Configurable client

[danielSanchezQ]

This PR adds the ability to allow overriding the reqwest::Client used by the proxy.

[danielSanchezQ]

Implements #44

[danielSanchezQ]

I made some basic benchmark to check what would be the penalisation:

use criterion::async_executor::AsyncExecutor;
use criterion::{criterion_group, criterion_main, Criterion};
use once_cell::sync::OnceCell;

async fn client_plain(client: &reqwest::Client) {
    client
        .get("https://www.google.es")
        .send()
        .await
        .expect("Couldn't get google response");
}

async fn client_once_cell(client: &OnceCell<reqwest::Client>) {
    client
        .get_or_init(reqwest::Client::new)
        .get("https://www.google.es")
        .send()
        .await
        .expect("Couldn't get google response");
}

pub fn client_plain_benchmark(c: &mut Criterion) {
    let client = reqwest::Client::new();
    let runtime = tokio::runtime::Runtime::new().unwrap();
    let mut group = c.benchmark_group("Clients");
    let once_client = OnceCell::new();
    group.bench_function("plain client 10", |b| {
        b.to_async(&runtime).iter(|| client_plain(&client))
    });
    group.bench_function("once client 10", |b| {
        b.to_async(&runtime).iter(|| client_once_cell(&once_client))
    });
}

criterion_group!(benches, client_plain_benchmark);
criterion_main!(benches);

Results show negligible performance differences:

Benchmarking Clients/plain client 10: Collecting 100 samples in estimated 23.022 s (100 iterations)
Benchmarking Clients/plain client 10: Analyzing
Clients/plain client 10 time:   [213.32 ms 216.64 ms 220.36 ms]
Found 10 outliers among 100 measurements (10.00%)
  6 (6.00%) high mild
  4 (4.00%) high severe
Benchmarking Clients/once client 10
Benchmarking Clients/once client 10: Warming up for 3.0000 s

Benchmarking Clients/once client 10: Collecting 100 samples in estimated 21.577 s (100 iterations)
Benchmarking Clients/once client 10: Analyzing
Clients/once client 10  time:   [206.35 ms 209.60 ms 213.54 ms]
Found 7 outliers among 100 measurements (7.00%)
  5 (5.00%) high mild
  2 (2.00%) high severe

Safe to merge ❤️