update_plot.py

from tkinter.tix import Tree
from tkinter.ttk import LabeledScale
from turtle import color
import matplotlib
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
from datetime import datetime
import pytz
import os
# args
# 选择提取哪些数据集，会往本地写数据，文件：'./result_update/[date]_[measurement].data'
datasets = ["indochina-2004"]
# "indochina-2004", "tags-stack-overflow", "threads-stack-overflow", "colisten-Spotify", "twitter-2010" , "temporal-reddit-reply",
log_measures = ['memory', 'memory_overhead', 'time', 'time_overhead']
plot_measures = ['memory', 'memory_overhead', 'time', 'time_overhead']

algos_to_read= ['CoinFlipWalk', 'PrefixWalk', 'RejectionWalk','readgraph']
algos = ['CoinFlipWalk', 'PrefixWalk', 'RejectionWalk']
algo_alias = ['CoinFlipWalk', 'PrefixWalk', 'RejectionWalk']
# algos = ['AliasWalk', 'CoinFlipWalk', 'PrefixWalk', 'RejectionWalk']
# algos = ['MCAM', 'MCSS', 'MCPS', 'MCAR']
# algo_alias = ['AliasWalk', 'CoinFlipWalk', 'PrefixWalk', 'RejectionWalk']
datasets_alias = {
    "twitter-2010": 'TW',
    "threads-stack-overflow": 'TH',
    "colisten-Spotify": 'CS',
    "indochina-2004": 'IC',
    "bitcoin-temporal": 'BT',
    "coauth-MAG": 'MAG',
    "coauth-AMiner": 'AM',
    "tags-stack-overflow": "TAG",
    "youtube": "YT",
    "soc-LiveJournal1": "LJ",
    "orkut-links": 'OL',
    "block_ethereum": 'ET',
    "blockchair": 'BC',
    "clickstream": "CL",
    "temporal-reddit-reply": 'RR'
}
tz = pytz.timezone('Asia/Shanghai')
# 避免生成的pdf在苹果系统中无法正确显示线形
matplotlib.rcParams['pdf.fonttype'] = 42
matplotlib.rcParams['ps.fonttype'] = 42

# RejectionWalk的位置，求overhead的时候需要减去
to_sub = 0
for i, algo in enumerate(algos_to_read):
    if  algo == 'readgraph':
        to_sub = i
        break
print("origin graph data at",to_sub)
time = []
memory = []
for dataset in datasets:
    tmp_time = []
    tmp_memory = []
    for algo in algos_to_read:
        print(dataset)
        print(algo)
        filename = './runlog/' + algo + '_' + dataset + '_update.log'
        tmp_del_time = 0
        tmp_add_time = 0
        if algo=='CoinFlipWalk':
            add_idx=5
            del_idx=6
            memory_idx=7
        else:
            add_idx=4
            del_idx=5
            memory_idx=6
        with open(filename, 'r') as f:
            for i, line in enumerate(f.readlines()):
                if i == del_idx:  # del time
                    tmp_del_time = float(line.split()[5])
                    print(tmp_del_time)
                if i == add_idx:
                    tmp_add_time = float(line.split()[5])
                    print(tmp_add_time)
                if i == memory_idx:
                    tmp_memory.append(float(line.split()[4]))
                    break
        tmp_time.append((tmp_del_time + tmp_add_time) / 2)
    time.append(tmp_time)
    memory.append(tmp_memory)
print(time)
print(memory)
time_overhead = [[j - i[to_sub] for j in i] for i in time]
memory_overhead = [[j - i[to_sub] for j in i] for i in memory]
print(time_overhead)
print(memory_overhead)

dt = datetime.now(tz).strftime("%m-%d")
path = "./result_update/" + dt + '/'
if (os.path.exists(path) == False):
    os.makedirs(path)
for measure in log_measures:
    isTime = False if measure.find("time") == -1 else True
    isOverhead = False if measure.find("overhead") == -1 else True
    if isTime and not isOverhead:
        filename = dt + "_" + "time.data"
        arr = time
    elif isTime and isOverhead:
        filename = dt + "_" + "time_overhead.data"
        arr = time_overhead
    elif not isTime and not isOverhead:
        filename = dt + "_" + "memory.data"
        arr = memory
    elif not isTime and isOverhead:
        filename = dt + "_" + "memory_overhead.data"
        arr = memory_overhead
    
    f = open(path + filename, 'w')
    for i, subarr in enumerate(arr):
        f.write(datasets[i] + ':')
        f.write(str(subarr))
        f.write('\n')
    f.close()

num_dataset = len(datasets)
# colors = [(0, 0.4470, 0.7410), (0.8500, 0.3250, 0.0980),
#           (0.4940, 0.1840, 0.5560), (0.6353, 0.07843, 0.1843)]
# colors = ['#263FA9', '#447B48', '#E85472', '#A686EC']
colors = ['#1F1DBF', '#E85472', '#447B48']
# ylims = [50, 500, 40, 6]#8382EB
#EC7F4B
#263FA9
#447B48
for measure in plot_measures:
    isTime = False if measure.find("time") == -1 else True
    isOverhead = False if measure.find("overhead") == -1 else True
    if isTime and not isOverhead:
        arr = time
    elif isTime and isOverhead:
        arr = time_overhead
    elif not isTime and not isOverhead:
        arr = memory
    elif not isTime and isOverhead:
        arr = memory_overhead
    for k, dataset in enumerate(datasets):
        fig, ax = plt.subplots(figsize=(10, 5))
        ax.bar(algos,
               arr[k][:3],
               tick_label=algo_alias,
               color='none',
               log=isTime,
               linewidth=3)
        ax.set_yticklabels([],minor=True)
        if measure=='memory_overhead' and datasets_alias[dataset]=='IC':
            ax.set_yticks([0,5,10,15,20])
        if measure=='time_overhead':
            if datasets_alias[dataset]=='IC' or datasets_alias[dataset]=='TAG':
                ax.set_yticks([1e-7,1e-6])
            elif datasets_alias[dataset]=='TW':
                ax.set_yticks([1e-6])
            elif datasets_alias[dataset]=='TH' :
                ax.set_yticks([1e-6,1e-5])
            elif datasets_alias[dataset]=='RR':
                ax.set_yticks([1e-6])
                ax.set_yticks([2*1e-6,3*1e-6,4*1e-6,5*1e-6,],labels=['2e-6','3e-6','','5e-6',],minor=True,fontsize=25)
        
        ax.tick_params(axis='x', labelsize=30, rotation=15)
        plt.yticks(fontsize=30)
        if isTime and isOverhead:
            ax.set_ylabel("update overhead (s)", fontsize=30)
        elif isTime and not isOverhead:
            ax.set_ylabel("time (s)", fontsize=30)
        elif not isTime and isOverhead:
            ax.set_ylabel("memory overhead (GB)", fontsize=26)
        else:
            ax.set_ylabel("memory (GB)", fontsize=30)
        hatches = ["/", "x", ".", "\\"]
        ax.set_title(datasets_alias[dataset], fontsize=35)
        for i, patch in enumerate(ax.patches):
            patch.set_hatch(hatches[i])
            patch.set_edgecolor(colors[i])
        plt.rcParams['hatch.linewidth'] = 3.0
        plt.tight_layout()
        plt.savefig(path +
                    "{}_{}.png".format(measure, datasets_alias[dataset]))
        plt.savefig(path + "{}_{}.pdf".format(measure, datasets_alias[dataset]))