main.py

from graph.PPINetwork import PPINetwork
from graph.StringLoader import StringLoader
from visual.ReportGenerator import ReportGenerator
from visual.plots_macros import *


def pipeline():

    network_type = "functional"  # "physical"
    protein_query = ["TP53", "BRCA1"]
    # protein_query = ["INS", "ALB", "APP", "SNCA", "CFTR"]
    nodes_to_add = len(protein_query)*10

    scatter_plot_metrics = [
        ("degree", "betweenness_centrality"),
        ("degree", "pagerank"),
        ("clustering_coefficient", "degree")
    ]
    top_edges_scores = 20

    loader = StringLoader(
        protein_query=protein_query,
        species=9606,
        add_nodes=nodes_to_add,
        required_score=400,
        network_type=network_type
    )

    print("Fetching data...")

    data = (
        loader
        .retrieve_data()
        .standardize_data_format()
        .get_data()
    )

    print("Construction PPI network...")
    ppin = PPINetwork(data=data)
    print("Computing network metrics...")
    ppin.compute_all_global_metrics(weighted=False)
    ppin.compute_all_edge_metrics(weighted=False)
    ppin.compute_all_node_metrics(weighted=False)

    exported_files = ppin.export_for_cytoscape(
        output_dir="cytoscape_export",
        basename=f"ppi_network_{'-'.join(protein_query)}",
    )

    print(f"Network exported as:")
    for file, path in exported_files.items():
        print(f"\t{file}: {path}")

    print("Generating plots...")
    node_labels = ppin.get_nodes_preferred_names()
    scatter_plot_html = "\n".join([
        f"""
        <div class="plotInGrid">
            <img src="{s}" class="imagePlotCard">
        </div>
        """
        for s in [
            create_scatter_plot_encoded_image(
                x_values=ppin.get_nodes_metric_value(metric_1)[0],
                y_values=ppin.get_nodes_metric_value(metric_2)[0],
                x_label=metric_1, y_label=metric_2,
                labels=node_labels
            )
            for metric_1, metric_2 in scatter_plot_metrics
        ]
    ])

    values_e_net, edges_net = ppin.get_edges_metric_value("confidence")
    edges_plot = create_barplot_encoded_image(
        edges_net[:top_edges_scores],
        values_e_net[:top_edges_scores],
        y_label="Edge", x_labebl="Confidence"
    )

    global_plot = create_graph_degree_distribution_encoded_image(
        ppin.get_nodes_metric_value("degree")[0],
        bins=20
    )

    weighted_degree_hubs_df = ppin.get_hub_proteins("weighted_degree", top_n=7)
    consensus_hubs_df = ppin.get_consensus_hub_proteins(top_n=7)

    bar_plot_hubs = create_barplot_encoded_image(
        weighted_degree_hubs_df['preferredName'].tolist(),
        weighted_degree_hubs_df['weighted_degree'].tolist(),
        title="Hubs by Weighted Degree",
        x_labebl='weighted_degree'
    )
    bar_plot_consensus_hubs = create_barplot_encoded_image(
        consensus_hubs_df['preferredName'].tolist(),
        consensus_hubs_df['weighted_degree'].tolist(),
        title="Consensus Hubs",
        x_labebl='weighted_degree'
    )

    ranked_genes_df, gsea_results_df = ppin.run_gsea(
        node_attribute="degree_centrality",
        gene_sets="GO_Biological_Process_2023",
        outdir="gsea_results"
    )

    print(ranked_genes_df.head())
    print(gsea_results_df.head(20))

    plot_gsea = ppin.plot_gsea_dotplot(
        gsea_results_df=gsea_results_df,
        column="FDR q-val",
        title="GO Biological Process enrichment",
        cutoff=0.25,
        top_term=10,
    )

    hubs_plots_html = "\n".join([
        f"""
            <div class="plotInGrid">
                <img src="{s}" class="imagePlotCard">
            </div>
            """ for s in [bar_plot_hubs, bar_plot_consensus_hubs, plot_gsea]
    ])

    print("generating html report...")

    report_name = "protein_network_report.html"
    (ReportGenerator()
        .add_cytoscape_html_report(
            ppin,
            title="STRING Protein Network",
            node_color_attr="katz_centrality",  # "betweenness_centrality",
            edge_width_attr="escore",  # "edge_betweenness_centrality",
            cmap_name="plasma",
            min_node_size=45,
            max_node_size=120,
            min_edge_width=1,
            max_edge_width=10
        )
        .add_section(html_to_add=f"""
            <h2>StringDB fetched network. Click on the logo for more!</h2>
            {loader.get_network_html_div()}
        """)
        .add_section_from_file(
            file="./visual/web/double_plots.html",
            replacements={
                "__global_plot__": global_plot,
                "__edges_plot__": edges_plot
            }
        )
        .add_section_from_file(
            file="visual/web/grid_plots.html",
            replacements={
                "__plots_html__": scatter_plot_html,
                "__title_grid__": "Scatter plots",
                "__description_grid__": (
                    "The scatter plots compare pairs of node-level network metrics. Each point"
                    " represents a protein, and the label identifies the corresponding preferred"
                    " protein name. These plots help reveal whether different centrality or"
                    " structural measures capture similar or complementary aspects of the network."
                )
            }
        )
        .add_section_from_file(
            file="visual/web/grid_plots.html",
            replacements={
                "__plots_html__": hubs_plots_html,
                "__title_grid__": "Hubs plots",
                "__description_grid__": (
                    "These plots summarize the main hub proteins identified in the protein-protein "
                    "interaction network. The first bar plot ranks proteins by weighted degree, "
                    "highlighting nodes with many high-confidence interactions. The second plot "
                    "shows the consensus hub ranking, which combines multiple centrality measures to "
                    "identify proteins that are consistently important across different network perspectives. "
                    "The enrichment plot provides a biological interpretation of the ranked genes by showing "
                    "Gene Ontology biological processes associated with the most central proteins."
                )
            }
        )
        .generate_report_file(output_file=report_name)
     )

    print(f"Report generated: {report_name}")


if __name__ == "__main__":
    pipeline()