Subplot with facet_col_wrap has unexpected behavior when referring to rows

[valankar]

Not sure what the problem here is but it is very confusing. I'm trying to add a hline to a subplot using facet_col_wrap. When I specify the row to add_hline(), it doesn't actually add it to the proper row.

The code I have is as follows:

#!/usr/bin/env python3
"""Test plot."""
import pandas as pd
import plotly.express as px


def main():
    """Test plot."""
    test_df = pd.read_csv('test.csv', index_col=0, parse_dates=True,
                          infer_datetime_format=True)
    line = px.line(
        test_df, x=test_df.index, y=test_df.columns, facet_col='variable', facet_col_wrap=2,
    )
    line.add_hline(y=100000, row=3, col=1)
    with open('test.html', 'w', encoding='utf-8') as out_file:
        out_file.write(line.to_html())


if __name__ == '__main__':
    main()

Here is a screenshot of the web page generated:

As you can see, the line gets added to row 1 instead of row 3.

The input CSV file is attached:

test.csv

[nicolaskruchten]

Rows are numbered from the bottom in Plotly, similar to how the origin for a cartesian plot is in the bottom-left. I'm sorry this has caused confusion for you!

[valankar]

Ah, I see. Thanks!

[nicolaskruchten]

Clarification: in Plotly Express they're numbered from the bottom, because PX calls make_subplots internally with start_cell='bottom-left'

[valankar]

I see, thanks. I think this is mainly what confused me, because I was also using make_subplots directly and expected the same behavior with Plotly Express.