plot_benchmark.py

import csv
from ast import literal_eval
import numpy as np
import matplotlib
import matplotlib.pyplot as plt
import sys
import os

gdv_name = "TFL"

# Large compare plot
fig_large, ax_large = plt.subplots()
large_i = 0

# This is filled by the plot_experiment_results
depth_ratio = {
    "cirq": [0] * 9,
    "qiskit": [0] * 9,
    "tket": [0] * 9,
    "jku": [0] * 9,
    "k7m": [0] * 9
}

# The optimal depth is known
depth_range = {
    "TFL" : [5 * x for x in range(1, 10)],
    "QSE" : [100 * x for x in range(1, 10)]
}

other_tools = ["cirq", "qiskit", "tket", "jku"]

def plot_experiment_results(benchmark_name):
    global large_i
    large_i += 1

    print(large_i, benchmark_name)
    folder_name = os.path.dirname(benchmark_name)

    global depth_ratio
    global depth_range

    # for future qiskit experiment
    dataset = list()

    # with open("_private_data/BNTF/{}".format(benchmark_name), 'r') as csvfile:
    with open(benchmark_name, 'r') as csvfile:
        for row in csv.reader(csvfile, delimiter=','):
            data = list()
            for i in range(len(row)):
                if i is not 1:
                    data.append(literal_eval(row[i]))
                else:
                    data.append(row[i])

            dataset.append(data)


    """
    Generate data for k7m
    """
    fig, ax = plt.subplots()

    # Reset
    depth_ratio["k7m"] = [0] * 9

    for tool in ["k7m"]:
        for i in range(10):
            depth = depth_range[gdv_name][i]
            # optimal_depth[i] = depth
            count_data = 0
            for data in dataset:
                if data[1] == tool and data[2] == depth:
                    count_data += 1
                    depth_ratio[tool][i] += data[3] / data[2]

                    ax.plot(data[2], data[3] / data[2], 'o', color="lightgreen")
                    # with open("_private_data/BNTF/{}_{}.csv".format(gdv_name, tool), 'a') as csvfile:
                    #     csv.writer(csvfile).writerow([data[2], data[3] / data[2]])
                    # csvfile.close()
            depth_ratio[tool][i] /= count_data

        """
        Save AVG
        """
        # with open("{}_{}.avg".format(benchmark_name, tool), 'w') as csvfile:
        with open("{}/{}_{}.avg".format(folder_name,large_i ,tool), 'w') as csvfile:
            writer = csv.writer(csvfile)
            for i in range(9):
                depth = depth_range[gdv_name][i]
                writer.writerow([depth, depth_ratio[tool][i]])
        csvfile.close()

    """
        Plot the graph
    """
    for tool in other_tools + ["k7m"]:
        ax.plot(depth_range[gdv_name], depth_ratio[tool], label=tool)

    #Include this plot also on the large one
    legend = os.path.splitext(os.path.basename(benchmark_name))[0][1:]
    ax_large.plot(depth_range[gdv_name], depth_ratio["k7m"],
                  label=legend)
                  # label=os.path.basename(benchmark_name))

    ax.set(xlabel='Optimal Depth', ylabel='Depth Ratio')
    if len(other_tools) > 0:
        ax.legend()

    # fig.savefig('_private_data/BNTF/{}.png'.format(gdv_name), dpi=300)
    png_name = os.path.basename(benchmark_name)
    folder_name = os.path.dirname(benchmark_name)
    # fig.savefig('{}/{}{}.png'.format(folder_name,large_i,png_name), dpi=150)
    fig.savefig('{}/{}.png'.format(folder_name, large_i), dpi=150)


def load_others():
    """
    Load other files
    """
    global depth_ratio

    for tool in other_tools:
        counts = [0] * 9
        with open("_private_data/BNTF/{}_{}.csv".format(gdv_name, tool), 'r') as csvfile:
            reader = csv.reader(csvfile)
            for row in reader:
                i = (int(row[0]) - 5)//5
                depth_ratio[tool][i] += float(row[1])
                counts[i] += 1
        for i in range(9):
            depth_ratio[tool][i] /= counts[i]


def plot_others():
    # for tool in other_tools:
    #     ax_large.plot(optimal_depth, depth_ratio[tool], label=large_i)

    """
    This is the large plot
    """
    ax_large.grid(True)
    ax_large.set(xlabel='Optimal Depth', ylabel='Depth Ratio')
    # ax_large.set_ylim(3.5, 6)
    ax_large.legend()

    fig_large.savefig('compare_all.png', dpi=150)



"""
    MAIN
"""
# load_others()
files = []
results_folder = "_private_data/BNTF/"
for file in os.listdir(results_folder):
    if file.startswith("_") and file.endswith(".csv") and (gdv_name in file):
        files.append(file)

for file in sorted(files):
    bench_name = os.path.join(results_folder, file)
    plot_experiment_results(bench_name)


plot_others()