Export graphs in a single PDF file

[lichengzhang1]

Hello,

I see that when converting all the images to PDF, they are all separate. In fact, in academic writing, sometimes we need to put them together (though usually not too many, typically less than 200 images). I find it relatively easy to do this using Mathematica.

g = Import["http://users.cecs.anu.edu.au/~bdm/data/highlyirregular12.g6"];

Grid[Partition[
  Table[Labeled[
    Graph[g[[i]], PlotTheme -> "Monochrome", 
     VertexSize -> .17 {1, 1}, ImageSize -> {50, 30}, 
     VertexStyle -> GrayLevel[.6]], 
    Style[i, FontFamily -> "Baskerville"], Bottom], {i, Length[g]}], 
  UpTo[10]]]

See also https://mathematica.stackexchange.com/questions/271553/how-to-fit-many-graphs-neatly-into-a-paper.

I'm wondering if this can also be done with this software or with Python. Of course, a single-page PDF may not accommodate all of them, so it needs to be intelligently adjusted, including the number of images in each row and the number of columns on each page.

[atilaajones]

Hello @lichengzhang1. We have already evaluated your proposal and will implement it for the next version of the software.

Thank you for your cooperation. It's always a pleasure to receive feedback and suggestions from users.

[lichengzhang1]

Now, I write a python code which can handle this.

import matplotlib.pyplot as plt
import networkx as nx
from matplotlib.backends.backend_pdf import PdfPages
import math

def read_graph6_file(filepath):
    with open(filepath, 'r') as file:
        graph_data = file.read().splitlines()
    graphs = [nx.from_graph6_bytes(line.encode('ascii')) for line in graph_data]
    return graphs

def draw_graphs_to_pdf(graphs, columns, pdf_path):
    rows_per_page = 6  # Number of rows per page
    graphs_per_page = rows_per_page * columns
    num_pages = math.ceil(len(graphs) / graphs_per_page)

    plt.rcParams['text.usetex'] = True  # Enable LaTeX rendering

    with PdfPages(pdf_path) as pdf:
        for page in range(num_pages):
            start = page * graphs_per_page
            end = start + graphs_per_page
            current_graphs = graphs[start:end]

            num_rows = math.ceil(len(current_graphs) / columns)
            fig, axes = plt.subplots(nrows=num_rows, ncols=columns, figsize=(15, num_rows * 3))

            # Ensure axes is always a 2D array
            if num_rows == 1:
                axes = [axes]
            if columns == 1:
                axes = [[ax] for ax in axes]

            for ax, graph, i in zip([ax for row in axes for ax in row], current_graphs, range(start + 1, end + 1)):
                #pos = nx.spring_layout(graph)
                pos = nx.circular_layout(graph)
                nx.draw_networkx(
                    graph,
                    pos,
                    ax=ax,
                    with_labels=True,
                    labels={node: rf'${node}$' for node in graph.nodes()},
                    node_size=250,
                    node_color='#D3D3D3',
                    edgecolors='black',
                    linewidths=2,
                    font_size=10,
                    font_color='black',
                    font_weight='bold',
                    edge_color='black',
                    width=1.5
                )
                ax.set_title(f'Graph {i}', fontsize=10)
                ax.axis('on')  # Display frame
                ax.set_xticks([])
                ax.set_yticks([])

            # Remove empty axes
            for ax in [ax for row in axes for ax in row][len(current_graphs):]:
                fig.delaxes(ax)

            plt.tight_layout(pad=3.0)
            if len(current_graphs) > 0:  # Avoid saving empty pages
                pdf.savefig(fig)
            plt.close()

# Example usage:
columns = 5      # Number of columns per page
pdf_path = './graph6c.pdf'  # Save to current working directory

graphs = read_graph6_file("graph6c.g6") # download from https://users.cecs.anu.edu.au/~bdm/data/graphs.html
draw_graphs_to_pdf(graphs, columns, pdf_path)

print(f"PDF saved to: {pdf_path}")

[fsoupimenta]

Hello @lichengzhang1, hope you're well!

Thanks for your implementation. We ended up not giving this issue that much attention because we are working on a complete refactoring of the software architecture #452, in order to ensure quality and facilitate the implementation of new features.

However, we've recently been working on some requests again, and we're about to update a new release including this new feature export functionality. You can follow its development on #451 , where it can be tested on your branch.

Feel free to request new features or report any problem, and also open pull requests.

Thank you for your support, it's great to have you here 😊!