we have a sample_rate of 100 meaning that in 1 time measure we take 100 values. Therefore having 10 times means that we have 1000 rows for the y columns.