"""
Plotting the scores
"""
import numpy as np
import pandas as pd
from matplotlib import pylab as plt
from matplotlib import use
from base64 import b64encode
from io import BytesIO
use("Agg") # Use non-interactive backend for mpl
[docs]
def plot_comparison_results(voter_model: str, results: dict, num_voters: int, num_candidates: int,
num_topn: int,
number_iterations: int, distortion_ratio = 0.0 , save_figure: bool = False):
"""
Plot the mean scores for all voting rules.
:param voter_model: The generative model to use.
:type voter_model: str
:param results: The voting rule result data to be plotted. Key is the index of iteration.
:type results: dict
:param num_voters: The number of voters in the model.
:type num_voters: int
:param num_candidates: The number of candidates in the model.
:type num_candidates: int
:param num_topn: The number of top candidates to consider.
:type num_topn: int
:param number_iterations: The number of iterations for the simulation.
:type number_iterations: int
:param save_figure: If True, saves the figure as a png file, defaults to False.
:type save_figure: bool, optional
:return: Base64 encoded string representation of the plot image.
:rtype: str
"""
# Aggregate results
voting_rules = list(results[0].keys())
results_agg = []
for rule in voting_rules:
for i in range(number_iterations):
r = results[i][rule]
r["rule"] = rule
results_agg.append(r)
df = pd.DataFrame(results_agg)
df = df.set_index("rule")
# Summarize the results using pivot_table
df_summary = df.pivot_table(values=["top", "topn"],
index=df.index.get_level_values(0),
aggfunc={"top": ["mean", "std"],
"topn": ["mean", "std"]})
df_summary.columns = df_summary.columns.map("_".join)
# Plot the scores for all voting rules
fig, axes = plt.subplots(1, 2, figsize=(10, 5))
fig.subplots_adjust(bottom=0.25)
df_summary.plot(y="top_mean", yerr="top_std", legend=False,
capsize=2, fmt="o-", color="b",
rot=90, ax=axes[0])
axes[0].set_xticks(np.arange(len(df_summary)))
axes[0].set_xticklabels(df_summary.index)
axes[0].set_xlabel("Voting rules")
axes[0].set_ylabel("Mean scores")
axes[0].set_title("Top mean")
axes[0].grid(color="gray", linestyle="dashed", linewidth=0.1)
df_summary.plot(y="topn_mean", yerr="topn_std", legend=False,
capsize=2, fmt="o-", color="g",
rot=90, ax=axes[1])
axes[1].set_xticks(np.arange(len(df_summary)))
axes[1].set_xticklabels(df_summary.index)
axes[1].set_xlabel("Voting rules")
axes[1].set_ylabel("Mean scores")
axes[1].set_title(f"Top{num_topn} mean")
axes[1].grid(color="gray", linestyle="dashed", linewidth=0.1)
fig.suptitle(
f"{num_voters} voters and {num_candidates} candidates. {number_iterations} iterations. " +
(f"Profiles are {voter_model} with distortion ratio {distortion_ratio}" if distortion_ratio != 0.0 else ""))
if save_figure:
plt.savefig(
f"figures/scores_{num_candidates}_{num_voters}_{voter_model}_{number_iterations}_{distortion_ratio if distortion_ratio != 0.0 else None}.png",
format="png", dpi=500)
# Instantiate tmp file
tmpfile = BytesIO()
plt.savefig(tmpfile, format="png", dpi=300)
encoded = b64encode(tmpfile.getvalue()).decode("utf-8")
return encoded