m2-traitement-benchmark/main.py

from elasticsearch import Elasticsearch
from psycopg2 import connect
from drivers.psql import PSQL_Testing
from drivers.elastic import Elastic_Testing
from sshtunnel import SSHTunnelForwarder
from utils.utils import preprocess_json
from typing_extensions import Dict
from collections import defaultdict
import numpy as np
import matplotlib.pyplot as plt

POSTGRES_HOST = "127.0.0.1"
POSTGRES_PORT = "5432"
POSTGRES_DB = "postgres"
POSTGRES_USER = "postgres"
POSTGRES_PASSWORD = "postgres"

ELASTIC_URL = "https://localhost:9200"
ELASTIC_API_KEY = 'Z18xZzFJOEJXTnUzZ2RiTk5YWkw6ekhiRVpQYnVTZ2FhRjFCR3NVUFB4UQ=='
RUN_ELASTIC_WITH_N_SHARDS = [1, 2, 4]

DOCUMENT_MULTIPLICATOR = 10

def test_psql():
            psqlClient = connect(
                database=POSTGRES_DB,
                host=POSTGRES_HOST,
                user=POSTGRES_USER,
                password=POSTGRES_PASSWORD,
                port=POSTGRES_PORT
            )
            psqlClient.autocommit = False
            return PSQL_Testing().do_tests(psqlClient)

def test_elasticsearch(shards = 1):
    es = Elasticsearch(
        ELASTIC_URL,
        api_key=ELASTIC_API_KEY,
        verify_certs=False, # just to not create certificates
        ssl_show_warn=False
    )
    return Elastic_Testing(shards).do_tests(es)


def plot(timings: Dict[str, Dict[str, float]]):
    functions = list(timings['PSQL'].keys())
    drivers = list(timings.keys())

    values = {func: [] for func in functions}

    for func in functions:
        values[func].append(timings['PSQL'][func]);

    for driver in [x for x in drivers if x != 'PSQL']:
        for func in functions:
            values[func].append(timings[driver][func] / values[func][0])

    for func in functions:
        values[func][0] = 1

    fig, ax = plt.subplots(figsize=(12, 8)) 
    index = np.arange(len(functions))
    bar_width = 0.2

    for i, driver in enumerate(drivers):
        ax.bar(index + i * bar_width, [values[func][i] for func in functions], bar_width, label=driver)


    ax.set_xlabel('Functions')
    ax.set_ylabel('Relative Time')
    ax.set_title('Performance of ES Relative to PSQL')
    ax.set_xticks(index + bar_width * (len(drivers) -1 ) / 2, functions, rotation=45, ha="right")
    ax.set_xticklabels(functions)
    ax.legend()

    for i, x in enumerate(drivers):
        for j, v in enumerate([values[func][i] for func in functions]):
            plt.text(j + (i * bar_width), v + 0.02, str(round(v, 2)), fontsize=6, ha='center', va='bottom')

    fig.tight_layout()
    plt.savefig("plot-orodruin-opti.png")
    plt.show()


if __name__ == "__main__":
    
    print("### Preprocessing JSON ###")
    preprocess_json(DOCUMENT_MULTIPLICATOR)
    
    
    timings = {}
    
    timings['PSQL'] = test_psql()
    for x in RUN_ELASTIC_WITH_N_SHARDS:
        timings[f'ES{x}shards'] = test_elasticsearch(x)

    plot(timings)
    print(timings)