app.py

from tipars.plot_tree import plot_fig
import gradio as gr
import os
import pandas as pd
import subprocess


def get_sequence_example(example):
    if example.startswith("H5"):
        return open("tipars/ref-tree/Influenza-A-H5/H5.fasta").read()
    elif example.startswith("SARS-CoV-2"):
        return open("tipars/ref-tree/sars-cov-2/sars2.fasta").read()


def mirror(a, b, c, d):
    return a, b, c, d


def insertion(
    task,
    tipars_file1,
    tipars_file2,
    tipars_file3,
    tipars_file4,
    input_textbox1,
    input_textbox2,
):
    # remove tipars.tree if exists
    if os.path.exists("tipars/tipars.tree"):
        os.remove("tipars/tipars.tree")
    if task == "tipars":
        print(
            tipars_file1.name, tipars_file2.name, tipars_file3.name, tipars_file4.name
        )
        query = tipars_file1.name
        tree = tipars_file2.name
        msa = tipars_file3.name
        anc = tipars_file4.name
    if task == "h5":

        with open("tipars/test.fasta", "w") as f:
            f.write(input_textbox2)
        query = "tipars/test.fasta"

        tree = "tipars/ref-tree/Influenza-A-H5/tree.nwk"
        msa = "tipars/ref-tree/Influenza-A-H5/taxa.fasta"
        anc = "tipars/ref-tree/Influenza-A-H5/anc.fas"
    elif task == "sars":

        with open("tipars/test.fasta", "w") as f:
            f.write(input_textbox1)
        query = "tipars/test.fasta"
        tree = "tipars/ref-tree/sars-cov-2/sars-cov-2_ready.tree"
        msa = "tipars/ref-tree/sars-cov-2/sars-cov-2_taxa.fasta"
        anc = "tipars/ref-tree/sars-cov-2/sars-cov-2_anc.fas"
    # open query fasta file
    # and add '(Query)' after '>'
    # print(query, tree, msa, anc)
    with open(query, "r") as f:
        lines = f.readlines()
        with open("tipars/query.fasta", "w") as f:
            for line in lines:
                if line.startswith(">"):
                    f.write(line.strip() + "_query\n")
                else:
                    f.write(line)
    output = "tipars/tipars.tree"
    query = "tipars/query.fasta"
    command = ["mafft-linux64/mafft.bat", "--add", query, "--keeplength", msa]
    with open("mafft_output.fas", "w") as output_file:
        subprocess.run(command, stdout=output_file, text=True)
    command = ["./seqkit", "seq", "-n", query]
    with open("query-name.txt", "w") as output_file:
        subprocess.run(command, stdout=output_file, text=True)
    command = ["./seqkit", "grep", "-f", "query-name.txt", "mafft_output.fas"]
    with open(query, "w") as output_file:
        subprocess.run(command, stdout=output_file, text=True)
    subprocess.run(
        [
            "tipars/tipars",
            "-t",
            tree,
            "-s",
            msa,
            "-a",
            anc,
            "-q",
            "tipars/query.fasta",
            "-o",
            output,
        ]
    )
    fig = plot_fig(output)
    return fig


css = """
h1 {
    text-align: center;
    display:block;
    
}
"""

with gr.Blocks(css=css) as demo:
    gr.Image(
        "tipars/tipars.svg",
        width=50,
        height=50,
        min_width=0,
        container=False,
        show_download_button=False,
    )

    task_selected = gr.State("sars")

    gr.Markdown(
        """# Phylogenetic Placement by TIPars
        <center>
        This is HuggingFace hosting of <a href=https://github.com/id-bioinfo/TIPars>TIPars</a>.
        </center>
        """
    )
    with gr.Row():

        with gr.Column():
            gr.Markdown(
                """## Input
                        Input your query sequences.
            """
            )
            with gr.Tab("SARS-Cov-2 Full Genome Tree") as sars_tab:

                gr.Markdown(
                    """
                ### Query Sequences
                Sequences to be insert to Reference Tree (.fasta)
                """
                )
                input_textbox1 = gr.Textbox(label="Query Sequences")
                gr.Examples(
                    examples=[
                        ["SARS-CoV-2"],
                    ],
                    inputs=input_textbox1,
                    fn=get_sequence_example,
                    cache_examples=True,
                    outputs=input_textbox1,
                )

            with gr.Tab("Influenza-A-H5 HA Tree") as h5_tab:

                gr.Markdown(
                    """
                ### Query Sequences
                Sequences to be insert to Reference Tree (.fasta)
                """
                )
                input_textbox2 = gr.Textbox(label="Query Sequences")
                gr.Examples(
                    examples=[["H5"]],
                    inputs=input_textbox2,
                    fn=get_sequence_example,
                    cache_examples=True,
                    outputs=input_textbox2,
                )

            with gr.Tab("TIPars") as tipars_tab:

                gr.Markdown(
                    """
                ### Query Sequences
                Sequences to be insert to Reference Tree (.fasta)
                """
                )
                tipars_file1 = gr.File(
                    label="Query Sequences",
                    file_types=["fasta", "fas", "fna", "ffn", "faa", "frn", "fa"],
                )
                gr.Markdown(
                    """
                ### Rooted Reference Tree
                The tree which new sample shall be inserted to (.nwk)
                """
                )
                tipars_file2 = gr.File(
                    label="Rooted Reference Tree", file_types=["nwk", "new", "tree"]
                )
                gr.Markdown(
                    """
                ### Tree MSA (Multiple Sequence Alignment)
                The multiple sequence alignment of taxa that built the Reference Tree (.fasta)
                """
                )
                tipars_file3 = gr.File(
                    label="Tree MSA (Multiple Sequence Alignment)",
                    file_types=["fasta", "fas", "fna", "ffn", "faa", "frn", "fa"],
                )
                gr.Markdown(
                    """
                ### Ancestral Sequence Alignment
                Ancestral sequence (aligned) obtained from PastML or TreeTime (.fasta)
                """
                )
                tipars_file4 = gr.File(
                    label="Ancestral Sequence Alignment",
                    file_types=["fasta", "fas", "fna", "ffn", "faa", "frn", "fa"],
                )
                gr.Examples(
                    examples=[
                        [
                            (
                                os.path.join(
                                    os.path.dirname(__file__),
                                    "tipars/Benchmark_datasets/NDV/NDV_query.fas",
                                )
                            ),
                            (
                                os.path.join(
                                    os.path.dirname(__file__),
                                    "tipars/Benchmark_datasets/NDV/NDV_tree.nwk",
                                )
                            ),
                            (
                                os.path.join(
                                    os.path.dirname(__file__),
                                    "tipars/Benchmark_datasets/NDV/NDV_taxa.fas",
                                )
                            ),
                            (
                                os.path.join(
                                    os.path.dirname(__file__),
                                    "tipars/Benchmark_datasets/NDV/NDV_anc.fas",
                                )
                            ),
                        ]
                    ],
                    inputs=[tipars_file1, tipars_file2, tipars_file3, tipars_file4],
                    fn=mirror,
                    # cache_examples=True,
                    outputs=[tipars_file1, tipars_file2, tipars_file3, tipars_file4],
                )

            btn = gr.Button("Run")
        with gr.Column():
            gr.Markdown(
                """## Output
                        TIPars insertion result:
            """
            )

            charts = gr.Plot(label="Tree")
    h5_tab.select(lambda: "h5", inputs=None, outputs=task_selected)
    sars_tab.select(lambda: "sars", inputs=None, outputs=task_selected)
    tipars_tab.select(lambda: "tipars", inputs=None, outputs=task_selected)

    btn.click(
        fn=insertion,
        inputs=[
            task_selected,
            tipars_file1,
            tipars_file2,
            tipars_file3,
            tipars_file4,
            input_textbox1,
            input_textbox2,
        ],
        outputs=[charts],
    )
if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)