Multiline text in mark_text?

[arogozhnikov]

There is a related issue for axes labels and titles that suggests passing lists of string for multilines, but it does not work for just pandas columns with multiline strings that are passed to mark_text.

If more context is needed, I want to display multiline text over the heatmap, like in this example.

[joelostblom]

mark_text works the same as axes labels and titles. Pandas columns that contain lists will be split onto multiple lines (works for both string and numbers):

import pandas as pd
import altair as alt


alt.Chart(pd.DataFrame({'a': [['one', 'two']]})).mark_text().encode(
    x=alt.value(0),
    y=alt.value(0),
    text='a'
)

[arogozhnikov]

thank you for your example. The same approach did not work for me, and after digging deeper I found that following example works differently:

• in jupyter+altair it plots multilines
• in jupyter+streamlit (same python env) it plots stringified version of array in a single line

I believe the reason is that streamlit uses old vega-lite maybe? I can confirm after using .to_html() in streamlit the porblem is gone

import pandas as pd
import altair as alt

def test_plot():
    df = pd.DataFrame([
        {'text': f'this is {row}', 'i': row // 10, 'j': row % 10, 'value': row}
        for row in range(100)
    ])
    df['text_split'] = df['text'].str.split(' ')

    base = alt.Chart(df).encode(
        x='i:N',
        y='j:N',
        tooltip=list(df.columns)
    )

    return alt.layer(
        base.mark_rect().encode(
            color='value:Q',
        ),
        base.mark_text(baseline='middle').encode(
            text='text_split'
        ),
    ).properties(
        width=500,
        height=500,
    )