diff options
Diffstat (limited to 'resources')
-rw-r--r-- | resources/tools/presentation/input_data_parser.py | 1167 |
1 files changed, 430 insertions, 737 deletions
diff --git a/resources/tools/presentation/input_data_parser.py b/resources/tools/presentation/input_data_parser.py index 869929adf2..1f175bc3d8 100644 --- a/resources/tools/presentation/input_data_parser.py +++ b/resources/tools/presentation/input_data_parser.py @@ -46,164 +46,11 @@ from pal_errors import PresentationError # Separator used in file names -SEPARATOR = u"__" +SEPARATOR = "__" class ExecutionChecker(ResultVisitor): """Class to traverse through the test suite structure. - - The functionality implemented in this class generates a json structure: - - Performance tests: - - { - "metadata": { - "generated": "Timestamp", - "version": "SUT version", - "job": "Jenkins job name", - "build": "Information about the build" - }, - "suites": { - "Suite long name 1": { - "name": Suite name, - "doc": "Suite 1 documentation", - "parent": "Suite 1 parent", - "level": "Level of the suite in the suite hierarchy" - } - "Suite long name N": { - "name": Suite name, - "doc": "Suite N documentation", - "parent": "Suite 2 parent", - "level": "Level of the suite in the suite hierarchy" - } - } - "tests": { - # NDRPDR tests: - "ID": { - "name": "Test name", - "parent": "Name of the parent of the test", - "doc": "Test documentation", - "msg": "Test message", - "conf-history": "DUT1 and DUT2 VAT History", - "show-run": "Show Run", - "tags": ["tag 1", "tag 2", "tag n"], - "type": "NDRPDR", - "status": "PASS" | "FAIL", - "throughput": { - "NDR": { - "LOWER": float, - "UPPER": float - }, - "PDR": { - "LOWER": float, - "UPPER": float - } - }, - "latency": { - "NDR": { - "direction1": { - "min": float, - "avg": float, - "max": float, - "hdrh": str - }, - "direction2": { - "min": float, - "avg": float, - "max": float, - "hdrh": str - } - }, - "PDR": { - "direction1": { - "min": float, - "avg": float, - "max": float, - "hdrh": str - }, - "direction2": { - "min": float, - "avg": float, - "max": float, - "hdrh": str - } - } - } - } - - # TCP tests: - "ID": { - "name": "Test name", - "parent": "Name of the parent of the test", - "doc": "Test documentation", - "msg": "Test message", - "tags": ["tag 1", "tag 2", "tag n"], - "type": "TCP", - "status": "PASS" | "FAIL", - "result": int - } - - # MRR, BMRR tests: - "ID": { - "name": "Test name", - "parent": "Name of the parent of the test", - "doc": "Test documentation", - "msg": "Test message", - "tags": ["tag 1", "tag 2", "tag n"], - "type": "MRR" | "BMRR", - "status": "PASS" | "FAIL", - "result": { - "receive-rate": float, - # Average of a list, computed using AvgStdevStats. - # In CSIT-1180, replace with List[float]. - } - } - - "ID" { - # next test - } - } - } - - - Functional tests: - - { - "metadata": { # Optional - "version": "VPP version", - "job": "Jenkins job name", - "build": "Information about the build" - }, - "suites": { - "Suite name 1": { - "doc": "Suite 1 documentation", - "parent": "Suite 1 parent", - "level": "Level of the suite in the suite hierarchy" - } - "Suite name N": { - "doc": "Suite N documentation", - "parent": "Suite 2 parent", - "level": "Level of the suite in the suite hierarchy" - } - } - "tests": { - "ID": { - "name": "Test name", - "parent": "Name of the parent of the test", - "doc": "Test documentation" - "msg": "Test message" - "tags": ["tag 1", "tag 2", "tag n"], - "conf-history": "DUT1 and DUT2 VAT History" - "show-run": "Show Run" - "status": "PASS" | "FAIL" - }, - "ID" { - # next test - } - } - } - - .. note:: ID is the lowercase full path to the test. """ REGEX_PLR_RATE = re.compile( @@ -312,8 +159,21 @@ class ExecutionChecker(ResultVisitor): :type process_oper: bool """ - # Type of message to parse out from the test messages - self._msg_type = None + # Save the provided metadata + for key, val in metadata.items(): + self._data["metadata"][key] = val + + # Mapping of TCs long names + self._mapping = mapping + + # Ignore list + self._ignore = ignore + + # Process operational data + self._process_oper = process_oper + + # Name of currently processed keyword + self._kw_name = None # VPP version self._version = None @@ -324,14 +184,6 @@ class ExecutionChecker(ResultVisitor): # Testbed. The testbed is identified by TG node IP address. self._testbed = None - # Mapping of TCs long names - self._mapping = mapping - - # Ignore list - self._ignore = ignore - - self._process_oper = process_oper - # Number of PAPI History messages found: # 0 - no message # 1 - PAPI History of DUT1 @@ -348,24 +200,9 @@ class ExecutionChecker(ResultVisitor): # The main data structure self._data = { - u"metadata": OrderedDict(), - u"suites": OrderedDict(), - u"tests": OrderedDict() - } - - # Save the provided metadata - for key, val in metadata.items(): - self._data[u"metadata"][key] = val - - # Dictionary defining the methods used to parse different types of - # messages - self.parse_msg = { - u"vpp-version": self._get_vpp_version, - u"dpdk-version": self._get_dpdk_version, - u"teardown-papi-history": self._get_papi_history, - u"test-show-runtime": self._get_show_run, - u"testbed": self._get_testbed, - u"test-telemetry": self._get_telemetry + "metadata": dict(), + "suites": dict(), + "tests": dict() } @property @@ -388,20 +225,20 @@ class ExecutionChecker(ResultVisitor): groups = re.search(self.REGEX_MRR_MSG_INFO, msg) if not groups or groups.lastindex != 1: - return u"Test Failed." + return "Test Failed." try: - data = groups.group(1).split(u", ") + data = groups.group(1).split(", ") except (AttributeError, IndexError, ValueError, KeyError): - return u"Test Failed." + return "Test Failed." - out_str = u"[" + out_str = "[" try: for item in data: out_str += f"{(float(item) / 1e6):.2f}, " - return out_str[:-2] + u"]" + return out_str[:-2] + "]" except (AttributeError, IndexError, ValueError, KeyError): - return u"Test Failed." + return "Test Failed." def _get_data_from_cps_test_msg(self, msg): """Get info from message of NDRPDR CPS tests. @@ -414,7 +251,7 @@ class ExecutionChecker(ResultVisitor): groups = re.search(self.REGEX_CPS_MSG_INFO, msg) if not groups or groups.lastindex != 2: - return u"Test Failed." + return "Test Failed." try: return ( @@ -422,7 +259,7 @@ class ExecutionChecker(ResultVisitor): f"2. {(float(groups.group(2)) / 1e6):5.2f}" ) except (AttributeError, IndexError, ValueError, KeyError): - return u"Test Failed." + return "Test Failed." def _get_data_from_pps_test_msg(self, msg): """Get info from message of NDRPDR PPS tests. @@ -435,7 +272,7 @@ class ExecutionChecker(ResultVisitor): groups = re.search(self.REGEX_PPS_MSG_INFO, msg) if not groups or groups.lastindex != 4: - return u"Test Failed." + return "Test Failed." try: return ( @@ -445,7 +282,7 @@ class ExecutionChecker(ResultVisitor): f"{float(groups.group(4)):5.2f}" ) except (AttributeError, IndexError, ValueError, KeyError): - return u"Test Failed." + return "Test Failed." def _get_data_from_perf_test_msg(self, msg): """Get info from message of NDRPDR performance tests. @@ -458,23 +295,23 @@ class ExecutionChecker(ResultVisitor): groups = re.search(self.REGEX_PERF_MSG_INFO, msg) if not groups or groups.lastindex != 10: - return u"Test Failed." + return "Test Failed." try: data = { - u"ndr_low": float(groups.group(1)), - u"ndr_low_b": float(groups.group(2)), - u"pdr_low": float(groups.group(3)), - u"pdr_low_b": float(groups.group(4)), - u"pdr_lat_90_1": groups.group(5), - u"pdr_lat_90_2": groups.group(6), - u"pdr_lat_50_1": groups.group(7), - u"pdr_lat_50_2": groups.group(8), - u"pdr_lat_10_1": groups.group(9), - u"pdr_lat_10_2": groups.group(10), + "ndr_low": float(groups.group(1)), + "ndr_low_b": float(groups.group(2)), + "pdr_low": float(groups.group(3)), + "pdr_low_b": float(groups.group(4)), + "pdr_lat_90_1": groups.group(5), + "pdr_lat_90_2": groups.group(6), + "pdr_lat_50_1": groups.group(7), + "pdr_lat_50_2": groups.group(8), + "pdr_lat_10_1": groups.group(9), + "pdr_lat_10_2": groups.group(10), } except (AttributeError, IndexError, ValueError, KeyError): - return u"Test Failed." + return "Test Failed." def _process_lat(in_str_1, in_str_2): """Extract P50, P90 and P99 latencies or min, avg, max values from @@ -495,13 +332,13 @@ class ExecutionChecker(ResultVisitor): if len(in_list_1) != 4 and len(in_list_2) != 4: return None - in_list_1[3] += u"=" * (len(in_list_1[3]) % 4) + in_list_1[3] += "=" * (len(in_list_1[3]) % 4) try: hdr_lat_1 = hdrh.histogram.HdrHistogram.decode(in_list_1[3]) except hdrh.codec.HdrLengthException: hdr_lat_1 = None - in_list_2[3] += u"=" * (len(in_list_2[3]) % 4) + in_list_2[3] += "=" * (len(in_list_2[3]) % 4) try: hdr_lat_2 = hdrh.histogram.HdrHistogram.decode(in_list_2[3]) except hdrh.codec.HdrLengthException: @@ -530,15 +367,15 @@ class ExecutionChecker(ResultVisitor): try: out_msg = ( - f"1. {(data[u'ndr_low'] / 1e6):5.2f} " - f"{data[u'ndr_low_b']:5.2f}" - f"\n2. {(data[u'pdr_low'] / 1e6):5.2f} " - f"{data[u'pdr_low_b']:5.2f}" + f"1. {(data['ndr_low'] / 1e6):5.2f} " + f"{data['ndr_low_b']:5.2f}" + f"\n2. {(data['pdr_low'] / 1e6):5.2f} " + f"{data['pdr_low_b']:5.2f}" ) latency = ( - _process_lat(data[u'pdr_lat_10_1'], data[u'pdr_lat_10_2']), - _process_lat(data[u'pdr_lat_50_1'], data[u'pdr_lat_50_2']), - _process_lat(data[u'pdr_lat_90_1'], data[u'pdr_lat_90_2']) + _process_lat(data['pdr_lat_10_1'], data['pdr_lat_10_2']), + _process_lat(data['pdr_lat_50_1'], data['pdr_lat_50_2']), + _process_lat(data['pdr_lat_90_1'], data['pdr_lat_90_2']) ) if all(latency): max_len = len(str(max((max(item) for item in latency)))) @@ -546,7 +383,7 @@ class ExecutionChecker(ResultVisitor): for idx, lat in enumerate(latency): if not idx: - out_msg += u"\n" + out_msg += "\n" out_msg += ( f"\n{idx + 3}. " f"{lat[0]:{max_len}d} " @@ -560,7 +397,7 @@ class ExecutionChecker(ResultVisitor): return out_msg except (AttributeError, IndexError, ValueError, KeyError): - return u"Test Failed." + return "Test Failed." def _get_testbed(self, msg): """Called when extraction of testbed IP is required. @@ -571,8 +408,8 @@ class ExecutionChecker(ResultVisitor): :returns: Nothing. """ - if msg.message.count(u"Setup of TG node") or \ - msg.message.count(u"Setup of node TG host"): + if msg.message.count("Setup of TG node") or \ + msg.message.count("Setup of node TG host"): reg_tg_ip = re.compile( r'.*TG .* (\d{1,3}.\d{1,3}.\d{1,3}.\d{1,3}).*') try: @@ -580,8 +417,7 @@ class ExecutionChecker(ResultVisitor): except (KeyError, ValueError, IndexError, AttributeError): pass finally: - self._data[u"metadata"][u"testbed"] = self._testbed - self._msg_type = None + self._data["metadata"]["testbed"] = self._testbed def _get_vpp_version(self, msg): """Called when extraction of VPP version is required. @@ -591,14 +427,12 @@ class ExecutionChecker(ResultVisitor): :returns: Nothing. """ - if msg.message.count(u"VPP version:") or \ - msg.message.count(u"VPP Version:"): + if msg.message.count("VPP version:") or \ + msg.message.count("VPP Version:"): self._version = str( re.search(self.REGEX_VERSION_VPP, msg.message).group(2) ) - self._data[u"metadata"][u"version"] = self._version - self._msg_type = None - logging.info(self._version) + self._data["metadata"]["version"] = self._version def _get_dpdk_version(self, msg): """Called when extraction of DPDK version is required. @@ -608,15 +442,13 @@ class ExecutionChecker(ResultVisitor): :returns: Nothing. """ - if msg.message.count(u"DPDK Version:"): + if msg.message.count("DPDK Version:"): try: self._version = str(re.search( self.REGEX_VERSION_DPDK, msg.message).group(2)) - self._data[u"metadata"][u"version"] = self._version + self._data["metadata"]["version"] = self._version except IndexError: pass - finally: - self._msg_type = None def _get_papi_history(self, msg): """Called when extraction of PAPI command history is required. @@ -625,21 +457,18 @@ class ExecutionChecker(ResultVisitor): :type msg: Message :returns: Nothing. """ - if msg.message.count(u"PAPI command history:"): + if msg.message.count("PAPI command history:"): self._conf_history_lookup_nr += 1 if self._conf_history_lookup_nr == 1: - self._data[u"tests"][self._test_id][u"conf-history"] = str() - else: - self._msg_type = None + self._data["tests"][self._test_id]["conf-history"] = str() text = re.sub( r"\d{1,3}.\d{1,3}.\d{1,3}.\d{1,3} PAPI command history:", - u"", + "", msg.message, count=1 - ).replace(u'"', u"'") - self._data[u"tests"][self._test_id][u"conf-history"] += ( + ).replace('"', "'") + self._data["tests"][self._test_id]["conf-history"] += \ f"**DUT{str(self._conf_history_lookup_nr)}:** {text}" - ) def _get_show_run(self, msg): """Called when extraction of VPP operational data (output of CLI command @@ -650,15 +479,15 @@ class ExecutionChecker(ResultVisitor): :returns: Nothing. """ - if not msg.message.count(u"stats runtime"): + if not msg.message.count("stats runtime"): return # Temporary solution if self._sh_run_counter > 1: return - if u"show-run" not in self._data[u"tests"][self._test_id].keys(): - self._data[u"tests"][self._test_id][u"show-run"] = dict() + if "show-run" not in self._data["tests"][self._test_id].keys(): + self._data["tests"][self._test_id]["show-run"] = dict() groups = re.search(self.REGEX_TC_PAPI_CLI, msg.message) if not groups: @@ -666,24 +495,24 @@ class ExecutionChecker(ResultVisitor): try: host = groups.group(1) except (AttributeError, IndexError): - host = u"" + host = "" try: sock = groups.group(2) except (AttributeError, IndexError): - sock = u"" + sock = "" - dut = u"dut{nr}".format( - nr=len(self._data[u'tests'][self._test_id][u'show-run'].keys()) + 1) + dut = "dut{nr}".format( + nr=len(self._data['tests'][self._test_id]['show-run'].keys()) + 1) - self._data[u'tests'][self._test_id][u'show-run'][dut] = \ + self._data['tests'][self._test_id]['show-run'][dut] = \ copy.copy( { - u"host": host, - u"socket": sock, - u"runtime": str(msg.message).replace(u' ', u''). - replace(u'\n', u'').replace(u"'", u'"'). - replace(u'b"', u'"').replace(u'u"', u'"'). - split(u":", 1)[1] + "host": host, + "socket": sock, + "runtime": str(msg.message).replace(' ', ''). + replace('\n', '').replace("'", '"'). + replace('b"', '"').replace('"', '"'). + split(":", 1)[1] } ) @@ -697,12 +526,12 @@ class ExecutionChecker(ResultVisitor): if self._telemetry_kw_counter > 1: return - if not msg.message.count(u"# TYPE vpp_runtime_calls"): + if not msg.message.count("# TYPE vpp_runtime_calls"): return - if u"telemetry-show-run" not in \ - self._data[u"tests"][self._test_id].keys(): - self._data[u"tests"][self._test_id][u"telemetry-show-run"] = dict() + if "telemetry-show-run" not in \ + self._data["tests"][self._test_id].keys(): + self._data["tests"][self._test_id]["telemetry-show-run"] = dict() self._telemetry_msg_counter += 1 groups = re.search(self.REGEX_SH_RUN_HOST, msg.message) @@ -711,49 +540,49 @@ class ExecutionChecker(ResultVisitor): try: host = groups.group(1) except (AttributeError, IndexError): - host = u"" + host = "" try: sock = groups.group(2) except (AttributeError, IndexError): - sock = u"" + sock = "" runtime = { - u"source_type": u"node", - u"source_id": host, - u"msg_type": u"metric", - u"log_level": u"INFO", - u"timestamp": msg.timestamp, - u"msg": u"show_runtime", - u"host": host, - u"socket": sock, - u"data": list() + "source_type": "node", + "source_id": host, + "msg_type": "metric", + "log_level": "INFO", + "timestamp": msg.timestamp, + "msg": "show_runtime", + "host": host, + "socket": sock, + "data": list() } for line in msg.message.splitlines(): - if not line.startswith(u"vpp_runtime_"): + if not line.startswith("vpp_runtime_"): continue try: - params, value, timestamp = line.rsplit(u" ", maxsplit=2) - cut = params.index(u"{") - name = params[:cut].split(u"_", maxsplit=2)[-1] + params, value, timestamp = line.rsplit(" ", maxsplit=2) + cut = params.index("{") + name = params[:cut].split("_", maxsplit=2)[-1] labels = eval( - u"dict" + params[cut:].replace('{', '(').replace('}', ')') + "dict" + params[cut:].replace('{', '(').replace('}', ')') ) - labels[u"graph_node"] = labels.pop(u"name") - runtime[u"data"].append( + labels["graph_node"] = labels.pop("name") + runtime["data"].append( { - u"name": name, - u"value": value, - u"timestamp": timestamp, - u"labels": labels + "name": name, + "value": value, + "timestamp": timestamp, + "labels": labels } ) except (TypeError, ValueError, IndexError): continue - self._data[u'tests'][self._test_id][u'telemetry-show-run']\ + self._data['tests'][self._test_id]['telemetry-show-run']\ [f"dut{self._telemetry_msg_counter}"] = copy.copy( { - u"host": host, - u"socket": sock, - u"runtime": runtime + "host": host, + "socket": sock, + "runtime": runtime } ) @@ -768,19 +597,19 @@ class ExecutionChecker(ResultVisitor): """ throughput = { - u"NDR": {u"LOWER": -1.0, u"UPPER": -1.0}, - u"PDR": {u"LOWER": -1.0, u"UPPER": -1.0} + "NDR": {"LOWER": -1.0, "UPPER": -1.0}, + "PDR": {"LOWER": -1.0, "UPPER": -1.0} } - status = u"FAIL" + status = "FAIL" groups = re.search(self.REGEX_NDRPDR_RATE, msg) if groups is not None: try: - throughput[u"NDR"][u"LOWER"] = float(groups.group(1)) - throughput[u"NDR"][u"UPPER"] = float(groups.group(2)) - throughput[u"PDR"][u"LOWER"] = float(groups.group(3)) - throughput[u"PDR"][u"UPPER"] = float(groups.group(4)) - status = u"PASS" + throughput["NDR"]["LOWER"] = float(groups.group(1)) + throughput["NDR"]["UPPER"] = float(groups.group(2)) + throughput["PDR"]["LOWER"] = float(groups.group(3)) + throughput["PDR"]["UPPER"] = float(groups.group(4)) + status = "PASS" except (IndexError, ValueError): pass @@ -797,19 +626,19 @@ class ExecutionChecker(ResultVisitor): """ gbps = { - u"NDR": {u"LOWER": -1.0, u"UPPER": -1.0}, - u"PDR": {u"LOWER": -1.0, u"UPPER": -1.0} + "NDR": {"LOWER": -1.0, "UPPER": -1.0}, + "PDR": {"LOWER": -1.0, "UPPER": -1.0} } - status = u"FAIL" + status = "FAIL" groups = re.search(self.REGEX_NDRPDR_GBPS, msg) if groups is not None: try: - gbps[u"NDR"][u"LOWER"] = float(groups.group(1)) - gbps[u"NDR"][u"UPPER"] = float(groups.group(2)) - gbps[u"PDR"][u"LOWER"] = float(groups.group(3)) - gbps[u"PDR"][u"UPPER"] = float(groups.group(4)) - status = u"PASS" + gbps["NDR"]["LOWER"] = float(groups.group(1)) + gbps["NDR"]["UPPER"] = float(groups.group(2)) + gbps["PDR"]["LOWER"] = float(groups.group(3)) + gbps["PDR"]["UPPER"] = float(groups.group(4)) + status = "PASS" except (IndexError, ValueError): pass @@ -826,17 +655,17 @@ class ExecutionChecker(ResultVisitor): """ throughput = { - u"LOWER": -1.0, - u"UPPER": -1.0 + "LOWER": -1.0, + "UPPER": -1.0 } - status = u"FAIL" + status = "FAIL" groups = re.search(self.REGEX_PLR_RATE, msg) if groups is not None: try: - throughput[u"LOWER"] = float(groups.group(1)) - throughput[u"UPPER"] = float(groups.group(2)) - status = u"PASS" + throughput["LOWER"] = float(groups.group(1)) + throughput["UPPER"] = float(groups.group(2)) + status = "PASS" except (IndexError, ValueError): pass @@ -851,35 +680,35 @@ class ExecutionChecker(ResultVisitor): :rtype: tuple(dict, str) """ latency_default = { - u"min": -1.0, - u"avg": -1.0, - u"max": -1.0, - u"hdrh": u"" + "min": -1.0, + "avg": -1.0, + "max": -1.0, + "hdrh": "" } latency = { - u"NDR": { - u"direction1": copy.copy(latency_default), - u"direction2": copy.copy(latency_default) + "NDR": { + "direction1": copy.copy(latency_default), + "direction2": copy.copy(latency_default) }, - u"PDR": { - u"direction1": copy.copy(latency_default), - u"direction2": copy.copy(latency_default) + "PDR": { + "direction1": copy.copy(latency_default), + "direction2": copy.copy(latency_default) }, - u"LAT0": { - u"direction1": copy.copy(latency_default), - u"direction2": copy.copy(latency_default) + "LAT0": { + "direction1": copy.copy(latency_default), + "direction2": copy.copy(latency_default) }, - u"PDR10": { - u"direction1": copy.copy(latency_default), - u"direction2": copy.copy(latency_default) + "PDR10": { + "direction1": copy.copy(latency_default), + "direction2": copy.copy(latency_default) }, - u"PDR50": { - u"direction1": copy.copy(latency_default), - u"direction2": copy.copy(latency_default) + "PDR50": { + "direction1": copy.copy(latency_default), + "direction2": copy.copy(latency_default) }, - u"PDR90": { - u"direction1": copy.copy(latency_default), - u"direction2": copy.copy(latency_default) + "PDR90": { + "direction1": copy.copy(latency_default), + "direction2": copy.copy(latency_default) }, } @@ -887,7 +716,7 @@ class ExecutionChecker(ResultVisitor): if groups is None: groups = re.search(self.REGEX_NDRPDR_LAT_BASE, msg) if groups is None: - return latency, u"FAIL" + return latency, "FAIL" def process_latency(in_str): """Return object with parsed latency values. @@ -904,42 +733,42 @@ class ExecutionChecker(ResultVisitor): in_list = in_str.split('/', 3) rval = { - u"min": float(in_list[0]), - u"avg": float(in_list[1]), - u"max": float(in_list[2]), - u"hdrh": u"" + "min": float(in_list[0]), + "avg": float(in_list[1]), + "max": float(in_list[2]), + "hdrh": "" } if len(in_list) == 4: - rval[u"hdrh"] = str(in_list[3]) + rval["hdrh"] = str(in_list[3]) return rval try: - latency[u"NDR"][u"direction1"] = process_latency(groups.group(1)) - latency[u"NDR"][u"direction2"] = process_latency(groups.group(2)) - latency[u"PDR"][u"direction1"] = process_latency(groups.group(3)) - latency[u"PDR"][u"direction2"] = process_latency(groups.group(4)) + latency["NDR"]["direction1"] = process_latency(groups.group(1)) + latency["NDR"]["direction2"] = process_latency(groups.group(2)) + latency["PDR"]["direction1"] = process_latency(groups.group(3)) + latency["PDR"]["direction2"] = process_latency(groups.group(4)) if groups.lastindex == 4: - return latency, u"PASS" + return latency, "PASS" except (IndexError, ValueError): pass try: - latency[u"PDR90"][u"direction1"] = process_latency(groups.group(5)) - latency[u"PDR90"][u"direction2"] = process_latency(groups.group(6)) - latency[u"PDR50"][u"direction1"] = process_latency(groups.group(7)) - latency[u"PDR50"][u"direction2"] = process_latency(groups.group(8)) - latency[u"PDR10"][u"direction1"] = process_latency(groups.group(9)) - latency[u"PDR10"][u"direction2"] = process_latency(groups.group(10)) - latency[u"LAT0"][u"direction1"] = process_latency(groups.group(11)) - latency[u"LAT0"][u"direction2"] = process_latency(groups.group(12)) + latency["PDR90"]["direction1"] = process_latency(groups.group(5)) + latency["PDR90"]["direction2"] = process_latency(groups.group(6)) + latency["PDR50"]["direction1"] = process_latency(groups.group(7)) + latency["PDR50"]["direction2"] = process_latency(groups.group(8)) + latency["PDR10"]["direction1"] = process_latency(groups.group(9)) + latency["PDR10"]["direction2"] = process_latency(groups.group(10)) + latency["LAT0"]["direction1"] = process_latency(groups.group(11)) + latency["LAT0"]["direction2"] = process_latency(groups.group(12)) if groups.lastindex == 12: - return latency, u"PASS" + return latency, "PASS" except (IndexError, ValueError): pass - return latency, u"FAIL" + return latency, "FAIL" @staticmethod def _get_hoststack_data(msg, tags): @@ -953,23 +782,23 @@ class ExecutionChecker(ResultVisitor): :rtype: tuple(dict, str) """ result = dict() - status = u"FAIL" + status = "FAIL" - msg = msg.replace(u"'", u'"').replace(u" ", u"") - if u"LDPRELOAD" in tags: + msg = msg.replace("'", '"').replace(" ", "") + if "LDPRELOAD" in tags: try: result = loads(msg) - status = u"PASS" + status = "PASS" except JSONDecodeError: pass - elif u"VPPECHO" in tags: + elif "VPPECHO" in tags: try: - msg_lst = msg.replace(u"}{", u"} {").split(u" ") + msg_lst = msg.replace("}{", "} {").split(" ") result = dict( client=loads(msg_lst[0]), server=loads(msg_lst[1]) ) - status = u"PASS" + status = "PASS" except (JSONDecodeError, IndexError): pass @@ -986,23 +815,23 @@ class ExecutionChecker(ResultVisitor): :rtype: tuple(dict, str) """ result = dict() - status = u"FAIL" + status = "FAIL" groups = re.search(self.REGEX_VSAP_MSG_INFO, msg) if groups is not None: try: - result[u"transfer-rate"] = float(groups.group(1)) * 1e3 - result[u"latency"] = float(groups.group(2)) - result[u"completed-requests"] = int(groups.group(3)) - result[u"failed-requests"] = int(groups.group(4)) - result[u"bytes-transferred"] = int(groups.group(5)) - if u"TCP_CPS"in tags: - result[u"cps"] = float(groups.group(6)) - elif u"TCP_RPS" in tags: - result[u"rps"] = float(groups.group(6)) + result["transfer-rate"] = float(groups.group(1)) * 1e3 + result["latency"] = float(groups.group(2)) + result["completed-requests"] = int(groups.group(3)) + result["failed-requests"] = int(groups.group(4)) + result["bytes-transferred"] = int(groups.group(5)) + if "TCP_CPS"in tags: + result["cps"] = float(groups.group(6)) + elif "TCP_RPS" in tags: + result["rps"] = float(groups.group(6)) else: return result, status - status = u"PASS" + status = "PASS" except (IndexError, ValueError): pass @@ -1016,8 +845,10 @@ class ExecutionChecker(ResultVisitor): :returns: Nothing. """ if self.start_suite(suite) is not False: + suite.setup.visit(self) suite.suites.visit(self) suite.tests.visit(self) + suite.teardown.visit(self) self.end_suite(suite) def start_suite(self, suite): @@ -1027,31 +858,18 @@ class ExecutionChecker(ResultVisitor): :type suite: Suite :returns: Nothing. """ - try: parent_name = suite.parent.name except AttributeError: return - self._data[u"suites"][suite.longname.lower(). - replace(u'"', u"'"). - replace(u" ", u"_")] = { - u"name": suite.name.lower(), - u"doc": suite.doc, - u"parent": parent_name, - u"level": len(suite.longname.split(u".")) - } - - suite.setup.visit(self) - suite.teardown.visit(self) - - def end_suite(self, suite): - """Called when suite ends. - - :param suite: Suite to process. - :type suite: Suite - :returns: Nothing. - """ + self._data["suites"][suite.longname.lower().replace('"', "'").\ + replace(" ", "_")] = { + "name": suite.name.lower(), + "doc": suite.doc, + "parent": parent_name, + "level": len(suite.longname.split(".")) + } def visit_test(self, test): """Implements traversing through the test. @@ -1090,54 +908,50 @@ class ExecutionChecker(ResultVisitor): # Change the TC long name and name if defined in the mapping table longname = self._mapping.get(longname_orig, None) if longname is not None: - name = longname.split(u'.')[-1] - logging.debug( - f"{self._data[u'metadata']}\n{longname_orig}\n{longname}\n" - f"{name}" - ) + name = longname.split('.')[-1] else: longname = longname_orig name = test.name.lower() # Remove TC number from the TC long name (backward compatibility): - self._test_id = re.sub(self.REGEX_TC_NUMBER, u"", longname) + self._test_id = re.sub(self.REGEX_TC_NUMBER, "", longname) # Remove TC number from the TC name (not needed): - test_result[u"name"] = re.sub(self.REGEX_TC_NUMBER, "", name) + test_result["name"] = re.sub(self.REGEX_TC_NUMBER, "", name) - test_result[u"parent"] = test.parent.name.lower() - test_result[u"tags"] = tags + test_result["parent"] = test.parent.name.lower() + test_result["tags"] = tags test_result["doc"] = test.doc - test_result[u"type"] = u"" - test_result[u"status"] = test.status - test_result[u"starttime"] = test.starttime - test_result[u"endtime"] = test.endtime - - if test.status == u"PASS": - if u"NDRPDR" in tags: - if u"TCP_PPS" in tags or u"UDP_PPS" in tags: - test_result[u"msg"] = self._get_data_from_pps_test_msg( + test_result["type"] = "" + test_result["status"] = test.status + test_result["starttime"] = test.starttime + test_result["endtime"] = test.endtime + + if test.status == "PASS": + if "NDRPDR" in tags: + if "TCP_PPS" in tags or "UDP_PPS" in tags: + test_result["msg"] = self._get_data_from_pps_test_msg( test.message) - elif u"TCP_CPS" in tags or u"UDP_CPS" in tags: - test_result[u"msg"] = self._get_data_from_cps_test_msg( + elif "TCP_CPS" in tags or "UDP_CPS" in tags: + test_result["msg"] = self._get_data_from_cps_test_msg( test.message) else: - test_result[u"msg"] = self._get_data_from_perf_test_msg( + test_result["msg"] = self._get_data_from_perf_test_msg( test.message) - elif u"MRR" in tags or u"FRMOBL" in tags or u"BMRR" in tags: - test_result[u"msg"] = self._get_data_from_mrr_test_msg( + elif "MRR" in tags or "FRMOBL" in tags or "BMRR" in tags: + test_result["msg"] = self._get_data_from_mrr_test_msg( test.message) else: - test_result[u"msg"] = test.message + test_result["msg"] = test.message else: - test_result[u"msg"] = test.message + test_result["msg"] = test.message - if u"PERFTEST" in tags and u"TREX" not in tags: + if "PERFTEST" in tags and "TREX" not in tags: # Replace info about cores (e.g. -1c-) with the info about threads # and cores (e.g. -1t1c-) in the long test case names and in the # test case names if necessary. tag_count = 0 tag_tc = str() - for tag in test_result[u"tags"]: + for tag in test_result["tags"]: groups = re.search(self.REGEX_TC_TAG, tag) if groups: tag_count += 1 @@ -1148,106 +962,101 @@ class ExecutionChecker(ResultVisitor): self.REGEX_TC_NAME_NEW, f"-{tag_tc.lower()}-", self._test_id, count=1 ) - test_result[u"name"] = re.sub( + test_result["name"] = re.sub( self.REGEX_TC_NAME_NEW, f"-{tag_tc.lower()}-", test_result["name"], count=1 ) else: - test_result[u"status"] = u"FAIL" - self._data[u"tests"][self._test_id] = test_result + test_result["status"] = "FAIL" + self._data["tests"][self._test_id] = test_result logging.debug( f"The test {self._test_id} has no or more than one " f"multi-threading tags.\n" - f"Tags: {test_result[u'tags']}" + f"Tags: {test_result['tags']}" ) return - if u"DEVICETEST" in tags: - test_result[u"type"] = u"DEVICETEST" - elif u"NDRPDR" in tags: - if u"TCP_CPS" in tags or u"UDP_CPS" in tags: - test_result[u"type"] = u"CPS" + if "DEVICETEST" in tags: + test_result["type"] = "DEVICETEST" + elif "NDRPDR" in tags: + if "TCP_CPS" in tags or "UDP_CPS" in tags: + test_result["type"] = "CPS" else: - test_result[u"type"] = u"NDRPDR" - if test.status == u"PASS": - test_result[u"throughput"], test_result[u"status"] = \ + test_result["type"] = "NDRPDR" + if test.status == "PASS": + test_result["throughput"], test_result["status"] = \ self._get_ndrpdr_throughput(test.message) - test_result[u"gbps"], test_result[u"status"] = \ + test_result["gbps"], test_result["status"] = \ self._get_ndrpdr_throughput_gbps(test.message) - test_result[u"latency"], test_result[u"status"] = \ + test_result["latency"], test_result["status"] = \ self._get_ndrpdr_latency(test.message) - elif u"MRR" in tags or u"FRMOBL" in tags or u"BMRR" in tags: - if u"MRR" in tags: - test_result[u"type"] = u"MRR" + elif "MRR" in tags or "FRMOBL" in tags or "BMRR" in tags: + if "MRR" in tags: + test_result["type"] = "MRR" else: - test_result[u"type"] = u"BMRR" - if test.status == u"PASS": - test_result[u"result"] = dict() + test_result["type"] = "BMRR" + if test.status == "PASS": + test_result["result"] = dict() groups = re.search(self.REGEX_BMRR, test.message) if groups is not None: items_str = groups.group(1) items_float = [ - float(item.strip().replace(u"'", u"")) + float(item.strip().replace("'", "")) for item in items_str.split(",") ] # Use whole list in CSIT-1180. stats = jumpavg.AvgStdevStats.for_runs(items_float) - test_result[u"result"][u"samples"] = items_float - test_result[u"result"][u"receive-rate"] = stats.avg - test_result[u"result"][u"receive-stdev"] = stats.stdev + test_result["result"]["samples"] = items_float + test_result["result"]["receive-rate"] = stats.avg + test_result["result"]["receive-stdev"] = stats.stdev else: groups = re.search(self.REGEX_MRR, test.message) - test_result[u"result"][u"receive-rate"] = \ + test_result["result"]["receive-rate"] = \ float(groups.group(3)) / float(groups.group(1)) - elif u"SOAK" in tags: - test_result[u"type"] = u"SOAK" - if test.status == u"PASS": - test_result[u"throughput"], test_result[u"status"] = \ + elif "SOAK" in tags: + test_result["type"] = "SOAK" + if test.status == "PASS": + test_result["throughput"], test_result["status"] = \ self._get_plr_throughput(test.message) - elif u"LDP_NGINX" in tags: - test_result[u"type"] = u"LDP_NGINX" - test_result[u"result"], test_result[u"status"] = \ + elif "LDP_NGINX" in tags: + test_result["type"] = "LDP_NGINX" + test_result["result"], test_result["status"] = \ self._get_vsap_data(test.message, tags) - elif u"HOSTSTACK" in tags: - test_result[u"type"] = u"HOSTSTACK" - if test.status == u"PASS": - test_result[u"result"], test_result[u"status"] = \ + elif "HOSTSTACK" in tags: + test_result["type"] = "HOSTSTACK" + if test.status == "PASS": + test_result["result"], test_result["status"] = \ self._get_hoststack_data(test.message, tags) - elif u"RECONF" in tags: - test_result[u"type"] = u"RECONF" - if test.status == u"PASS": - test_result[u"result"] = None + elif "RECONF" in tags: + test_result["type"] = "RECONF" + if test.status == "PASS": + test_result["result"] = None try: grps_loss = re.search(self.REGEX_RECONF_LOSS, test.message) grps_time = re.search(self.REGEX_RECONF_TIME, test.message) - test_result[u"result"] = { - u"loss": int(grps_loss.group(1)), - u"time": float(grps_time.group(1)) + test_result["result"] = { + "loss": int(grps_loss.group(1)), + "time": float(grps_time.group(1)) } except (AttributeError, IndexError, ValueError, TypeError): - test_result[u"status"] = u"FAIL" + test_result["status"] = "FAIL" else: - test_result[u"status"] = u"FAIL" - - self._data[u"tests"][self._test_id] = test_result + test_result["status"] = "FAIL" - def end_test(self, test): - """Called when test ends. + self._data["tests"][self._test_id] = test_result - :param test: Test to process. - :type test: Test - :returns: Nothing. - """ - - def visit_keyword(self, keyword): + def visit_keyword(self, kw): """Implements traversing through the keyword and its child keywords. :param keyword: Keyword to process. :type keyword: Keyword :returns: Nothing. """ - if self.start_keyword(keyword) is not False: - self.end_keyword(keyword) + if self.start_keyword(kw) is not False: + if hasattr(kw, "body"): + kw.body.visit(self) + kw.teardown.visit(self) + self.end_keyword(kw) def start_keyword(self, keyword): """Called when keyword starts. Default implementation does nothing. @@ -1256,15 +1065,7 @@ class ExecutionChecker(ResultVisitor): :type keyword: Keyword :returns: Nothing. """ - try: - if keyword.type in ("setup", "SETUP"): - self.visit_setup_kw(keyword) - elif keyword.type in ("teardown", "TEARDOWN"): - self.visit_teardown_kw(keyword) - else: - self.visit_test_kw(keyword) - except AttributeError: - pass + self._kw_name = keyword.name def end_keyword(self, keyword): """Called when keyword ends. Default implementation does nothing. @@ -1273,126 +1074,8 @@ class ExecutionChecker(ResultVisitor): :type keyword: Keyword :returns: Nothing. """ - - def visit_test_kw(self, test_kw): - """Implements traversing through the test keyword and its child - keywords. - - :param test_kw: Keyword to process. - :type test_kw: Keyword - :returns: Nothing. - """ - for keyword in test_kw.body: - if self.start_test_kw(keyword) is not False: - self.visit_test_kw(keyword) - self.end_test_kw(keyword) - - def start_test_kw(self, test_kw): - """Called when test keyword starts. Default implementation does - nothing. - - :param test_kw: Keyword to process. - :type test_kw: Keyword - :returns: Nothing. - """ - if not self._process_oper: - return - - if test_kw.name.count(u"Run Telemetry On All Duts"): - self._msg_type = u"test-telemetry" - self._telemetry_kw_counter += 1 - elif test_kw.name.count(u"Show Runtime On All Duts"): - self._msg_type = u"test-show-runtime" - self._sh_run_counter += 1 - else: - return - test_kw.messages.visit(self) - - def end_test_kw(self, test_kw): - """Called when keyword ends. Default implementation does nothing. - - :param test_kw: Keyword to process. - :type test_kw: Keyword - :returns: Nothing. - """ - - def visit_setup_kw(self, setup_kw): - """Implements traversing through the teardown keyword and its child - keywords. - - :param setup_kw: Keyword to process. - :type setup_kw: Keyword - :returns: Nothing. - """ - for keyword in setup_kw.setup: - if self.start_setup_kw(keyword) is not False: - self.visit_setup_kw(keyword) - self.end_setup_kw(keyword) - for keyword in setup_kw.body: - if self.start_setup_kw(keyword) is not False: - self.visit_setup_kw(keyword) - self.end_setup_kw(keyword) - - def start_setup_kw(self, setup_kw): - """Called when teardown keyword starts. Default implementation does - nothing. - - :param setup_kw: Keyword to process. - :type setup_kw: Keyword - :returns: Nothing. - """ - if setup_kw.name.count(u"Show Vpp Version On All Duts") \ - and not self._version: - self._msg_type = u"vpp-version" - elif setup_kw.name.count(u"Install Dpdk Framework On All Duts") and \ - not self._version: - self._msg_type = u"dpdk-version" - elif setup_kw.name.count(u"Setup Framework") and not self._testbed: - self._msg_type = u"testbed" - else: - return - setup_kw.messages.visit(self) - - def end_setup_kw(self, setup_kw): - """Called when keyword ends. Default implementation does nothing. - - :param setup_kw: Keyword to process. - :type setup_kw: Keyword - :returns: Nothing. - """ - - def visit_teardown_kw(self, teardown_kw): - """Implements traversing through the teardown keyword and its child - keywords. - - :param teardown_kw: Keyword to process. - :type teardown_kw: Keyword - :returns: Nothing. - """ - for keyword in teardown_kw.body: - if self.start_teardown_kw(keyword) is not False: - self.visit_teardown_kw(keyword) - self.end_teardown_kw(keyword) - - def start_teardown_kw(self, teardown_kw): - """Called when teardown keyword starts - - :param teardown_kw: Keyword to process. - :type teardown_kw: Keyword - :returns: Nothing. - """ - if teardown_kw.name.count(u"Show Papi History On All Duts"): - self._conf_history_lookup_nr = 0 - self._msg_type = u"teardown-papi-history" - teardown_kw.messages.visit(self) - - def end_teardown_kw(self, teardown_kw): - """Called when keyword ends. Default implementation does nothing. - - :param teardown_kw: Keyword to process. - :type teardown_kw: Keyword - :returns: Nothing. - """ + _ = keyword + self._kw_name = None def visit_message(self, msg): """Implements visiting the message. @@ -1412,16 +1095,26 @@ class ExecutionChecker(ResultVisitor): :type msg: Message :returns: Nothing. """ - if self._msg_type: - self.parse_msg[self._msg_type](msg) - - def end_message(self, msg): - """Called when message ends. Default implementation does nothing. - - :param msg: Message to process. - :type msg: Message - :returns: Nothing. - """ + if self._kw_name is None: + return + elif self._kw_name.count("Run Telemetry On All Duts"): + self._telemetry_kw_counter += 1 + self._get_telemetry(msg) + elif self._kw_name.count("Show Runtime On All Duts"): + self._sh_run_counter += 1 + self._get_show_run(msg) + elif self._kw_name.count("Show Vpp Version On All Duts"): + if not self._version: + self._get_vpp_version(msg) + elif self._kw_name.count("Install Dpdk Framework On All Duts"): + if not self._version: + self._get_dpdk_version(msg) + elif self._kw_name.count("Setup Framework"): + if not self._testbed: + self._get_testbed(msg) + elif self._kw_name.count("Show Papi History On All Duts"): + self._conf_history_lookup_nr = 0 + self._get_papi_history(msg) class InputData: @@ -1477,7 +1170,7 @@ class InputData: :returns: Metadata :rtype: pandas.Series """ - return self.data[job][build][u"metadata"] + return self.data[job][build]["metadata"] def suites(self, job, build): """Getter - suites @@ -1489,7 +1182,7 @@ class InputData: :returns: Suites. :rtype: pandas.Series """ - return self.data[job][str(build)][u"suites"] + return self.data[job][str(build)]["suites"] def tests(self, job, build): """Getter - tests @@ -1501,7 +1194,7 @@ class InputData: :returns: Tests. :rtype: pandas.Series """ - return self.data[job][build][u"tests"] + return self.data[job][build]["tests"] def _parse_tests(self, job, build): """Process data from robot output.xml file and return JSON structured @@ -1516,11 +1209,11 @@ class InputData: """ metadata = { - u"job": job, - u"build": build + "job": job, + "build": build } - with open(build[u"file-name"], u'r') as data_file: + with open(build["file-name"], 'r') as data_file: try: result = ExecutionResult(data_file) except errors.DataError as err: @@ -1530,11 +1223,11 @@ class InputData: return None process_oper = False - if u"-vpp-perf-report-coverage-" in job: + if "-vpp-perf-report-coverage-" in job: process_oper = True - # elif u"-vpp-perf-report-iterative-" in job: + # elif "-vpp-perf-report-iterative-" in job: # # Exceptions for TBs where we do not have coverage data: - # for item in (u"-2n-icx", ): + # for item in ("-2n-icx", ): # if item in job: # process_oper = True # break @@ -1543,14 +1236,14 @@ class InputData: ) result.visit(checker) - checker.data[u"metadata"][u"tests_total"] = \ + checker.data["metadata"]["tests_total"] = \ result.statistics.total.total - checker.data[u"metadata"][u"tests_passed"] = \ + checker.data["metadata"]["tests_passed"] = \ result.statistics.total.passed - checker.data[u"metadata"][u"tests_failed"] = \ + checker.data["metadata"]["tests_failed"] = \ result.statistics.total.failed - checker.data[u"metadata"][u"elapsedtime"] = result.suite.elapsedtime - checker.data[u"metadata"][u"generated"] = result.suite.endtime[:14] + checker.data["metadata"]["elapsedtime"] = result.suite.elapsedtime + checker.data["metadata"]["generated"] = result.suite.endtime[:14] return checker.data @@ -1570,9 +1263,9 @@ class InputData: :type repeat: int """ - logging.info(f"Processing the job/build: {job}: {build[u'build']}") + logging.info(f"Processing the job/build: {job}: {build['build']}") - state = u"failed" + state = "failed" success = False data = None do_repeat = repeat @@ -1584,54 +1277,54 @@ class InputData: if not success: logging.error( f"It is not possible to download the input data file from the " - f"job {job}, build {build[u'build']}, or it is damaged. " + f"job {job}, build {build['build']}, or it is damaged. " f"Skipped." ) if success: - logging.info(f" Processing data from build {build[u'build']}") + logging.info(f" Processing data from build {build['build']}") data = self._parse_tests(job, build) if data is None: logging.error( f"Input data file from the job {job}, build " - f"{build[u'build']} is damaged. Skipped." + f"{build['build']} is damaged. Skipped." ) else: - state = u"processed" + state = "processed" try: - remove(build[u"file-name"]) + remove(build["file-name"]) except OSError as err: logging.error( - f"Cannot remove the file {build[u'file-name']}: {repr(err)}" + f"Cannot remove the file {build['file-name']}: {repr(err)}" ) # If the time-period is defined in the specification file, remove all # files which are outside the time period. is_last = False - timeperiod = self._cfg.environment.get(u"time-period", None) + timeperiod = self._cfg.environment.get("time-period", None) if timeperiod and data: now = dt.utcnow() timeperiod = timedelta(int(timeperiod)) - metadata = data.get(u"metadata", None) + metadata = data.get("metadata", None) if metadata: - generated = metadata.get(u"generated", None) + generated = metadata.get("generated", None) if generated: - generated = dt.strptime(generated, u"%Y%m%d %H:%M") + generated = dt.strptime(generated, "%Y%m%d %H:%M") if (now - generated) > timeperiod: # Remove the data and the file: - state = u"removed" + state = "removed" data = None is_last = True logging.info( - f" The build {job}/{build[u'build']} is " + f" The build {job}/{build['build']} is " f"outdated, will be removed." ) return { - u"data": data, - u"state": state, - u"job": job, - u"build": build, - u"last": is_last + "data": data, + "state": state, + "job": job, + "build": build, + "last": is_last } def download_and_parse_data(self, repeat=1): @@ -1643,30 +1336,30 @@ class InputData: :type repeat: int """ - logging.info(u"Downloading and parsing input files ...") + logging.info("Downloading and parsing input files ...") for job, builds in self._cfg.input.items(): for build in builds: result = self._download_and_parse_build(job, build, repeat) - if result[u"last"]: + if result["last"]: break - build_nr = result[u"build"][u"build"] + build_nr = result["build"]["build"] - if result[u"data"]: - data = result[u"data"] + if result["data"]: + data = result["data"] build_data = pd.Series({ - u"metadata": pd.Series( - list(data[u"metadata"].values()), - index=list(data[u"metadata"].keys()) + "metadata": pd.Series( + list(data["metadata"].values()), + index=list(data["metadata"].keys()) ), - u"suites": pd.Series( - list(data[u"suites"].values()), - index=list(data[u"suites"].keys()) + "suites": pd.Series( + list(data["suites"].values()), + index=list(data["suites"].keys()) ), - u"tests": pd.Series( - list(data[u"tests"].values()), - index=list(data[u"tests"].keys()) + "tests": pd.Series( + list(data["tests"].values()), + index=list(data["tests"].keys()) ) }) @@ -1674,27 +1367,27 @@ class InputData: self._input_data[job] = pd.Series(dtype="float64") self._input_data[job][str(build_nr)] = build_data self._cfg.set_input_file_name( - job, build_nr, result[u"build"][u"file-name"] + job, build_nr, result["build"]["file-name"] ) - self._cfg.set_input_state(job, build_nr, result[u"state"]) + self._cfg.set_input_state(job, build_nr, result["state"]) mem_alloc = \ resource.getrusage(resource.RUSAGE_SELF).ru_maxrss / 1000 logging.info(f"Memory allocation: {mem_alloc:.0f}MB") - logging.info(u"Done.") + logging.info("Done.") msg = f"Successful downloads from the sources:\n" - for source in self._cfg.environment[u"data-sources"]: - if source[u"successful-downloads"]: + for source in self._cfg.environment["data-sources"]: + if source["successful-downloads"]: msg += ( - f"{source[u'url']}/{source[u'path']}/" - f"{source[u'file-name']}: " - f"{source[u'successful-downloads']}\n" + f"{source['url']}/{source['path']}/" + f"{source['file-name']}: " + f"{source['successful-downloads']}\n" ) logging.info(msg) - def process_local_file(self, local_file, job=u"local", build_nr=1, + def process_local_file(self, local_file, job="local", build_nr=1, replace=True): """Process local XML file given as a command-line parameter. @@ -1714,14 +1407,14 @@ class InputData: raise PresentationError(f"The file {local_file} does not exist.") try: - build_nr = int(local_file.split(u"/")[-1].split(u".")[0]) + build_nr = int(local_file.split("/")[-1].split(".")[0]) except (IndexError, ValueError): pass build = { - u"build": build_nr, - u"status": u"failed", - u"file-name": local_file + "build": build_nr, + "status": "failed", + "file-name": local_file } if replace: self._cfg.input = dict() @@ -1735,17 +1428,17 @@ class InputData: ) build_data = pd.Series({ - u"metadata": pd.Series( - list(data[u"metadata"].values()), - index=list(data[u"metadata"].keys()) + "metadata": pd.Series( + list(data["metadata"].values()), + index=list(data["metadata"].keys()) ), - u"suites": pd.Series( - list(data[u"suites"].values()), - index=list(data[u"suites"].keys()) + "suites": pd.Series( + list(data["suites"].values()), + index=list(data["suites"].keys()) ), - u"tests": pd.Series( - list(data[u"tests"].values()), - index=list(data[u"tests"].keys()) + "tests": pd.Series( + list(data["tests"].values()), + index=list(data["tests"].keys()) ) }) @@ -1753,7 +1446,7 @@ class InputData: self._input_data[job] = pd.Series(dtype="float64") self._input_data[job][str(build_nr)] = build_data - self._cfg.set_input_state(job, build_nr, u"processed") + self._cfg.set_input_state(job, build_nr, "processed") def process_local_directory(self, local_dir, replace=True): """Process local directory with XML file(s). The directory is processed @@ -1816,7 +1509,7 @@ class InputData: self.process_local_file(local_file, job, idx + 1, replace=False) @staticmethod - def _end_of_tag(tag_filter, start=0, closer=u"'"): + def _end_of_tag(tag_filter, start=0, closer="'"): """Return the index of character in the string which is the end of tag. :param tag_filter: The string where the end of tag is being searched. @@ -1849,9 +1542,9 @@ class InputData: if index is None: return tag_filter index += 1 - tag_filter = tag_filter[:index] + u" in tags" + tag_filter[index:] + tag_filter = tag_filter[:index] + " in tags" + tag_filter[index:] - def filter_data(self, element, params=None, data=None, data_set=u"tests", + def filter_data(self, element, params=None, data=None, data_set="tests", continue_on_error=False): """Filter required data from the given jobs and builds. @@ -1891,22 +1584,22 @@ class InputData: try: if data_set == "suites": - cond = u"True" - elif element[u"filter"] in (u"all", u"template"): - cond = u"True" + cond = "True" + elif element["filter"] in ("all", "template"): + cond = "True" else: - cond = InputData._condition(element[u"filter"]) + cond = InputData._condition(element["filter"]) logging.debug(f" Filter: {cond}") except KeyError: - logging.error(u" No filter defined.") + logging.error(" No filter defined.") return None if params is None: - params = element.get(u"parameters", None) + params = element.get("parameters", None) if params: - params.extend((u"type", u"status")) + params.extend(("type", "status")) - data_to_filter = data if data else element[u"data"] + data_to_filter = data if data else element["data"] data = pd.Series(dtype="float64") try: for job, builds in data_to_filter.items(): @@ -1922,7 +1615,7 @@ class InputData: return None for test_id, test_data in data_dict.items(): - if eval(cond, {u"tags": test_data.get(u"tags", u"")}): + if eval(cond, {"tags": test_data.get("tags", "")}): data[job][str(build)][test_id] = \ pd.Series(dtype="float64") if params is None: @@ -1935,7 +1628,7 @@ class InputData: test_data[param] except KeyError: data[job][str(build)][test_id][param] =\ - u"No Data" + "No Data" return data except (KeyError, IndexError, ValueError) as err: @@ -1954,7 +1647,7 @@ class InputData: ) return None - def filter_tests_by_name(self, element, params=None, data_set=u"tests", + def filter_tests_by_name(self, element, params=None, data_set="tests", continue_on_error=False): """Filter required data from the given jobs and builds. @@ -1989,17 +1682,17 @@ class InputData: :rtype pandas.Series """ - include = element.get(u"include", None) + include = element.get("include", None) if not include: - logging.warning(u"No tests to include, skipping the element.") + logging.warning("No tests to include, skipping the element.") return None if params is None: - params = element.get(u"parameters", None) - if params and u"type" not in params: - params.append(u"type") + params = element.get("parameters", None) + if params and "type" not in params: + params.append("type") - cores = element.get(u"core", None) + cores = element.get("core", None) if cores: tests = list() for core in cores: @@ -2010,7 +1703,7 @@ class InputData: data = pd.Series(dtype="float64") try: - for job, builds in element[u"data"].items(): + for job, builds in element["data"].items(): data[job] = pd.Series(dtype="float64") for build in builds: data[job][str(build)] = pd.Series(dtype="float64") @@ -2036,7 +1729,7 @@ class InputData: test_data[param] except KeyError: data[job][str(build)][ - test_id][param] = u"No Data" + test_id][param] = "No Data" except KeyError as err: if continue_on_error: logging.debug(repr(err)) @@ -2076,7 +1769,7 @@ class InputData: :rtype: pandas.Series """ - logging.info(u" Merging data ...") + logging.info(" Merging data ...") merged_data = pd.Series(dtype="float64") for builds in data.values: @@ -2091,63 +1784,63 @@ class InputData: for job in self._input_data.values: for build in job.values: - for test_id, test_data in build[u"tests"].items(): + for test_id, test_data in build["tests"].items(): print(f"{test_id}") - if test_data.get(u"show-run", None) is None: + if test_data.get("show-run", None) is None: continue - for dut_name, data in test_data[u"show-run"].items(): - if data.get(u"runtime", None) is None: + for dut_name, data in test_data["show-run"].items(): + if data.get("runtime", None) is None: continue - runtime = loads(data[u"runtime"]) + runtime = loads(data["runtime"]) try: - threads_nr = len(runtime[0][u"clocks"]) + threads_nr = len(runtime[0]["clocks"]) except (IndexError, KeyError): continue threads = OrderedDict( {idx: list() for idx in range(threads_nr)}) for item in runtime: for idx in range(threads_nr): - if item[u"vectors"][idx] > 0: - clocks = item[u"clocks"][idx] / \ - item[u"vectors"][idx] - elif item[u"calls"][idx] > 0: - clocks = item[u"clocks"][idx] / \ - item[u"calls"][idx] - elif item[u"suspends"][idx] > 0: - clocks = item[u"clocks"][idx] / \ - item[u"suspends"][idx] + if item["vectors"][idx] > 0: + clocks = item["clocks"][idx] / \ + item["vectors"][idx] + elif item["calls"][idx] > 0: + clocks = item["clocks"][idx] / \ + item["calls"][idx] + elif item["suspends"][idx] > 0: + clocks = item["clocks"][idx] / \ + item["suspends"][idx] else: clocks = 0.0 - if item[u"calls"][idx] > 0: - vectors_call = item[u"vectors"][idx] / \ - item[u"calls"][idx] + if item["calls"][idx] > 0: + vectors_call = item["vectors"][idx] / \ + item["calls"][idx] else: vectors_call = 0.0 - if int(item[u"calls"][idx]) + int( - item[u"vectors"][idx]) + \ - int(item[u"suspends"][idx]): + if int(item["calls"][idx]) + int( + item["vectors"][idx]) + \ + int(item["suspends"][idx]): threads[idx].append([ - item[u"name"], - item[u"calls"][idx], - item[u"vectors"][idx], - item[u"suspends"][idx], + item["name"], + item["calls"][idx], + item["vectors"][idx], + item["suspends"][idx], clocks, vectors_call ]) - print(f"Host IP: {data.get(u'host', '')}, " - f"Socket: {data.get(u'socket', '')}") + print(f"Host IP: {data.get('host', '')}, " + f"Socket: {data.get('socket', '')}") for thread_nr, thread in threads.items(): txt_table = prettytable.PrettyTable( ( - u"Name", - u"Nr of Vectors", - u"Nr of Packets", - u"Suspends", - u"Cycles per Packet", - u"Average Vector Size" + "Name", + "Nr of Vectors", + "Nr of Packets", + "Suspends", + "Cycles per Packet", + "Average Vector Size" ) ) avg = 0.0 @@ -2155,14 +1848,14 @@ class InputData: txt_table.add_row(row) avg += row[-1] if len(thread) == 0: - avg = u"" + avg = "" else: avg = f", Average Vector Size per Node: " \ f"{(avg / len(thread)):.2f}" - th_name = u"main" if thread_nr == 0 \ + th_name = "main" if thread_nr == 0 \ else f"worker_{thread_nr}" print(f"{dut_name}, {th_name}{avg}") - txt_table.float_format = u".2" - txt_table.align = u"r" - txt_table.align[u"Name"] = u"l" + txt_table.float_format = ".2" + txt_table.align = "r" + txt_table.align["Name"] = "l" print(f"{txt_table.get_string()}\n") |