config.py

#!/usr/bin/env python

"""
Configuration, constants and helper functions for the project.
"""
from __future__ import annotations
import typing as tp

import numpy as np
import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
import seaborn as sns

from imc.types import Path, DataFrame  # type: ignore


figkws = dict(dpi=300, bbox_inches="tight")

metadata_dir = Path("metadata")
data_dir = Path("data")
results_dir = Path("results")

X_FILE = data_dir / "assay_data.csv"
Y_FILE = data_dir / "metadata.csv"


attributes = [
    "age",
    "sex",
    "obesity",
    "bmi",
    "race",
    "hospitalized",
    "intubated",
    "patient_group",
    "WHO_score_sample",
    "WHO_score_patient",
    "alive",
    "days_since_symptoms",
    "days_since_hospitalization",
]

palettes = dict(
    sex=sns.color_palette("Pastel1")[3:5],
    obesity=sns.color_palette("tab10")[3:6],
    race=sns.color_palette("tab10"),
    hospitalized=sns.color_palette("Set2")[:2],
    intubated=sns.color_palette("Set2")[:2],
    patient_group=np.asarray(sns.color_palette("Set1"))[[2, 1, 7, 3, 0]].tolist(),
    WHO_score_sample=sns.color_palette("inferno", 9),
    WHO_score_patient=sns.color_palette("inferno", 9),
    alive=sns.color_palette("Dark2")[:2],
)
cmaps = dict(
    age="winter_r",
    bmi="copper",
    days_since_symptoms="cividis",
    days_since_hospitalization="cividis",
)

_q = np.random.choice(range(40), 40, replace=False)
tab40 = matplotlib.colors.ListedColormap(
    colors=np.concatenate(
        [
            plt.get_cmap("tab20c")(range(20)),
            plt.get_cmap("tab20b")(range(20)),
        ]
    )[_q],
    name="tab40",
)

patches = (
    matplotlib.collections.PatchCollection,
    matplotlib.collections.PathCollection,
)