Cut off data with `layout.auto_size(true)`

[Redhawk18]

This data I'm working on has 35 or 36 subjects on the y axis in my data, but theres only 18 in the graph? Also I think the same thing is happening on the x axis as well, I just don't know how many columns are missing there.

    let trace = HeatMap::new(
        subject_labels[0].clone(),
        subject_names.clone(),
        subject_values,
    );
    // let layout = Layout::new().auto_size(true).height(HEIGHT);
    let layout = Layout::new().auto_size(true);
    let mut plot = Plot::new();
    plot.add_trace(trace);
    plot.set_layout(layout);

    println!("{}", subject_names.len());
    let filename = "No Wait";
    plot.write_html(&format!("{}/html/{}.html", full_path, filename));

[Redhawk18]

Is there a method I'm missing? In my experience the height and width changed how big the graph was it didn't show more data for example.

[andrei-ng]

Have you tried setting the layout number of ticks per axis? Something like

    let layout = Layout::new()
        .x_axis(Axis::new().n_ticks(100))
        .y_axis(Axis::new().n_ticks(36));
    plot.set_layout(layout);

But I doubt what you can visually extract from such a heat map. Maybe a better plot would be better for your task. 🤷‍♂️

[Redhawk18]

I came up with a good pruning strategy, I will also be using other plots. But Why is the default behavior to cut data points off?

[andrei-ng]

I came up with a good pruning strategy, I will also be using other plots. But Why is the default behavior to cut data points off?

My educated guess would be that plotly.js defaults to some auto strategy for selecting ticks and axis labels when none is provided.

[Redhawk18]

Couldn't we just enforce x and y axis since the size is almost always known in rust?

[andrei-ng]

Feel free to close the issue if you found yourself a fix.

[Redhawk18]

No I meant as a pr

[Redhawk18]

in my example we would never run into this issue if the library just used the length of the slice/vec as n ticks

[andrei-ng]

in my example we would never run into this issue if the library just used the length of the slice/vec as n ticks

This is not in the spirit of any of the Plotly libraries which are meant for dynamic visualizations. The libraries provide some sensible defaults, which work nicely for most use cases, which the users can further customize as per their particular needs.

Your suggestion won't play nice if the data has tens of thousands of points or millions.