diff options
Diffstat (limited to 'resources/tools/presentation/input_data_parser.py')
| -rw-r--r-- | resources/tools/presentation/input_data_parser.py | 205 |
1 files changed, 158 insertions, 47 deletions
diff --git a/resources/tools/presentation/input_data_parser.py b/resources/tools/presentation/input_data_parser.py index 7adc9c085b..beec34c106 100644 --- a/resources/tools/presentation/input_data_parser.py +++ b/resources/tools/presentation/input_data_parser.py @@ -18,15 +18,20 @@ - provide access to the data. """ +import multiprocessing +import os import re import pandas as pd import logging -import xml.etree.ElementTree as ET from robot.api import ExecutionResult, ResultVisitor from robot import errors from collections import OrderedDict from string import replace +from os import remove + +from input_data_files import download_and_unzip_data_file +from utils import Worker class ExecutionChecker(ResultVisitor): @@ -171,14 +176,14 @@ class ExecutionChecker(ResultVisitor): REGEX_TOLERANCE = re.compile(r'^[\D\d]*LOSS_ACCEPTANCE:\s(\d*\.\d*)\s' r'[\D\d]*') - REGEX_VERSION = re.compile(r"(stdout: 'vat# vat# Version:)(\s*)(.*)") + REGEX_VERSION = re.compile(r"(return STDOUT Version:\s*)(.*)") REGEX_TCP = re.compile(r'Total\s(rps|cps|throughput):\s([0-9]*).*$') REGEX_MRR = re.compile(r'MaxReceivedRate_Results\s\[pkts/(\d*)sec\]:\s' r'tx\s(\d*),\srx\s(\d*)') - def __init__(self, **metadata): + def __init__(self, metadata): """Initialisation. :param metadata: Key-value pairs to be included in "metadata" part of @@ -244,14 +249,13 @@ class ExecutionChecker(ResultVisitor): :returns: Nothing. """ - if msg.message.count("stdout: 'vat# vat# Version:"): + if msg.message.count("return STDOUT Version:"): self._version = str(re.search(self.REGEX_VERSION, msg.message). - group(3)) + group(2)) self._data["metadata"]["version"] = self._version + self._data["metadata"]["generated"] = msg.timestamp self._msg_type = None - logging.debug(" VPP version: {0}".format(self._version)) - def _get_vat_history(self, msg): """Called when extraction of VAT command history is required. @@ -585,7 +589,7 @@ class ExecutionChecker(ResultVisitor): :type setup_kw: Keyword :returns: Nothing. """ - if setup_kw.name.count("Vpp Show Version Verbose") \ + if setup_kw.name.count("Show Vpp Version On All Duts") \ and not self._version: self._msg_type = "setup-version" setup_kw.messages.visit(self) @@ -696,7 +700,7 @@ class InputData(object): self._cfg = spec # Data store: - self._input_data = None + self._input_data = pd.Series() @property def data(self): @@ -747,76 +751,186 @@ class InputData(object): return self.data[job][build]["tests"] @staticmethod - def _parse_tests(job, build): + def _parse_tests(job, build, log): """Process data from robot output.xml file and return JSON structured data. :param job: The name of job which build output data will be processed. :param build: The build which output data will be processed. + :param log: List of log messages. :type job: str :type build: dict + :type log: list of tuples (severity, msg) :returns: JSON data structure. :rtype: dict """ - tree = ET.parse(build["file-name"]) - root = tree.getroot() - generated = root.attrib["generated"] + metadata = { + "job": job, + "build": build + } with open(build["file-name"], 'r') as data_file: try: result = ExecutionResult(data_file) except errors.DataError as err: - logging.error("Error occurred while parsing output.xml: {0}". - format(err)) + log.append(("ERROR", "Error occurred while parsing output.xml: " + "{0}".format(err))) return None - checker = ExecutionChecker(job=job, build=build, generated=generated) + checker = ExecutionChecker(metadata) result.visit(checker) return checker.data - def read_data(self): - """Parse input data from input files and store in pandas' Series. + def _download_and_parse_build(self, pid, data_queue, job, build, repeat): + """Download and parse the input data file. + + :param pid: PID of the process executing this method. + :param data_queue: Shared memory between processes. Queue which keeps + the result data. This data is then read by the main process and used + in further processing. + :param job: Name of the Jenkins job which generated the processed input + file. + :param build: Information about the Jenkins build which generated the + processed input file. + :param repeat: Repeat the download specified number of times if not + successful. + :type pid: int + :type data_queue: multiprocessing.Manager().Queue() + :type job: str + :type build: dict + :type repeat: int + """ + + logs = list() + + logging.info(" Processing the job/build: {0}: {1}". + format(job, build["build"])) + + logs.append(("INFO", " Processing the job/build: {0}: {1}". + format(job, build["build"]))) + + state = "failed" + success = False + data = None + do_repeat = repeat + while do_repeat: + success = download_and_unzip_data_file(self._cfg, job, build, pid, + logs) + if success: + break + do_repeat -= 1 + if not success: + logs.append(("ERROR", "It is not possible to download the input " + "data file from the job '{job}', build " + "'{build}', or it is damaged. Skipped.". + format(job=job, build=build["build"]))) + if success: + logs.append(("INFO", " Processing data from the build '{0}' ...". + format(build["build"]))) + data = InputData._parse_tests(job, build, logs) + if data is None: + logs.append(("ERROR", "Input data file from the job '{job}', " + "build '{build}' is damaged. Skipped.". + format(job=job, build=build["build"]))) + else: + state = "processed" + + try: + remove(build["file-name"]) + except OSError as err: + logs.append(("ERROR", "Cannot remove the file '{0}': {1}". + format(build["file-name"], err))) + logs.append(("INFO", " Done.")) + + result = { + "data": data, + "state": state, + "job": job, + "build": build, + "logs": logs + } + data_queue.put(result) + + def download_and_parse_data(self, repeat=1): + """Download the input data files, parse input data from input files and + store in pandas' Series. + + :param repeat: Repeat the download specified number of times if not + successful. + :type repeat: int """ - logging.info("Parsing input files ...") + logging.info("Downloading and parsing input files ...") + + work_queue = multiprocessing.JoinableQueue() + manager = multiprocessing.Manager() + data_queue = manager.Queue() + cpus = multiprocessing.cpu_count() + + workers = list() + for cpu in range(cpus): + worker = Worker(work_queue, + data_queue, + self._download_and_parse_build) + worker.daemon = True + worker.start() + workers.append(worker) + os.system("taskset -p -c {0} {1} > /dev/null 2>&1". + format(cpu, worker.pid)) - job_data = dict() for job, builds in self._cfg.builds.items(): - logging.info(" Extracting data from the job '{0}' ...'". - format(job)) - builds_data = dict() for build in builds: - if build["status"] == "failed" \ - or build["status"] == "not found": - continue - logging.info(" Extracting data from the build '{0}'". - format(build["build"])) - logging.info(" Processing the file '{0}'". - format(build["file-name"])) - data = InputData._parse_tests(job, build) - if data is None: - logging.error("Input data file from the job '{job}', build " - "'{build}' is damaged. Skipped.". - format(job=job, build=build["build"])) - continue + work_queue.put((job, build, repeat)) + + work_queue.join() + + logging.info("Done.") + + while not data_queue.empty(): + result = data_queue.get() + + job = result["job"] + build_nr = result["build"]["build"] + if result["data"]: + data = result["data"] build_data = pd.Series({ "metadata": pd.Series(data["metadata"].values(), index=data["metadata"].keys()), "suites": pd.Series(data["suites"].values(), index=data["suites"].keys()), "tests": pd.Series(data["tests"].values(), - index=data["tests"].keys()), - }) - builds_data[str(build["build"])] = build_data - logging.info(" Done.") + index=data["tests"].keys())}) - job_data[job] = pd.Series(builds_data.values(), - index=builds_data.keys()) - logging.info(" Done.") + if self._input_data.get(job, None) is None: + self._input_data[job] = pd.Series() + self._input_data[job][str(build_nr)] = build_data + + self._cfg.set_input_file_name(job, build_nr, + result["build"]["file-name"]) + + self._cfg.set_input_state(job, build_nr, result["state"]) + + for item in result["logs"]: + if item[0] == "INFO": + logging.info(item[1]) + elif item[0] == "ERROR": + logging.error(item[1]) + elif item[0] == "DEBUG": + logging.debug(item[1]) + elif item[0] == "CRITICAL": + logging.critical(item[1]) + elif item[0] == "WARNING": + logging.warning(item[1]) + + del data_queue + + # Terminate all workers + for worker in workers: + worker.terminate() + worker.join() - self._input_data = pd.Series(job_data.values(), index=job_data.keys()) logging.info("Done.") @staticmethod @@ -893,9 +1007,6 @@ class InputData(object): :rtype pandas.Series """ - logging.info(" Creating the data set for the {0} '{1}'.". - format(element.get("type", ""), element.get("title", ""))) - try: if element["filter"] in ("all", "template"): cond = "True" |