scripts/script.py

import pandas as pd
import math
import plotly.subplots as sp
import plotly.graph_objects as go

import os


def plot_pie_chart(df):
    ARG_prediction_counts = dict(df["ARG_prediction"].value_counts())
    ARG_prediction_df = pd.DataFrame.from_dict(
        ARG_prediction_counts, orient="index", columns=["count"]
    )
    resistance_category_counts = dict(df["resistance_category"].value_counts())
    resistance_category_df = pd.DataFrame.from_dict(
        resistance_category_counts, orient="index", columns=["count"]
    )

    number_of_catgeory = len(df["resistance_category"].value_counts())

    colors = [
        "#f9b4ab",
        "#fdebd3",
        "#264e70",
        "#679186",
        "#bbd4ce",
    ]
    full_colors = []
    for i in range(math.ceil(number_of_catgeory / 5)):
        full_colors += colors

    # colors = ['gold', 'mediumturquoise', 'darkorange', 'lightgreen']

    fig = sp.make_subplots(
        rows=1,
        cols=2,
        subplot_titles=("ARG/non-ARG", "Resistance category"),
        specs=[[{"type": "domain"}, {"type": "domain"}]],
    )

    fig.add_trace(
        go.Pie(
            labels=ARG_prediction_df.index,
            values=ARG_prediction_df["count"],
            legendgroup="1",
            title="ARG/non-ARG",
        ),
        row=1,
        col=1,
    )
    fig.add_trace(
        go.Pie(
            labels=resistance_category_df.index,
            values=resistance_category_df["count"],
            legendgroup="2",
            title="Resistance category",
        ),
        row=1,
        col=2,
    )

    fig.update_layout(showlegend=False, margin=dict(l=200, r=200, t=100, b=100))
    fig.update_traces(
        textposition="inside",
        hoverinfo="label+percent",
        textinfo="label",
        marker=dict(colors=full_colors, line=dict(color="#38496e", width=1)),
    )

    return fig


def view_stat(output_name):
    df = pd.read_csv(f"results/{output_name}", delimiter="\t")
    # change df header
    new_headers = ["Test ID", "ARG Prediction", "Resistance Category", "Probability"]
    fig = plot_pie_chart(df)

    # delete the output file
    if os.path.exists(f"results/{output_name}"):
        os.remove(f"results/{output_name}")

    return [df.rename(columns=dict(zip(df.columns, new_headers))), fig]


def run_argnet(input, output_name, sequence_type, sequence_length_type):
    with open("input.txt", "w") as f:
        f.write(input)

    # delete the output file
    if os.path.exists(f"results/{output_name}"):
        os.remove(f"results/{output_name}")

    if sequence_type == "aa" and sequence_length_type == "s":
        from . import argnet_ssaa_chunk as ssaa

        ssaa.argnet_ssaa("input.txt", output_name)
    elif sequence_type == "nt" and sequence_length_type == "s":
        from . import argnet_ssnt_new_chunk as ssnt

        ssnt.argnet_ssnt("input.txt", output_name)
    elif sequence_type == "aa" and sequence_length_type == "l":
        from . import argnet_lsaa_speed_sgpu as lsaa

        lsaa.argnet_lsaa("input.txt", output_name)
    elif sequence_type == "nt" and sequence_length_type == "l":
        from . import argnet_lsnt as lsnt

        lsnt.argnet_lsnt("input.txt", output_name)

    if os.path.exists("input.txt"):
        os.remove("input.txt")