speaker-confusion-model/code/plots/corpora_biases.py

#!/usr/bin/env python3

import pandas as pd
import pickle
import numpy as np
from scipy.special import logit, expit

import argparse
import matplotlib
import matplotlib.pyplot as plt

matplotlib.use("pgf")
matplotlib.rcParams.update(
    {
        "pgf.texsystem": "pdflatex",
        "font.family": "serif",
        "font.serif": "Times New Roman",
        "text.usetex": True,
        "pgf.rcfonts": False,
    }
)


def set_size(width, fraction=1, ratio=None):
    fig_width_pt = width * fraction
    inches_per_pt = 1 / 72.27
    if ratio is None:
        ratio = (5**0.5 - 1) / 2
    fig_width_in = fig_width_pt * inches_per_pt
    fig_height_in = fig_width_in * ratio
    return fig_width_in, fig_height_in


parser = argparse.ArgumentParser(description="plot_pred")
parser.add_argument("data")
parser.add_argument("fit")
parser.add_argument("output")
parser.add_argument("--selected-corpus", type=int, default=-1)

args = parser.parse_args()

with open(args.data, "rb") as fp:
    data = pickle.load(fp)

fit = pd.read_parquet(args.fit)

fig = plt.figure(figsize=set_size(450, 1, 1))
axes = [fig.add_subplot(4, 4, i + 1) for i in range(4 * 4)]

speakers = ["CHI", "OCH", "FEM", "MAL"]

n_groups = data["n_groups"]
n_corpora = data["n_corpora"]

for i in range(4 * 4):
    ax = axes[i]
    row = i // 4 + 1
    col = i % 4 + 1
    label = f"{col}.{row}"

    corpora = np.zeros((n_corpora, len(fit)))
    for c in range(n_corpora):
        corpora[c, :] = fit[f"corpus_bias.{c+1}.{label}"].values

    ax.set_xticks([])
    ax.set_xticklabels([])
    ax.set_yticks([])
    ax.set_yticklabels([])
    ax.set_ylim(0, 3)
    ax.set_xlim(-2, 2)

    if row == 1:
        ax.xaxis.tick_top()
        ax.set_xticks([0])
        ax.set_xticklabels([speakers[col - 1]])

    if row == 4:
        ax.set_xticks(np.linspace(-1, 1, 3, endpoint=True))
        ax.set_xticklabels(np.linspace(-1, 1, 3, endpoint=True))

    if col == 1:
        ax.set_yticks([1.5])
        ax.set_yticklabels([speakers[row - 1]])

    for c in range(n_corpora):
        if c != args.selected_corpus:
            ax.hist(
                corpora[c, :],
                bins=np.linspace(-2, 2, 40),
                density=True,
                histtype="step",
                color="gray",
                lw=0.5,
                alpha=0.5,
            )

    if args.selected_corpus >= 0:
        ax.hist(
            corpora[args.selected_corpus, :],
            bins=np.linspace(-2, 2, 40),
            density=True,
            histtype="step",
            color="red",
            lw=1,
        )

    # ax.axvline(np.mean(data), linestyle="--", linewidth=0.5, color="#333", alpha=1)
    # ax.text(0.5, 4.5, f"{low:.2f} - {high:.2f}", ha="center", va="center")

fig.suptitle("Corpora biases $b_{ij}$ distributions")
fig.subplots_adjust(wspace=0, hspace=0)

plt.savefig(args.output)

plt.show()