performance.py

from RaiseWikibase.dbconnection import DBConnection
from RaiseWikibase.datamodel import snak, claim, entity
from RaiseWikibase.raiser import batch
import random
import string
import time
import pandas as pd
import seaborn as sns


def item_claim(nc=1, extra=0):
    s = ''.join(random.choice(letters) for i in range(30))
    if extra == 0:
        q, r = [], []
    if extra == 1:
        q, r = [snak('string', s, 'P1')], []
    if extra == 2:
        q, r = [snak('string', s, 'P1')], [snak('string', s, 'P1')]
    edict = entity(labels={'en': {'language': 'en', 'value': s}},
                   aliases={'en': [{'language': 'en', 'value': s}]},
                   descriptions={'en': {'language': 'en', 'value': s}},
                   claims=claim('P1',
                                snak('string', s, 'P1'),
                                qualifiers=q,
                                references=r),
                   etype="item",
                   datatype='string')
    if nc >= 2:
        for n in range(2, nc + 1):
            edict['claims'].update(claim('P' + str(n),
                                         snak('string', s, 'P' + str(n)),
                                         qualifiers=q,
                                         references=r))
    return edict


if __name__ == "__main__":
    time1 = time.time()
    batch_lengths = [100] # a list of batch lenghts used in experiments; 
    # the length of batch 10000 makes it time consuming;
    # start with 100, then running both experiments takes 80 seconds
    dir_experiments = './experiments/'
    connection = DBConnection()
    ## Experiment 1
    letters = string.ascii_lowercase
    res = []
    for k in [1, 2, 3, 4, 5, 6]:  # number of a test
        for nc in [500, 1000, 1500, 2000, 2500, 3000, 3500, 4000, 4500, 5000, 5500, 6000]:  # length of a page
            t = ''.join(random.choice(letters) for i in range(nc))
            for nn in batch_lengths:  # length of the batch
                time0 = time.time()
                batch('wikitext', [t for m in range(nn)], 0, [str(random.random()) for m in range(nn)])
                nt = time.time() - time0
                res.append([nc, nn, nt, k])
    d = pd.DataFrame(res, columns=['Number of characters per page', 'Number of pages', 'Time, [s]', 'Run'])
    d['Speed in pages per second'] = d['Number of pages'] / d['Time, [s]']
    # Save data
    d.to_csv(dir_experiments + 'exp1.csv')
    # Plot data and save the figure
    sns.set_theme(style="darkgrid")
    sns.set_context("paper", font_scale=1.5, rc={"font.size": 15, "axes.titlesize": 15, "axes.labelsize": 15})
    ax = sns.scatterplot(data=d, y="Speed in pages per second", x="Number of characters per page", hue="Run",
                         style="Run", s=180, legend=False)
    ax.figure.savefig(dir_experiments + 'exp1.pdf', dpi=300, bbox_inches='tight')

    # Experiment 2
    letters = string.ascii_lowercase
    res = []  # [['nc','nn','nt']]
    for e in [0, 1, 2]:  # 0 - each claim with no qualifiers and no references,
        # 1 - each claim with one qualifer and no reference,
        # 2 - each claim with one qualifer and one reference.
        for k in [1, 2]:  # number of a test
            for nc in [1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21]:  # number of claims per item
                for nn in batch_lengths:  # length of the batch
                    it = item_claim(nc, e)
                    time0 = time.time()
                    batch('wikibase-item', [it for m in range(nn)])
                    nt = time.time() - time0
                    res.append([nc, nn, nt, k, e])
    d = pd.DataFrame(res, columns=['Number of claims per page', 'Number of pages', 'Time, [s]', 'Run', 'Extra'])
    d['Speed in pages per second'] = d['Number of pages'] / d['Time, [s]']
    # Save data
    d.to_csv(dir_experiments + 'exp2.csv')
    # Plot data and save the figure
    sns.set_theme(style="darkgrid")
    sns.set_context("paper", font_scale=1.5, rc={"font.size": 15, "axes.titlesize": 15, "axes.labelsize": 15})
    ax = sns.scatterplot(data=d, y="Speed in pages per second", x="Number of claims per page", hue="Run", style="Extra",
                         s=180, legend=False)
    ax.figure.savefig(dir_experiments + 'exp2.pdf', dpi=300, bbox_inches='tight')

    print(time.time()-time1)