Dynamic Book Chapter Creation

[DeanVogt]

Description

I have a situation where I’d like to dynamically create chapters for a Quarto HTML book. In the sample code below, I’m producing results for five different segments using a loop. I’d like each iteration to produce a new book chapter. I can produce the desired book results by creating one notebook for each segment, but that’s not ideal since any changes need to be made in each notebook. Again using the sample code below, I'd like to produce a new chapter for each of the five segments (A-E).

I’m using Jupyter Lab, Python, and Quarto version 1.2.475 in a Windows environment.

# Set Up
import numpy as np
import pandas as pd
import plotly.express as px

# Make Data
np.random.seed(42)

n_observations = 1000

df = pd.DataFrame({
    'segment': np.random.choice(['A', 'B', 'C', 'D', 'E'], n_observations),
    'var1': np.random.rand(n_observations),
    'var2': np.random.rand(n_observations)
})

# Results Across All Segments.  Ideally, this content would be in a chapter.
segment_results = df.groupby('segment').apply(lambda x: pd.Series({
    '# of observations': (x['segment'].count()),
    'var1 mean': (x['var1'].mean()),
    'var2 mean': (x['var2'].mean()),
    'correlation': (np.corrcoef(x['var1'], x['var2'])[0, 1])
}))

total_row = pd.Series({
    '# of observations': df['segment'].count(),
    'var1 mean': df['var1'].mean(),
    'var2 mean': df['var2'].mean(),
    'correlation': np.corrcoef(df['var1'], df['var2'])[0, 1]
}, name='Total')

total_row = pd.DataFrame(total_row).transpose()

results = pd.concat([segment_results, total_row]).rename_axis('Segment')

results['# of observations'] = results['# of observations'].map('{:,.0f}'.format)

display(results)

px.scatter(df.sort_values('segment'), x='var1', y='var2', color='segment', title='All Observations')

# Results by Segment.  Ideally, results for each segment would be in a separate chapter.
for segment in df['segment'].sort_values().unique():
    print(f"Segment: {segment}")
    print(results[results.index==segment].reset_index(drop=True).to_string(index=False))
    display(px.scatter(df[df['segment']==segment], x='var1', y='var2', title=f"Segment {segment}"))

[mcanouil]

Use pre-render script and generate the chapter files.
Note that the first run might get it wrong as all the files does not exist yet.

Also, I suggest you upgrade to 1.3, especially since 1.4 is on the horizon.

[mcanouil]

Does globing not work in chapters? I would assume it does as with render.

[jdlom]

There is an error if you don't explicitly set the chapters in the chapter section.

[mcanouil]

In some aspect, it makes sense to not have globing to define chapters but wonder if that could be a good enhancement?🤔
CC @dragonstyle What do you think about allowing globing in chapters? Is it possible? A not bad idea?

[dragonstyle]

I think it is a good idea and something I will take a look at!

#1917

(I'll mark this discussion closed in favor of the above issue)

[jdlom]

Would be a great feature because for now, I generate with a script a _quarto-dyn.yml file and use quarto -profile dyn