[add_feature] export figures

[shlff]

Related code:

export_figures = False

and

if export_figures == True:
	plt.savefig("figures/xx.pdf")

where xx is the figure’s name.

export_figures name lists for each chapter:

ch_intro

• crude_oil_2019.pdf
• commercial_aircraft_2019_1.pdf
• simplex_1.pdf
• heavy_tailed_draws.pdf
• ccdf_comparison_1.pdf
• empirical_powerlaw_plots_firms_forbes.pdf
• zeta_1.pdf
• networkx_basics_1.pdf
• financial_network_analysis_visualization.pdf
• financial_network_analysis_centrality.pdf
• commercial_aircraft_2019_2.pdf
• rand_graph_experiments_1.pdf
• rand_graph_experiments_2.pdf

ch_production

• input_output_analysis_15.pdf
• input_output_analysis_71.pdf
• input_output_analysis_15_leo.pdf
• input_output_analysis_15_shocks.pdf
• input_output_analysis_15_ec.pdf
• input_output_analysis_aus_114.pdf
• input_output_analysis_15_omult.pdf
• input_output_analysis_15_fwd.pdf
• input_output_analysis_15_up.pdf
• gdp_growth.pdf
• input_output_analysis_15_katz.pdf

ch_opt

• shortest_path_iter_1.pdf
• betweenness_centrality_1.pdf
• linear_programming_1.pdf
• ot_figs_1.pdf
• ot_large_scale_1.pdf

ch_mcs

• markov_matrix_visualization.pdf
• benhabib_mobility_mixing.pdf
• quah_gdppc_prediction.pdf
• simplex_2.pdf
• simplex_3.pdf
• benhabib_mobility_dists.pdf
• benhabib_ergodicity_1.pdf

ch_fpms

• fin_network_sims_1.pdf

appendix

• func_types_1.pdf
• func_types_2.pdf
• three_fixed_points.pdf
• complex_number.pdf
• euclidean_convergence_1.pdf
• span1.pdf
• func_types_3.pdf
• polyhedron1.pdf
• saddle_1.pdf

[shlff]

Hi @jstac the figure saddle_1.pdf in the current online website has been separated into 4 plots.

PR #39 combines these 4 figures together again, the same as the figure used in the textbook.

[shlff]

Hi @mmcky , @jstac and I decide to generate the figures used in the network textbook directly from this repo.

To make the figures aligned with our original ones, we need lines 6-13 in a file called initialize.py to control the figure styles.

Do you think we should

• incorporate the corresponding code into somewhere of the current repo or
• push it to the new repo (https://github.com/QuantEcon/quantecon-book-networks)?

Or other options?

Looking forward to your advice.

[jstac]

@mmcky , we would appreciate your advice here. The lines to be added require a latex installation. Will this be a problem for the GH build?

[mmcky]

@shlff if you want that configuration to matplotlib to exist for all users then it is probably best to go into https://github.com/QuantEcon/quantecon-book-networks

You could:

1. add it in __init__.py, or
2. add it as a function that you call to setup matplotlib when the package is imported such as quantecon_book_networks.configure("matplotlib")

Option 2 is probably best as your not changing matplotlib configuration by default when importing that package.

[mmcky]

@jstac the lines are matplotlib configurations right?

[mmcky]

@jstac to answer your question adding latex is no problem. We build pdf now for the live sites via GitHub actions etc.

[jstac]

OK, thanks @mmcky . @shlff , let's remember to check that the figures look OK after this PR is merged.

[shlff]

Thanks for your great suggestion and comment @mmcky and @jstac .

Sure @jstac I will set up the environment using environment.yml and compile the files on my local machine before you review the PR.

[Fdu-lijiachen]

Question from 'Chapter 2 - Production -Backward linkages for 15 US sectors in 2019 '
Is hub-based eigenvector with centrality backward linkage coefficient matrix right?
backward linkage may view the sight from supply side, like i is supplier of j, so I think may be use authority-based is better?

[jstac]

Thanks for your comment @Fdu-lijiachen . Are you referring to the color scheme of that figure?

[Fdu-lijiachen]

Thanks for your comment @Fdu-lijiachen . Are you referring to the color scheme of that figure?

Sorry，I mean this is a question I raised about Chapter 2, but I think I put it in the wrong place. I wonder if there is a place to discuss it with you?

[jstac]

Hi @Fdu-lijiachen

If you click on the green button near the top that says "new issue", you can open a new issue to describe the problem. That is our standard procedure. Otherwise, if you prefer, you can reach me by email via john.stachurski@anu.edu.au.

Your comments will be appreciated. It's very helpful to us if you can carefully describe the issue.