python - Plotly: How to avoid text in scatter plot being cut when exporting to PDF?

Question

I have the following code to make an scatter plot using plotly as px

fig1 = px.scatter(utilities, x="Rating", y="Spread", facet_col="Duration",text="Security_Name")
fig1.update_layout(width=1200,title_text='Spread por rating y duración en el sector Utilities')
fig1.update_traces(textfont_size=8,textposition='top right')
fig1.show()

The problem is that when I export fig1 to PDF using this code:

fig1.write_image("Utilities.pdf")

Then most of the texts in "text" are cut because names are too long (they cannot be shortened). I have already tried using "top center" and "top left" but in either case the text is cut. I would really appreciate your help.

Screenshot

Update: Status after upgrading plotly: I have shortened the strings a little bit.

question from:https://stackoverflow.com/questions/66055382/plotly-how-to-avoid-text-in-scatter-plot-being-cut-when-exporting-to-pdf

Answer 1

You haven't provided a dataset, but I've been able to reproduce your problem with a subset of a similar dataset px.data.stocks and an approach that resembles your use of px.scatter. And I'm assuming that your data has a long format, but that's not too important. What's important is this:

1. You're absolutely right, the text will be cut using px.Scatter([...] text = <text>)
2. There does surprisingly not seem to be a direct way to fix this.
3. I've produced a work-around that's more than a bit hacky.
4. The work-around turns this:

...into this:

The following snippet will reproduce the plot above. If this is something you can use, I'd be happy to elaborate on the details. But the short story is this: Faceted plots makes room for annotations made from add_annotations(), but not from annotations using px.Scatter([...] text = <text>).

Complete code:

# imports
import pandas as pd
import plotly.express as px
import plotly.io as pio

# data
df = px.data.stocks()
df = df[df.columns[:4]]
df2 = pd.melt(df.iloc[0:3],
              id_vars=['date'],
              value_vars=df.columns[1:],
              var_name='stock',
              value_name='value')

# your approach
fig = px.scatter(df2, x="date", y="value", facet_col="stock", text = 'stock')
fig.update_layout(width=1200,title_text='Spread por rating y duración en el sector Utilities')
fig.update_traces(textfont_size=8, textposition='top right')

# annotations
for c, d in enumerate(fig.data):
    for i, y in enumerate(d.y):      
        fig.add_annotation(x=d.x[i], y=d.y[i], text = str(d.x[i]),
                           font=dict(color="rgba(0,0,0,0)",size=10),
                           showarrow = False,
                           xanchor = 'left',
                           row = 1, col = c + 1
                          )
        
fig.write_image('c:plotlyplotssubnotation.pdf')
fig.show()