Initial push

praxis3/requirements.txt

@@ -0,0 +1,6 @@
+numpy>=1.20.3
+pandas
+requests
+zmq
+pytest
+pytest-timeout

praxis3/test/check_submission.sh

@@ -0,0 +1,43 @@
+#!/bin/bash
+
+set -e
+
+if [[ ! $# -eq 1 ]]; then
+    echo "Missing argument."
+    echo "Usage: '$0 praxis{0,1,2,3}'"
+    exit 1
+fi
+
+TEST_SCRIPT="$(dirname $0)/test_$1.py"
+
+if [[ ! -e ${TEST_SCRIPT} ]]; then
+    echo "Requested test file ${TEST_SCRIPT} does not exist!"
+    exit 1
+fi
+
+echo "Packaging submission..."
+PACKAGING_DIR=$(mktemp -d)
+cmake -B${PACKAGING_DIR}
+cmake --build ${PACKAGING_DIR} -t package_source
+
+echo "Extracting submission..."
+SOURCE_DIR=$(mktemp -d)
+(cd ${SOURCE_DIR} && find ${PACKAGING_DIR} -name '*.tar.gz' | head -n1 | xargs tar -xzf)
+EXTRACTED_SOURCE_DIR=$(find ${SOURCE_DIR} -maxdepth 1 -mindepth 1 -type d | head -n1)
+
+echo "Building submission..."
+BUILD_DIR=$(mktemp -d)
+cmake -S"${EXTRACTED_SOURCE_DIR}" -B${BUILD_DIR}
+TARGET="webserver"
+EXECUTABLE="--executable ${BUILD_DIR}/webserver"
+if [[ "$1" == "praxis0" ]]; then
+	TARGET="hello_world"
+	EXECUTABLE="--executable ${BUILD_DIR}/hello_world"
+elif [[ "$1" == "praxis3" ]]; then
+	TARGET="zmq_distributor zmq_worker"
+	EXECUTABLE="--dist_exec ${BUILD_DIR}/zmq_distributor --work_exec ${BUILD_DIR}/zmq_worker"
+fi
+cmake --build ${BUILD_DIR} -t ${TARGET}
+
+echo "Executing tests..."
+python3 -m pytest -o cache_dir=${BUILD_DIR} ${TEST_SCRIPT} ${EXECUTABLE}

praxis3/test/conftest.py

@@ -0,0 +1,7 @@
+import pytest
+
+
+def pytest_addoption(parser):
+    parser.addoption('--dist_exec', action='store', default='build/zmq_distributor')
+    parser.addoption('--work_exec', action='store', default='build/zmq_worker')
+    parser.addoption('--debug_test', action='store_true', default=False)

praxis3/test/test_praxis3.py

@@ -0,0 +1,515 @@
+"""
+Tests for RN Praxis 3
+"""
+
+import multiprocessing
+from sys import stderr
+
+import numpy as np
+import pytest
+import subprocess
+import sys
+import time
+import util
+import zmq
+
+zmq.Context()
+
+test_args = None
+distributor_valgrind_test_passed = False
+debug_tests = False
+
+
+@pytest.fixture
+def program_args(request):
+    d_exec = request.config.option.dist_exec
+    w_exec = request.config.option.work_exec
+    global debug_tests
+    debug_tests = request.config.option.debug_test
+    # cache = request.config.option.cache_texts
+    # cache_dir = request.config.option.cache_dir
+    return {"distributor": d_exec, "worker": w_exec}
+
+
+@pytest.mark.timeout(15)
+def test_workers_reachable(program_args):
+    global test_args
+    test_args = util.init_tests(program_args)
+
+    base_port = test_args["base_port"]
+
+    for num_workers in [1, 2, 4, 8]:
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        # launch worker here
+        port_list = [str(x) for x in workers]
+        worker_procs = util.start_threaded_workers(test_args["worker"], port_list)
+
+        workers_reached = 0
+
+        test_message = "map\0"
+        rip_message = "rip\0"
+        context = zmq.Context.instance()
+        for w in workers:
+            socket = context.socket(zmq.REQ)
+            socket.setsockopt(zmq.LINGER, 0)
+            socket.setsockopt(zmq.RCVTIMEO, 500)
+            socket.connect("tcp://127.0.0.1:" + str(w))
+            socket.send(bytes(test_message, "ascii"))
+
+            message = None
+            try:
+                message = socket.recv()
+            except zmq.ZMQError:
+                print("Failed to receive reply from worker " + str(w))
+                socket.close()
+                continue
+
+            reply = message[:-1].decode("ascii")
+            if reply == "":
+                workers_reached += 1
+
+            # rip worker
+            socket.send(bytes(rip_message, "ascii"))
+            try:
+                socket.recv()
+            except zmq.ZMQError:
+                print("Failed to receive rip from worker " + str(w))
+
+            socket.close()
+
+        util.join_workers(worker_procs)
+
+        assert workers_reached == num_workers, f"Failed to reach all workers for {num_workers} worker threads! Expected {num_workers} but reached {workers_reached}."
+
+
+@pytest.mark.timeout(15)
+def test_distributor_message_size():
+    word_list = util.get_part_of_word_list(test_args["word_list"], 250)
+
+    simple_delimiters = test_args["simple_delimiters"]
+
+    max_msg_size = 1500
+    max_string_length = 50 * max_msg_size
+
+    test_string = util.generate_text_from_word_list(word_list, simple_delimiters, max_string_length)
+
+    filename = test_args["filename_simple"]
+    f = open(filename, "w")
+    f.write(test_string)
+    f.close()
+
+    port = str(test_args["base_port"])
+
+    # kill any zmq procs currently running
+    util.kill_zmq_distributor_and_worker()
+
+    proc = util.start_distributor([test_args["distributor"], filename, port])
+
+    context = zmq.Context.instance()
+    socket = context.socket(zmq.REP)
+    socket.bind("tcp://*:" + port)
+
+    map_response = "test1" * (int(max_msg_size/5)-1) + "\0"
+
+    message_lengths_received = []
+
+    while True:
+        msg = socket.recv()
+        message_lengths_received.append(len(msg))
+        req_type = msg[0:3].decode("ascii")
+
+        if req_type == "rip":
+            socket.send(b"rip\0")
+            break
+        elif req_type == "map":
+            socket.send(map_response.encode("ascii"))
+        elif req_type == "red":
+            socket.send(b"test1\0")
+    socket.close()
+    proc.wait()
+
+    for l in message_lengths_received:
+        assert l <= max_msg_size, f"Received message size exceeds max message size of {max_msg_size} bytes."
+
+
+@pytest.mark.timeout(60)
+def test_simple_text():
+    filename_simple = test_args["filename_simple"]
+
+    f = open(filename_simple, "r")
+    simple_text = f.read()
+    f.close()
+
+    base_port = test_args["base_port"]
+
+    for num_workers in [1, 2, 4, 8]:
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+        port_list = [str(x) for x in workers]
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        # launch worker and distributor
+        worker_procs = util.start_threaded_workers(test_args["worker"], port_list)
+        proc_distributor = util.start_distributor([test_args["distributor"], filename_simple] +
+                                      port_list)
+
+        util.join_workers(worker_procs)
+
+        distributor_output, distributor_err = proc_distributor.communicate()
+        correct_word_count = util.count_words(simple_text)
+
+        if debug_tests:
+            util.create_test_debug_output("test_simple_text", num_workers, correct_word_count, distributor_output)
+
+        assert distributor_output == correct_word_count, f"{num_workers} workers failed simple text test."
+
+
+@pytest.mark.timeout(60)
+def test_complex_text():
+    word_list = util.get_part_of_word_list(test_args["word_list"], 2500)
+    all_delimiters = test_args["all_delimiters"]
+    max_string_length = 2.5e5
+
+    test_string = util.generate_text_from_word_list(word_list, all_delimiters, max_string_length)
+
+    filename = test_args["filename_complex"]
+
+    f = open(filename, "w")
+    f.write(test_string)
+    f.close()
+
+    base_port = test_args["base_port"]
+
+    for num_workers in [2, 4, 8]:
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+        port_list = [str(x) for x in workers]
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        # launch worker and distributor
+        worker_procs = util.start_threaded_workers(test_args["worker"], port_list)
+        proc_distributor = util.start_distributor([test_args["distributor"], filename] +
+                                      port_list)
+
+        util.join_workers(worker_procs)
+
+        distributor_output, distributor_err = proc_distributor.communicate()
+        correct_word_count = util.count_words(test_string)
+
+        if debug_tests:
+            util.create_test_debug_output("test_complex_text", num_workers, correct_word_count, distributor_output)
+
+        assert distributor_output == correct_word_count, f"{num_workers} workers failed complex text test."
+
+
+@pytest.mark.timeout(90)
+def test_book_1():
+    filename = test_args["filename_book_1"]
+    base_port = test_args["base_port"]
+    book_text = test_args["books"][0]
+
+    file_out = open(filename, "wb")
+    file_out.write(book_text)
+    file_out.close()
+
+    for num_workers in [2, 4, 8]:
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+        port_list = [str(x) for x in workers]
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        # launch worker and distributor
+        worker_procs = util.start_threaded_workers(test_args["worker"], port_list)
+        proc_distributor = util.start_distributor([test_args["distributor"], filename] +
+                                           port_list)
+
+        util.join_workers(worker_procs)
+
+        distributor_output, distributor_err = proc_distributor.communicate()
+
+        book_text_str = book_text.decode("ascii", errors="ignore")
+        correct_word_count = util.count_words(book_text_str)
+
+        if debug_tests:
+            util.create_test_debug_output("test_book_1", num_workers, correct_word_count, distributor_output)
+
+        assert distributor_output == correct_word_count, f"{num_workers} workers failed book 1 test."
+
+
+@pytest.mark.timeout(90)
+def test_book_2():
+    filename = test_args["filename_book_2"]
+    base_port = test_args["base_port"]
+    book_text = test_args["books"][1]
+
+    file_out = open(filename, "wb")
+    file_out.write(book_text)
+    file_out.close()
+
+    for num_workers in [2, 4, 8]:
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+        port_list = [str(x) for x in workers]
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        # launch worker and distributor
+        worker_procs = util.start_threaded_workers(test_args["worker"], port_list)
+        proc_distributor = util.start_distributor([test_args["distributor"], filename] +
+                                           port_list)
+
+        util.join_workers(worker_procs)
+
+        distributor_output, distributor_err = proc_distributor.communicate()
+
+        book_text_str = book_text.decode("ascii", errors="ignore")
+        correct_word_count = util.count_words(book_text_str)
+
+        if debug_tests:
+            util.create_test_debug_output("test_book_2", num_workers, correct_word_count, distributor_output)
+
+        assert distributor_output == correct_word_count, f"{num_workers} workers failed book 2 test."
+
+
+@pytest.mark.timeout(30)
+def test_interoperability():
+    base_port = test_args["base_port"]
+
+    # test workers first
+    for num_workers in [2, 4, 8]:
+
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+        port_list = [str(x) for x in workers]
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        worker_procs = util.start_threaded_workers(test_args["worker"], port_list)
+
+        context = zmq.Context.instance()
+
+        map_message = "mapInteroperability test. Test uses python distributor.\0"
+        map_message_expected_return = "interoperability1test11uses1python1distributor1\0"
+        reduce_message_expected_return = "interoperability1test2uses1python1distributor1\0"
+        rip_message = "rip\0"
+
+        worker_responses = dict()
+
+        for port in port_list:
+            curr_responses = []
+
+            socket = context.socket(zmq.REQ)
+            socket.connect("tcp://127.0.0.1:" + port)
+
+            socket.send(bytes(map_message, "ascii"))
+            message = socket.recv()
+            message = message.decode("ascii")
+            curr_responses.append(message)
+            # assert message == map_message_expected_return
+
+            socket.send(bytes("red" + map_message_expected_return, "ascii"))
+            message = socket.recv()
+            message = message.decode("ascii")
+            curr_responses.append(message)
+            # assert message == reduce_message_expected_return
+
+            socket.send(bytes(rip_message, "ascii"))
+            message = socket.recv()
+            message = message.decode("ascii")
+            curr_responses.append(message)
+            # assert message == rip_message
+
+            socket.close()
+
+            worker_responses[port] = curr_responses
+
+        util.join_workers(worker_procs)
+
+        for res in worker_responses.values():
+            assert res[0] == map_message_expected_return
+            assert res[1] == reduce_message_expected_return
+            assert res[2] == rip_message
+
+    # test distributor second
+    port_list = [str(base_port)]
+
+    # kill any zmq procs currently running
+    util.kill_zmq_distributor_and_worker()
+
+    context = zmq.Context.instance()
+    socket = context.socket(zmq.REP)
+    socket.bind("tcp://*:" + str(base_port))
+
+    map_message = "mapInteroperability test. Test uses python distributor.\0"
+    map_message_expected_return = "interoperability1test11uses1python1distributor1\0"
+    reduce_message_expected_return = "interoperability1test2uses1python1distributor1\0"
+    rip_message = "rip\0"
+
+    distributor_messages = []
+
+    out_filename = test_args["filename_interop"]
+    f = open(out_filename, "w")
+    f.write(map_message[3:-1])
+    f.close()
+    proc_distributor = util.start_distributor([test_args["distributor"], out_filename] + port_list)
+
+    message = socket.recv()
+    message = message.decode("ascii")
+    # print(message)
+    # assert message == map_message
+    distributor_messages.append(message)
+    socket.send(bytes(map_message_expected_return, "ascii"))
+
+    message = socket.recv()
+    message = message.decode("ascii")
+    # print(message)
+    # assert message == "red" + map_message_expected_return
+    distributor_messages.append(message)
+    socket.send(bytes(reduce_message_expected_return, "ascii"))
+
+    message = socket.recv()
+    message = message.decode("ascii")
+    # print(message)
+    # assert message == rip_message
+    distributor_messages.append(message)
+    socket.send(bytes(rip_message, "ascii"))
+
+    socket.close()
+
+    distributor_output, distributor_err = proc_distributor.communicate()
+
+    assert distributor_messages[0] == map_message
+    assert distributor_messages[1] == "red" + map_message_expected_return
+    assert distributor_messages[2] == rip_message
+
+    assert distributor_output == util.count_words(map_message[3:])
+
+
+@pytest.mark.timeout(60)
+def test_load_distribution():
+    base_port = test_args["base_port"]
+
+    for num_workers in [2, 4, 8]:
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+        port_list = [str(x) for x in workers]
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        stds = []
+
+        # run test 3 times, because sometimes bad scheduling may affect load distribution
+        for i in range(3):
+            manager = multiprocessing.Manager()
+            return_dict = manager.dict()
+            worker_threads = []
+            for p in port_list:
+                proc = multiprocessing.Process(target=util.run_worker_load_distribution, args=(p, return_dict))
+                worker_threads.append(proc)
+                proc.start()
+
+            proc_distributor = util.start_distributor([test_args["distributor"], test_args["filename_complex"]] + port_list)
+
+            proc_distributor.wait()
+            for t in worker_threads:
+                t.join()
+
+            requests_per_worker = return_dict.values()
+            curr_std = np.std(requests_per_worker)
+            stds.append(curr_std)
+            # stop because std is already satisfying condition, no need for extra runs
+            if curr_std <= 2:
+                break
+
+        std = np.min(stds)
+        assert std <= 2
+
+
+@pytest.mark.timeout(120)
+def test_memory_leaks():
+    global test_args
+
+    valgrind_output_file = "valgrind_output"
+    valgrind_command_base = ["valgrind", "--tool=memcheck", "--xml=yes", "--xml-file=" + valgrind_output_file,
+                        "--gen-suppressions=all", "--leak-check=full", "--leak-resolution=med", "--track-origins=yes",
+                        "--trace-children=yes", "--vgdb=no", "--fair-sched=yes"]
+
+    valgrind_command_distributor = valgrind_command_base.copy()
+    valgrind_command_distributor[3] = "--xml-file=" + valgrind_output_file + "_distributor.xml"
+
+    test_args["filename_valgrind_distributor"] = valgrind_command_distributor[3].split("=")[1]
+
+    valgrind_commands_worker = []
+    valgrind_worker_out_files = []
+    for i in range(8):
+        curr_command = valgrind_command_base.copy()
+        curr_filename = valgrind_output_file + "_worker_" + str(i) + ".xml"
+        test_args["filename_valgrind_worker" + str(i)] = curr_filename
+        valgrind_worker_out_files.append(curr_filename)
+        curr_command[3] = "--xml-file=" + curr_filename
+        curr_command.append(test_args["worker"])
+        valgrind_commands_worker.append(curr_command)
+
+    word_list = util.get_part_of_word_list(test_args["word_list"], 500)
+    simple_delimiters = test_args["simple_delimiters"]
+    text_length = 100e3
+
+    test_string = util.generate_text_from_word_list(word_list, simple_delimiters, text_length)
+
+    valgrind_test_filename = test_args["filename_valgrind"]
+    f = open(valgrind_test_filename, "w")
+    f.write(test_string)
+    f.close()
+
+    base_port = test_args["base_port"]
+
+    worker_results = []
+    distributor_results = []
+
+    for num_workers in [2, 4, 8]:
+
+        workers = np.arange(base_port, base_port + num_workers).tolist()
+        port_list = [str(x) for x in workers]
+
+        # kill any zmq procs currently running
+        util.kill_zmq_distributor_and_worker()
+
+        # launch worker and distributor
+        proc_workers = util.start_threaded_workers(valgrind_commands_worker[:num_workers], port_list)
+        proc_distributor = util.start_distributor(valgrind_command_distributor + [test_args["distributor"], valgrind_test_filename] +
+                                      port_list)
+
+        proc_distributor.wait()
+        util.join_workers(proc_workers)
+
+        for i in range(num_workers):
+            worker_results.append(util.check_valgrind_output_no_errors(valgrind_worker_out_files[i]))
+        distributor_results.append(util.check_valgrind_output_no_errors(valgrind_command_distributor[3].split("=")[1]))
+
+    assert int(np.mean(worker_results)) == 1, f"Worker failed to pass valgrind test!"
+
+    global distributor_valgrind_test_passed
+    distributor_valgrind_test_passed = int(np.mean(distributor_results)) == 1
+
+
+@pytest.mark.timeout(3)
+def test_valgrind_second_point_and_cleanup():
+    time.sleep(1)
+    # kill any zmq procs currently running, including memcheck this time
+    util.kill_zmq_distributor_and_worker(kill_memcheck=True)
+
+    context = zmq.Context.instance()
+    context.destroy()
+
+    if not debug_tests:
+        remove_valgrind_xml_files = True
+        util.cleanup_files(test_args, remove_valgrind_xml_files)
+
+    assert distributor_valgrind_test_passed, f"Distributor failed to pass valgrind test!"

praxis3/test/util.py

@@ -0,0 +1,301 @@
+import multiprocessing
+import os
+import random
+import re
+import signal
+import string
+from io import StringIO
+import subprocess
+import sys
+import time
+import xml.etree.ElementTree as ET
+
+from collections import Counter
+from typing import List, Union
+
+import numpy as np
+import pandas as pd
+import requests
+import shutil
+import zmq
+
+
+def init_tests(args):
+    random.seed()
+
+    distributor_exec = args["distributor"]
+    worker_exec = args["worker"]
+    # cache = args["cache"]
+    # cache_dir = args["cache_dir"]
+
+    simple_delimiters = [", ", ". ", " "]
+    all_delimiters = [x + " " for x in string.punctuation]
+
+    filename_simple = "test_simple_text.txt"
+    filename_complex = "test_complex_text.txt"
+
+    base_port = 6001
+    base_port = check_ports_free(base_port)
+
+    word_list_url = "https://raw.githubusercontent.com/dwyl/english-words/master/words_alpha.txt"
+    word_list = None
+    for i in range(4):
+        try:
+            url_content = requests.get(word_list_url, timeout=5)
+            word_list = url_content.text.splitlines()
+            break
+        except requests.exceptions.Timeout:
+            continue
+    if word_list is None:
+        sys.exit("Failed to get word list. Can not run tests without it. Quitting...")
+
+    book_base_url = "https://www.gutenberg.org/cache/epub/"
+    book_options = ["1513/pg1513.txt", "2701/pg2701.txt", "1342/pg1342.txt", "84/pg84.txt", "145/pg145.txt"]
+    book_1 = random.choice(book_options)
+    book_options.remove(book_1)
+    book_2 = random.choice(book_options)
+    book_text_1 = get_book_text(book_base_url + book_1)
+    book_list = [book_text_1]
+    book_text_2 = get_book_text(book_base_url + book_2)
+    book_list.append(book_text_2)
+
+    filename_book_1 = "book_1.txt"
+    filename_book_2 = "book_2.txt"
+
+    filename_interop = "interop_test.txt"
+
+    filename_valgrind = "valgrind_test.txt"
+
+
+    test_args = {"distributor": distributor_exec,
+                 "worker": worker_exec,
+                 # "cache": cache,
+                 # "cache_dir": cache_dir,
+                 "simple_delimiters": simple_delimiters,
+                 "all_delimiters": all_delimiters,
+                 "filename_simple": filename_simple,
+                 "filename_complex": filename_complex,
+                 "base_port": base_port,
+                 "word_list": word_list,
+                 "books": book_list,
+                 "filename_book_1": filename_book_1,
+                 "filename_book_2": filename_book_2,
+                 "filename_interop": filename_interop,
+                 "filename_valgrind": filename_valgrind,
+                 }
+
+    return test_args
+
+
+# checks if the base port and all following 8 ports are actually free for testing
+# if not, check the next 10
+def check_ports_free(base_port):
+    if not shutil.which("netstat") or sys.platform != "linux":
+        print("netstat not found. "
+              "To ensure all used ports during testing are free, please install net-tools with: sudo apt install net-tools",
+              file=sys.stderr)
+        return base_port
+
+    proc = subprocess.Popen(["netstat -tuna | tail -n +2 | tr -s [:blank:] ','"], stdout=subprocess.PIPE, shell=True)
+    netstat_output, netstat_err = proc.communicate()
+
+    netstat_output = netstat_output.decode("ascii")
+    df = pd.read_csv(StringIO(netstat_output))
+    df = df["Local"]
+    df = df.apply(lambda x: x.split(":")[-1])
+
+    ports = np.arange(base_port, base_port + 8).tolist()
+    port_list = [str(x) for x in ports]
+
+    if sum((df == i).any() for i in port_list) > 0:
+        base_port += 10
+        base_port = check_ports_free(base_port)
+    return base_port
+
+
+def kill_zmq_distributor_and_worker(kill_memcheck=False):
+    try:
+        distributor_procs = subprocess.check_output(["pidof", "zmq_distributor"]).decode("utf-8").split()
+        distributor_procs = [int(x) for x in distributor_procs]
+    except subprocess.CalledProcessError:
+        distributor_procs = []
+    try:
+        worker_procs = subprocess.check_output(["pidof", "zmq_worker"]).decode("utf-8").split()
+        worker_procs = [int(x) for x in worker_procs]
+    except subprocess.CalledProcessError:
+        worker_procs = []
+    if kill_memcheck:
+        # pgrep required, because pidof only returns exact matches
+        try:
+            memcheck_procs = subprocess.check_output(["pgrep", "memcheck"]).decode("utf-8").split()
+            memcheck_procs = [int(x) for x in memcheck_procs]
+        except subprocess.CalledProcessError:
+            memcheck_procs = []
+    else:
+        memcheck_procs = []
+    procs = distributor_procs + worker_procs + memcheck_procs
+    for p in procs:
+        try:
+            os.kill(p, signal.SIGKILL)
+        except ProcessLookupError:
+            continue
+
+
+def get_part_of_word_list(word_list, n_words):
+    word_list_part = random.sample(word_list, n_words)
+    return word_list_part
+
+
+def random_capitalization(words):
+    for i in range(len(words)):
+        words[i] = ''.join(map(random.choice, zip(words[i].lower(), words[i].upper())))
+
+
+def generate_text_from_word_list(w_list, delim, text_size):
+    word_lengths = [len(x) for x in w_list]
+    median_length = np.median(word_lengths)
+    words_to_pick = int(text_size / median_length)
+    words = random.choices(w_list, k=words_to_pick)
+    random_capitalization(words)
+
+    i = random.randint(25,100)
+    while i < len(words):
+        words.insert(i, "\n")
+        i += random.randint(25,100)
+
+    test_string = "".join(x + random.choice(delim) for x in words)
+
+    return test_string
+
+def count_words(input_string):
+    input_string = input_string.lower()
+
+    words = re.findall("[a-z]+", input_string)
+
+    words = Counter(words)
+    words = sorted(words.items(), key=lambda a: (-a[1], a[0]))
+
+    output_string = "word,frequency\n"
+
+    for w in words:
+        output_string += w[0] + "," + str(w[1]) + "\n"
+
+    return output_string
+
+def get_book_text(url):
+    book_text = None
+    for i in range(4):
+        try:
+            url_content = requests.get(url, timeout=5)
+            book_text = url_content.text.encode(encoding="ascii", errors="ignore")
+            break
+        except requests.exceptions.Timeout:
+            continue
+    if book_text is None:
+        sys.exit("Failed to get book text. Can not run tests without it. Quitting...")
+    return book_text
+
+
+def create_test_debug_output(test_name, num_w, expected_output, dist_output):
+    debug_base_path = "debug/"
+    if not os.path.isdir(debug_base_path):
+        os.mkdir(debug_base_path)
+    out_name = debug_base_path + test_name + "_n-workers=" + str(num_w) + "_"
+
+    f = open(out_name + "expected.txt", "w")
+    f.write(expected_output)
+    f.close()
+
+    f = open(out_name + "distributor_output.txt", "w")
+    f.write(dist_output)
+    f.close()
+
+
+def run_worker(w_exec : Union[str, List[str]], port : str):
+    proc_worker = None
+    if isinstance(w_exec, str):
+        proc_worker = subprocess.Popen([w_exec, port])
+    elif isinstance(w_exec, list):
+        w_exec.append(port)
+        proc_worker = subprocess.Popen(w_exec)
+    proc_worker.wait()
+
+
+def join_workers(workers: list):
+    for w in workers:
+        if isinstance(w, subprocess.Popen):
+            w.wait()
+        elif isinstance(w, multiprocessing.Process):
+            w.join()
+
+
+def start_threaded_workers(worker_exec : Union[str, List[List[str]]], port_list : list, use_multiprocessing = True):
+    # use_multiprocessing = False
+    if use_multiprocessing:
+        worker_procs = []
+        if isinstance(worker_exec, str):
+            for p in port_list:
+                proc = multiprocessing.Process(target=run_worker, args=(worker_exec, p))
+                worker_procs.append(proc)
+                proc.start()
+        elif isinstance(worker_exec, list):
+            for p, exe in zip(port_list, worker_exec):
+                proc = multiprocessing.Process(target=run_worker, args=(exe, p))
+                worker_procs.append(proc)
+                proc.start()
+        return worker_procs
+    else:
+        proc_workers = None
+        if isinstance(worker_exec, str):
+            proc_workers = subprocess.Popen([worker_exec] + port_list)
+        elif isinstance(worker_exec, list):
+            proc_workers = subprocess.Popen(worker_exec[0] + port_list)
+        return [proc_workers]
+
+
+def start_distributor(dist_args : List[str]):
+    return subprocess.Popen(dist_args, stdout=subprocess.PIPE, encoding="ascii")
+
+
+def run_worker_load_distribution(port, return_dict):
+    return_message = "test1worker1load1distribution1\0"
+
+    context = zmq.Context.instance()
+    socket = context.socket(zmq.REP)
+    socket.bind("tcp://*:" + str(port))
+
+    received_requests = 0
+
+    while True:
+        message = socket.recv()
+        message = message.decode("ascii")
+
+        msg_type = message[:3]
+        if msg_type == "map" or msg_type == "red":
+            received_requests += 1
+            socket.send(bytes(return_message, encoding="ascii"))
+        elif msg_type == "rip":
+            socket.send(bytes("rip\0", encoding="ascii"))
+            break
+        else:
+            print("Unknown message type!")
+            break
+    socket.close()
+
+    return_dict[port] = received_requests
+
+
+def check_valgrind_output_no_errors(filename):
+    if not os.path.isfile(filename):
+        return True
+    tree = ET.parse(filename)
+    errors = tree.findall("error")
+    return len(errors) == 0
+
+
+def cleanup_files(test_args, remove_valgrind_files=True):
+    for k, v in test_args.items():
+        if not remove_valgrind_files and "valgrind" in k and ".xml" in v:
+            continue
+        if "filename" in k and os.path.isfile(v):
+            os.remove(v)

praxis3/tox.ini

@@ -0,0 +1,10 @@
+[tox]
+env_list = py{38,39,310,311}
+skipsdist = true
+
+[testenv]
+description = run unit tests
+deps =
+    pytest
+commands =
+    pytest