plot.py

import json
import matplotlib.pyplot as plt
import math
import numpy as np

# Read the results from the file
with open('results.jsonl', 'r') as f:
    results = [json.loads(line) for line in f.readlines()]

# Organize data by filter size
filter_size_to_ratios = {}
for result in results:
    filter_size = result['args']['filter_size']
    log2_size = int(math.log2(filter_size))

    unsorted_ms = result['unsorted']['milliseconds']
    sorted_ms = result['sorted']['milliseconds']
    ratio = unsorted_ms / sorted_ms

    if log2_size not in filter_size_to_ratios:
        filter_size_to_ratios[log2_size] = []

    filter_size_to_ratios[log2_size].append(ratio)

# Calculate mean and standard deviation for each filter size
x_values = []
y_values = []
y_errors = []

for log2_size, ratios in sorted(filter_size_to_ratios.items()):
    x_values.append(log2_size)
    y_values.append(np.mean(ratios))
    y_errors.append(np.std(ratios))

# Create the plot
plt.figure(figsize=(10, 6))
plt.errorbar(x_values, y_values, yerr=y_errors, fmt='o-', linewidth=2, markersize=8, capsize=5)
plt.axhline(y=1.0, color='r', linestyle='--', alpha=0.7)  # Reference line at ratio=1

# Add labels and title with increased font sizes
plt.xlabel('Bloom Filter Size (log2 bits)', fontsize=14)
plt.ylabel('Time Ratio (Unsorted/Sorted)', fontsize=14)
plt.grid(True, alpha=0.3)

# Format the y-axis to show ratio properly
plt.tight_layout()

# Save the plot
plt.savefig('speed_ratio.pdf')