import numpy as np
import plotly.graph_objects as go
from plotly.subplots import make_subplots
[docs]def plot(
data: np.ndarray,
scores: np.ndarray,
date: np.ndarray = None,
features: np.ndarray = None,
label: np.ndarray = None,
):
"""Plot data, score and ground truth (if exists).
Args:
data (np.array): Original data stream.
scores (np.array): Anomaly scores of the data stream.
date (np.array, optional): Timestamp of the data. Defaults to None.
features (np.array, optional): Features name. Defaults to None.
label (np.array, optional): Ground truth. Defaults to None.
"""
if features is None:
features = ["f" + str(i) for i in range(np.array(data).shape[1])]
else:
assert (
len(features) == data.shape[1]
), "Number of features must match data dimension."
if date is None:
date = [i for i in range(np.array(data).shape[0])]
else:
assert (
len(date) == data.shape[0]
), "Number of date must match data dimension."
height = 100 * len(features) + 80
row_heights = [100 / height for _ in range(len(features))]
row_heights.append(80 / height)
fig = make_subplots(
rows=len(features) + 1,
cols=1,
shared_xaxes=True,
vertical_spacing=20 / height,
row_heights=row_heights,
)
# Plot data by features
for i, feature in enumerate(features):
fig.add_trace(
go.Scatter(x=date, y=data[:, i], name=str(feature)),
row=i + 1,
col=1,
)
# Plot label by anomalies
if label is not None:
anomalies = date[np.where(label == 1)[0]]
for anomaly in anomalies:
fig.add_vrect(
x0=anomaly,
x1=anomaly,
fillcolor="red",
opacity=0.25,
row="all",
col=1,
name="Anomaly",
)
# Plot score
fig.add_trace(
go.Scatter(x=date, y=scores, name="anomaly score", marker_color="red"),
row=len(features) + 1,
col=1,
)
# fig.update_xaxes(rangeslider={"visible": True}, row=2, col=1)
fig.update_layout(
margin=dict(l=10, r=10, t=10, b=10),
legend=dict(
orientation="h", yanchor="bottom", y=1.02, xanchor="right", x=1
),
height=height,
)
return fig
