MATRIX_VIEW
Take a Matrix or OrderedPair DataContainer object as input, then visualize it in a Plotly table. Params: default : OrderedPair | Matrix the DataContainer to be visualized in matrix format. Returns: out : Plotly the DataContainer containing visualization of the input in matrix format
Python Code
import numpy as np
import plotly.graph_objects as go
from flojoy import DCNpArrayType, Matrix, OrderedPair, Plotly, Vector, flojoy
CELL_SIZE = 50
FONT_SIZE = 10
MAX_ALLOWED_SHAPE = 8
v_dot = "$\\vdots$"
d_dot = "$\\ddots$"
l_dot = "$\\ldots$"
def numpy_2d_array_as_table(
arr: DCNpArrayType,
arr_row_shape: int,
arr_col_shape: int,
placeholder: str,
):
new_arr = arr
if arr_row_shape > MAX_ALLOWED_SHAPE or arr_col_shape > MAX_ALLOWED_SHAPE:
new_arr = np.full(
(MAX_ALLOWED_SHAPE, MAX_ALLOWED_SHAPE), placeholder, dtype=object
)
new_arr[:-2, :-2] = arr[: MAX_ALLOWED_SHAPE - 2, : MAX_ALLOWED_SHAPE - 2]
last_row = arr[arr_row_shape - 1, :]
first_cols = last_row[: MAX_ALLOWED_SHAPE - 2]
new_arr[MAX_ALLOWED_SHAPE - 1, : MAX_ALLOWED_SHAPE - 2] = first_cols
last_col = arr[:, arr.shape[1] - 1]
first_rows = last_col[: MAX_ALLOWED_SHAPE - 2]
new_arr[: MAX_ALLOWED_SHAPE - 2, MAX_ALLOWED_SHAPE - 1] = first_rows
new_arr[MAX_ALLOWED_SHAPE - 1, MAX_ALLOWED_SHAPE - 1 :] = arr[
arr_row_shape - 1, arr.shape[1] - 1 :
]
new_arr[0, MAX_ALLOWED_SHAPE - 2] = l_dot
new_arr[MAX_ALLOWED_SHAPE - 1, MAX_ALLOWED_SHAPE - 2] = l_dot
new_arr[MAX_ALLOWED_SHAPE - 2, 0] = v_dot
new_arr[MAX_ALLOWED_SHAPE - 2, MAX_ALLOWED_SHAPE - 1] = v_dot
return new_arr.T
def numpy_1d_array_as_table(arr: DCNpArrayType):
if arr.size > MAX_ALLOWED_SHAPE:
converted_type = arr.astype(object)
new_arr = converted_type[:MAX_ALLOWED_SHAPE]
new_arr[MAX_ALLOWED_SHAPE - 2] = l_dot
else:
new_arr = arr
return new_arr.reshape(-1, 1)
def numpy_array_as_table(arr: DCNpArrayType):
ndim = arr.ndim
if ndim == 1:
cell_values = numpy_1d_array_as_table(arr)
elif ndim > 2:
raise ValueError("MATRIX_VIEW can process only 2D arrays!")
else:
row_shape, col_shape = arr.shape
cell_values = numpy_2d_array_as_table(arr, row_shape, col_shape, d_dot)
return cell_values
@flojoy
def MATRIX_VIEW(default: OrderedPair | Matrix) -> Plotly:
"""Take a Matrix or OrderedPair DataContainer object as input, then visualize it in a Plotly table.
Parameters
----------
default : OrderedPair | Matrix
the DataContainer to be visualized in matrix format.
Returns
-------
Plotly
the DataContainer containing visualization of the input in matrix format
"""
if isinstance(default, Matrix):
np_arr = default.m
cell_values = numpy_array_as_table(np_arr)
elif isinstance(default, Vector):
np_arr = default.v
cell_values = numpy_array_as_table(np_arr)
else:
np_arr = default.y
cell_values = numpy_array_as_table(np_arr)
fig = go.Figure(
data=[
go.Table(
header=dict(line={"width": 0}, values=[]),
cells=dict(
values=cell_values,
line={"width": 3},
font={"size": FONT_SIZE},
height=CELL_SIZE,
align="center",
format=[".3"],
),
)
]
)
width = MAX_ALLOWED_SHAPE * CELL_SIZE + 80
height = width + 80
fig.layout = go.Layout(
autosize=False,
width=width,
height=height,
margin=dict(l=0, r=0, t=0, b=0),
xaxis=dict(visible=False),
yaxis=dict(visible=False),
hovermode="closest",
font=dict(size=FONT_SIZE),
)
return Plotly(fig=fig)
Example
Having problems with this example app? Join our Discord community and we will help you out!
In this example, MATRIX node generates 8x8 matrix and passes it MATRIX_VIEW node.