How to 'hold' values instead of linearly interpolating with a mark_line

[kevinlinxc]

Here is my minimal reproducible example:

# main.py
# pip install altair pandas streamlit
# streamlit run main.py
import altair as alt
import pandas as pd
import datetime as dt
import streamlit as st

now = dt.datetime.now()
yesterday = now - dt.timedelta(days=1)
ereyesterday = yesterday - dt.timedelta(days=1)

d = {'Date': [now, yesterday, ereyesterday], 'data1': [109, 34, 109]}
df = pd.DataFrame(data=d)
df.set_index('Date', inplace=True)


def get_alt(df, signal):
    source = df.reset_index(names="Date").melt(
        id_vars="Date", var_name="sig_value", value_name="value", value_vars=signal
    )
    return alt.Chart(source).mark_line().encode(x='Date:T', y='value:Q')


graph_dict: alt.Chart = get_alt(df, 'data1')

# show the chart - streamlit run main.py
st.altair_chart(graph_dict)

This creates the following, where the line is drawn from point to point:

I'm wondering if it's possible to instead make something like this, in red, where the value is held until the next value appears:

I want everything else to remain the same if possible.

[kevinlinxc]

Okay I think https://altair-viz.github.io/gallery/step_chart.html is what I want, my bad