ClickHouse/tests/ci/functional_test_check.py

#!/usr/bin/env python3
#
from report import create_test_html_report
from s3_helper import S3Helper
import requests
from github import Github
from get_robot_token import get_best_robot_token, get_parameter_from_ssm
import csv
from pr_info import PRInfo
import logging
import subprocess
import os
import json
import time
from pr_info import PRInfo
import sys

DOWNLOAD_RETRIES_COUNT = 5

def process_results(result_folder):
    test_results = []
    additional_files = []
    # Just upload all files from result_folder.
    # If task provides processed results, then it's responsible for content of result_folder.
    if os.path.exists(result_folder):
        test_files = [f for f in os.listdir(result_folder) if os.path.isfile(os.path.join(result_folder, f))]
        additional_files = [os.path.join(result_folder, f) for f in test_files]

    status_path = os.path.join(result_folder, "check_status.tsv")
    logging.info("Found test_results.tsv")
    status = list(csv.reader(open(status_path, 'r'), delimiter='\t'))
    if len(status) != 1 or len(status[0]) != 2:
        return "error", "Invalid check_status.tsv", test_results, additional_files
    state, description = status[0][0], status[0][1]

    results_path = os.path.join(result_folder, "test_results.tsv")
    test_results = list(csv.reader(open(results_path, 'r'), delimiter='\t'))
    if len(test_results) == 0:
        raise Exception("Empty results")

    return state, description, test_results, additional_files


def process_logs(s3_client, additional_logs, s3_path_prefix):
    additional_urls = []
    for log_path in additional_logs:
        if log_path:
            additional_urls.append(
                s3_client.upload_test_report_to_s3(
                    log_path,
                    s3_path_prefix + "/" + os.path.basename(log_path)))

    return additional_urls


def upload_results(s3_client, pr_number, commit_sha, test_results, raw_log, additional_files, check_name):
    additional_files = [raw_log] + additional_files
    s3_path_prefix = f"{pr_number}/{commit_sha}/fasttest"
    additional_urls = process_logs(s3_client, additional_files, s3_path_prefix)

    branch_url = "https://github.com/ClickHouse/ClickHouse/commits/master"
    branch_name = "master"
    if pr_number != 0:
        branch_name = "PR #{}".format(pr_number)
        branch_url = "https://github.com/ClickHouse/ClickHouse/pull/" + str(pr_number)
    commit_url = f"https://github.com/ClickHouse/ClickHouse/commit/{commit_sha}"

    task_url = f"https://github.com/ClickHouse/ClickHouse/actions/runs/{os.getenv('GITHUB_RUN_ID')}"

    raw_log_url = additional_urls[0]
    additional_urls.pop(0)

    html_report = create_test_html_report(check_name, test_results, raw_log_url, task_url, branch_url, branch_name, commit_url, additional_urls, True)
    with open('report.html', 'w') as f:
        f.write(html_report)

    url = s3_client.upload_test_report_to_s3('report.html', s3_path_prefix + ".html")
    logging.info("Search result in url %s", url)
    return url

def get_commit(gh, commit_sha):
    repo = gh.get_repo(os.getenv("GITHUB_REPOSITORY", "ClickHouse/ClickHouse"))
    commit = repo.get_commit(commit_sha)
    return commit

def get_image_name(check_name):
    if 'stateless' in check_name.lower():
        return 'clickhouse/stateless-test'
    if 'stateful' in  check_name.lower():
        return 'clickhouse/stateful-test'


def dowload_build_with_progress(url, path):
    logging.info("Downloading from %s to temp path %s", url, path)
    for i in range(DOWNLOAD_RETRIES_COUNT):
        try:
            with open(path, 'wb') as f:
                response = requests.get(url, stream=True)
                response.raise_for_status()
                total_length = response.headers.get('content-length')
                if total_length is None or int(total_length) == 0:
                    logging.info("No content-length, will download file without progress")
                    f.write(response.content)
                else:
                    dl = 0
                    total_length = int(total_length)
                    logging.info("Content length is %ld bytes", total_length)
                    for data in response.iter_content(chunk_size=4096):
                        dl += len(data)
                        f.write(data)
                        if sys.stdout.isatty():
                            done = int(50 * dl / total_length)
                            percent = int(100 * float(dl) / total_length)
                            sys.stdout.write("\r[{}{}] {}%".format('=' * done, ' ' * (50-done), percent))
                            sys.stdout.flush()
            break
        except Exception as ex:
            sys.stdout.write("\n")
            time.sleep(3)
            logging.info("Exception while downloading %s, retry %s", ex, i + 1)
            if os.path.exists(path):
                os.remove(path)
    else:
        raise Exception("Cannot download dataset from {}, all retries exceeded".format(url))

    sys.stdout.write("\n")
    logging.info("Downloading finished")


def download_builds(result_path, build_urls):
    for url in build_urls:
        if url.endswith('.deb'):
            fname = os.path.basename(url)
            logging.info("Will download %s to %s", fname, result_path))
            download_build_with_progress(url, os.path.join(result_path, fname))


def get_build_config(build_number, repo_path):
    ci_config_path = os.path.join(repo_path, "tests/ci/ci_config.json")
    with open(ci_config_path, 'r') as ci_config:
        config_dict = json.load(ci_config)
        return config_dict['build_config'][build_number]

def get_build_urls(build_config_str, reports_path):
    for root, dirs, files in os.walk(reports_path):
        for f in files:
            if build_config_str in f :
                logging.info("Found build report json %s", f)
                with open(os.path.join(root, f), 'r') as file_handler:
                    build_report = json.load(file_handler)
                    return build_report['build_urls']
    return []

def build_config_to_string(build_config):
    if build_config["package-type"] == "performance":
        return "performance"

    return "_".join([
        build_config['compiler'],
        build_config['build-type'] if build_config['build-type'] else "relwithdebuginfo",
        build_config['sanitizer'] if build_config['sanitizer'] else "none",
        build_config['bundled'],
        build_config['splitted'],
        "tidy" if build_config['tidy'] == "enable" else "notidy",
        "with_coverage" if build_config['with_coverage'] else "without_coverage",
        build_config['package-type'],
    ])

def get_run_command(builds_path, result_path, server_log_path, kill_timeout, additional_envs, image):
    additional_options = ['--hung-check']
    additional_options.append('--print-time')
    additional_options_str = '-e ADDITIONAL_OPTIONS="' + ' '.join(additional_options) + '"'

    envs = ['-e MAX_RUN_TIME={int(0.9 * kill_timeout)}']
    envs += [f'-e {e}' for e in additional_envs]
    env_str = ' '.join(envs)

    return f"docker run --volume={builds_path}:/package_folder " \
        f"--volume={result_path}:/test_output --volume={server_log_path}:/var/log/clickhouse-server " \
        f"--cap-add=SYS_PTRACE {env_str} {additional_options} {image}"


def process_results(result_folder, server_log_path):
    test_results = []
    additional_files = []
    # Just upload all files from result_folder.
    # If task provides processed results, then it's responsible for content of result_folder.
    if os.path.exists(result_folder):
        test_files = [f for f in os.listdir(result_folder) if os.path.isfile(os.path.join(result_folder, f))]
        additional_files = [os.path.join(result_folder, f) for f in test_files]

    if os.path.exists(server_log_path):
        server_log_files = [f for f in os.listdir(server_log_path) if os.path.isfile(os.path.join(server_log_path, f))]
        additional_files = additional_files + [os.path.join(server_log_path, f) for f in server_log_files]

    status_path = os.path.join(result_folder, "check_status.tsv")
    logging.info("Found test_results.tsv")
    status = list(csv.reader(open(status_path, 'r'), delimiter='\t'))
    if len(status) != 1 or len(status[0]) != 2:
        return "error", "Invalid check_status.tsv", test_results, additional_files
    state, description = status[0][0], status[0][1]

    results_path = os.path.join(result_folder, "test_results.tsv")
    test_results = list(csv.reader(open(results_path, 'r'), delimiter='\t'))
    if len(test_results) == 0:
        raise Exception("Empty results")

    return state, description, test_results, additional_files


def process_logs(s3_client, additional_logs, s3_path_prefix):
    additional_urls = []
    for log_path in additional_logs:
        if log_path:
            additional_urls.append(
                s3_client.upload_test_report_to_s3(
                    log_path,
                    s3_path_prefix + "/" + os.path.basename(log_path)))

    return additional_urls


def upload_results(s3_client, pr_number, commit_sha, test_results, raw_log, additional_files):
    additional_files = [raw_log] + additional_files
    s3_path_prefix = f"{pr_number}/{commit_sha}/fasttest"
    additional_urls = process_logs(s3_client, additional_files, s3_path_prefix)

    branch_url = "https://github.com/ClickHouse/ClickHouse/commits/master"
    branch_name = "master"
    if pr_number != 0:
        branch_name = "PR #{}".format(pr_number)
        branch_url = "https://github.com/ClickHouse/ClickHouse/pull/" + str(pr_number)
    commit_url = f"https://github.com/ClickHouse/ClickHouse/commit/{commit_sha}"

    task_url = f"https://github.com/ClickHouse/ClickHouse/actions/runs/{os.getenv('GITHUB_RUN_ID')}"

    raw_log_url = additional_urls[0]
    additional_urls.pop(0)

    html_report = create_test_html_report(NAME, test_results, raw_log_url, task_url, branch_url, branch_name, commit_url, additional_urls, True)
    with open('report.html', 'w') as f:
        f.write(html_report)

    url = s3_client.upload_test_report_to_s3('report.html', s3_path_prefix + ".html")
    logging.info("Search result in url %s", url)
    return url

def get_commit(gh, commit_sha):
    repo = gh.get_repo(os.getenv("GITHUB_REPOSITORY", "ClickHouse/ClickHouse"))
    commit = repo.get_commit(commit_sha)
    return commit


if __name__ == "__main__":
    logging.basicConfig(level=logging.INFO)
    temp_path = os.getenv("TEMP_PATH", os.path.abspath("."))
    repo_path = os.getenv("REPO_COPY", os.path.abspath("../../"))
    reports_path = os.getenv("REPORTS_PATH", "./reports")

    check_name = sys.argv[1]
    build_number = int(sys.argv[2])
    kill_timeout = int(sys.argv[3])

    if not os.path.exists(temp_path):
        os.makedirs(temp_path)

    with open(os.getenv('GITHUB_EVENT_PATH'), 'r') as event_file:
        event = json.load(event_file)

    pr_info = PRInfo(event)

    gh = Github(get_best_robot_token())

    images_path = os.path.join(temp_path, 'changed_images.json')
    image_name = get_image_name(check_name)

    docker_image = image_name
    if os.path.exists(images_path):
        logging.info("Images file exists")
        with open(images_path, 'r') as images_fd:
            images = json.load(images_fd)
            logging.info("Got images %s", images)
            if image_name in images:
                docker_image += ':' + images[image_name]

    for i in range(10):
        try:
            logging.info(f"Pulling image {docker_image}")
            subprocess.check_output(f"docker pull {docker_image}", stderr=subprocess.STDOUT, shell=True)
            break
        except Exception as ex:
            time.sleep(i * 3)
            logging.info("Got execption pulling docker %s", ex)
    else:
        raise Exception(f"Cannot pull dockerhub for image docker pull {docker_image}")

    build_config = get_build_config(build_number, repo_path)
    build_config_str = build_config_to_string(build_config)
    urls = get_build_urls(build_config_str, reports_path)
    if not urls:
        raise Exception("No build URLs found")

    packages_path = os.path.join(temp_path, "packages")
    if not os.path.exists(packages_path):
        os.makedirs(packages_path)

    server_log_path = os.path.join(temp_path, "server_log")
    if not os.path.exists(server_log_path):
        os.makedirs(server_log_path)

    result_path = os.path.join(temp_path, "result_path")
    if not os.path.exists(result_path):
        os.makedirs(result_path)

    run_log_path = os.path.join(result_path, "runlog.log")
    if not os.path.exists(run_log_path):
        os.makedirs(run_log_path)

    download_builds(packages_path, urls)
    run_command = get_run_command(packages_path, result_path, server_log_path, kill_timeout, [], docker_image)
    logging.info("Going to run func tests: %s", run_command)

    with open(run_log_path, 'w') as log:
        retcode = subprocess.Popen(run_cmd, shell=True, stderr=log, stdout=log).wait()
        if retcode == 0:
            logging.info("Run successfully")
        else:
            logging.info("Run failed")

    subprocess.check_call(f"sudo chown -R ubuntu:ubuntu {temp_path}", shell=True)

    s3_helper = S3Helper('https://s3.amazonaws.com')
    state, description, test_results, additional_logs = process_results(output_path, server_log_path)
    report_url = upload_results(s3_helper, pr_info.number, pr_info.sha, test_results, run_log_path, additional_logs)
    print("::notice ::Report url: {}".format(report_url))
    commit = get_commit(gh, pr_info.sha)
    commit.create_status(context=check_name, description=description, state=state, target_url=report_url)