Extract Command Examples¶

Practical examples for using nbctl extract to save notebook outputs.

Basic Usage¶

Extract All Outputs (Interactive)¶

nbctl extract analysis.ipynb

Prompts:

What would you like to extract?
1. Both data and images
2. Only data outputs
3. Only image outputs
4. All outputs without prompting

Enter choice (1-4):

Extract Everything (Non-Interactive)¶

nbctl extract analysis.ipynb --all

Result:

outputs/
├── data/
│   └── cell_0_output_0_data_0.json
└── images/
    ├── cell_2_output_0_img_0.png
    └── cell_5_output_0_img_1.svg

Extract Only Images¶

nbctl extract analysis.ipynb --images

Perfect for saving plots and visualizations.

Extract Only Data¶

nbctl extract analysis.ipynb --data

Perfect for saving DataFrames and JSON outputs.

Custom Output Directory¶

Save to Custom Location¶

nbctl extract notebook.ipynb --all -o ./figures/

Organized Extraction¶

# Extract images to figures directory
nbctl extract analysis.ipynb --images -o ./report/figures/

# Extract data to data directory
nbctl extract analysis.ipynb --data -o ./report/data/

Workflow Examples¶

Create Report Package¶

Extract outputs for a report:

#!/bin/bash
# create-report-package.sh

notebook="analysis.ipynb"
report_dir="report-$(date +%Y%m%d)"

mkdir -p "$report_dir"

# Extract all outputs
nbctl extract "$notebook" --all -o "$report_dir/outputs/"

# Export to HTML
nbctl export "$notebook" -f html --output-dir "$report_dir/"

# Create archive
tar -czf "$report_dir.tar.gz" "$report_dir"

echo "Report package: $report_dir.tar.gz"

Extract for Publication¶

Save figures for academic paper:

# Create publication-ready figures directory
mkdir -p paper/figures

# Extract all plots
nbctl extract ml_results.ipynb --images -o paper/figures/

# Rename for paper
cd paper/figures
mv cell_3_output_0_img_0.png figure1_accuracy.png
mv cell_7_output_0_img_1.png figure2_confusion_matrix.png

Batch Extraction¶

Extract from multiple notebooks:

#!/bin/bash
# Extract outputs from all analysis notebooks

for nb in analysis_*.ipynb; do
    name=$(basename "$nb" .ipynb)
    nbctl extract "$nb" --all -o "outputs/$name/"
done

echo "Extracted outputs from all notebooks"

Data Extraction Examples¶

Save DataFrames¶

Extract pandas DataFrame HTML tables:

nbctl extract data_analysis.ipynb --data

Result: DataFrames saved as HTML tables.

Extract JSON Results¶

Save model evaluation metrics:

nbctl extract ml_training.ipynb --data -o results/

# results/data/ contains JSON files with metrics

Extract CSV Data¶

nbctl extract processing.ipynb --data

# CSVoutputs saved as .csv files

Image Extraction Examples¶

Save All Plots¶

Extract matplotlib/seaborn plots:

nbctl extract visualization.ipynb --images -o plots/

Save Vector Graphics¶

Extract SVG for scalable figures:

nbctl extract notebook.ipynb --images -o figures/

# Look for .svg files in figures/images/

Extract for Presentation¶

# Extract plots for slides
nbctl extract analysis.ipynb --images -o presentation/images/

# Use in PowerPoint, Google Slides, etc.

Advanced Examples¶

Conditional Extraction¶

Extract only if outputs exist:

if nbctl info notebook.ipynb | grep -q "outputs"; then
    nbctl extract notebook.ipynb --all
    echo "Outputs extracted"
else
    echo "ℹ No outputs to extract"
fi

Extract with Timestamp¶

timestamp=$(date +%Y%m%d-%H%M%S)
nbctl extract notebook.ipynb --all -o "outputs-$timestamp/"

Extract and Archive¶

# Extract outputs
nbctl extract notebook.ipynb --all -o temp-outputs/

# Create archive
tar -czf outputs-$(date +%Y%m%d).tar.gz temp-outputs/

# Cleanup
rm -rf temp-outputs/

echo "Outputs archived"

Automated Workflows¶

Daily Plot Extraction¶

#!/bin/bash
# daily-plots.sh - Extract plots daily

date=$(date +%Y-%m-%d)
plot_dir="plots/$date"

mkdir -p "$plot_dir"

for nb in dashboards/*.ipynb; do
    # Run notebook to generate latest plots
    nbctl run "$nb"

    # Extract plots
    name=$(basename "$nb" .ipynb)
    nbctl extract "$nb" --images -o "$plot_dir/$name/"
done

echo "Daily plots extracted to $plot_dir"

CI/CD Integration¶

.github/workflows/extract-outputs.yml:

name: Extract Outputs
on:
  push:
    paths:
      - '**.ipynb'

jobs:
  extract:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2

      - name: Setup Python
        uses: actions/setup-python@v2

      - name: Install nbctl
        run: pip install nbctl

      - name: Extract outputs
        run: |
          mkdir -p artifacts
          for nb in *.ipynb; do
            nbctl extract "$nb" --all -o "artifacts/${nb%.ipynb}/"
          done

      - name: Upload artifacts
        uses: actions/upload-artifact@v2
        with:
          name: notebook-outputs
          path: artifacts/

Organization Patterns¶

By Notebook¶

# Structure: outputs/{notebook_name}/data/ and /images/

for nb in *.ipynb; do
    name=$(basename "$nb" .ipynb)
    nbctl extract "$nb" --all -o "outputs/$name/"
done