diff options
Diffstat (limited to 'csit.infra.dash/app/cdash/utils')
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/__init__.py | 12 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/anomalies.py | 69 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/constants.py | 469 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/control_panel.py | 87 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/telemetry_data.py | 362 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/tooltips.yaml | 51 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/trigger.py | 65 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/url_processing.py | 99 | ||||
| -rw-r--r-- | csit.infra.dash/app/cdash/utils/utils.py | 895 |
9 files changed, 2109 insertions, 0 deletions
diff --git a/csit.infra.dash/app/cdash/utils/__init__.py b/csit.infra.dash/app/cdash/utils/__init__.py new file mode 100644 index 0000000000..c6a5f639fe --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/__init__.py @@ -0,0 +1,12 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. diff --git a/csit.infra.dash/app/cdash/utils/anomalies.py b/csit.infra.dash/app/cdash/utils/anomalies.py new file mode 100644 index 0000000000..3deece2e04 --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/anomalies.py @@ -0,0 +1,69 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""Functions used by Dash applications to detect anomalies. +""" + +from numpy import isnan + +from ..jumpavg import classify + + +def classify_anomalies(data): + """Process the data and return anomalies and trending values. + + Gather data into groups with average as trend value. + Decorate values within groups to be normal, + the first value of changed average as a regression, or a progression. + + :param data: Full data set with unavailable samples replaced by nan. + :type data: OrderedDict + :returns: Classification and trend values + :rtype: 3-tuple, list of strings, list of floats and list of floats + """ + # NaN means something went wrong. + # Use 0.0 to cause that being reported as a severe regression. + bare_data = [0.0 if isnan(sample) else sample for sample in data.values()] + # TODO: Make BitCountingGroupList a subclass of list again? + group_list = classify(bare_data).group_list + group_list.reverse() # Just to use .pop() for FIFO. + classification = list() + avgs = list() + stdevs = list() + active_group = None + values_left = 0 + avg = 0.0 + stdv = 0.0 + for sample in data.values(): + if isnan(sample): + classification.append("outlier") + avgs.append(sample) + stdevs.append(sample) + continue + if values_left < 1 or active_group is None: + values_left = 0 + while values_left < 1: # Ignore empty groups (should not happen). + active_group = group_list.pop() + values_left = len(active_group.run_list) + avg = active_group.stats.avg + stdv = active_group.stats.stdev + classification.append(active_group.comment) + avgs.append(avg) + stdevs.append(stdv) + values_left -= 1 + continue + classification.append("normal") + avgs.append(avg) + stdevs.append(stdv) + values_left -= 1 + return classification, avgs, stdevs diff --git a/csit.infra.dash/app/cdash/utils/constants.py b/csit.infra.dash/app/cdash/utils/constants.py new file mode 100644 index 0000000000..840766488a --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/constants.py @@ -0,0 +1,469 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""Constants used in CDash. + +"Constant" means a value that keeps its value since initialization. The value +does not need to be hard coded here, but can be read from environment variables. +""" + +import logging + +from dash import html + + +class Constants: + """Constants used in CDash. + """ + + ############################################################################ + # General, application wide constants. + + # Logging settings. + LOG_LEVEL = logging.INFO + LOG_FORMAT = "%(asctime)s: %(levelname)s: %(message)s" + LOG_DATE_FORMAT = "%Y/%m/%d %H:%M:%S" + + # The application title. + TITLE = "FD.io CSIT" + BRAND = "CSIT-Dash" + + # The application description. + DESCRIPTION = "Performance Dashboard" + + # External stylesheets. + EXTERNAL_STYLESHEETS = ["/static/dist/css/bootstrap.css", ] + + # URL to Jenkins + URL_JENKINS = "https://jenkins.fd.io/job/" + + # URL to logs + URL_LOGS = "https://s3-logs.fd.io/vex-yul-rot-jenkins-1/" + + # URL to the documentation + URL_DOC_TRENDING = "https://csit.fd.io/cdocs/methodology/trending/analysis/" + URL_DOC_REL_NOTES = "https://csit.fd.io/cdocs/release_notes/current/" + + # Path and name of the file specifying the HTML layout of the dash + # application. + MAIN_HTML_LAYOUT_FILE = "base_layout.jinja2" + + # Path and name of the file specifying the HTML layout of the dash + # application. + HTML_LAYOUT_FILE = "cdash/templates/dash_layout.jinja2" + + # Application root. + APPLICATIN_ROOT = "/" + + # Data to be downloaded from the parquets specification file. + DATA_SPEC_FILE = "cdash/data/data.yaml" + + # Path to schemas to use when reading data from the parquet. + PATH_TO_SCHEMAS = "cdash/data/_metadata/" + + # The file with tooltips. + TOOLTIP_FILE = "cdash/utils/tooltips.yaml" + + # Maximal value of TIME_PERIOD for data read from the parquets in days. + # Do not change without a good reason. + MAX_TIME_PERIOD = 250 + + # It defines the time period for data read from the parquets in days from + # now back to the past. + # TIME_PERIOD = None - means all data (max MAX_TIME_PERIOD days) is read. + # TIME_PERIOD = MAX_TIME_PERIOD - is the default value + TIME_PERIOD = MAX_TIME_PERIOD # [days] + + ############################################################################ + # General, application wide, layout affecting constants. + + # Add a time delay (in ms) to the spinner being shown + SPINNER_DELAY = 500 + + # If True, clear all inputs in control panel when button "ADD SELECTED" is + # pressed. + CLEAR_ALL_INPUTS = False + + # The element is disabled. + STYLE_DISABLED = {"visibility": "hidden"} + + # The element is enabled and visible. + STYLE_ENABLED = {"visibility": "visible"} + + # The element is not displayed. + STYLE_DONT_DISPLAY = {"display": "none"} + + # The element is displaed. + STYLE_DISPLAY = {"display": "flex"} + + # Checklist "All" is disabled. + CL_ALL_DISABLED = [ + { + "label": "All", + "value": "all", + "disabled": True + } + ] + + # Checklist "All" is enabled, visible and unchecked. + CL_ALL_ENABLED = [ + { + "label": "All", + "value": "all", + "disabled": False + } + ] + + # Placeholder for any element in the layout. + PLACEHOLDER = html.Nobr("") + + # List of drivers used in CSIT. + DRIVERS = ("avf", "af-xdp", "rdma", "dpdk", "mlx5") + + # Labels for input elements (dropdowns, ...). + LABELS = { + "dpdk": "DPDK", + "container_memif": "LXC/DRC Container Memif", + "crypto": "IPSec IPv4 Routing", + "gso": "GSO", + "ip4": "IPv4 Routing", + "ip4_tunnels": "IPv4 Tunnels", + "ip6": "IPv6 Routing", + "ip6_tunnels": "IPv6 Tunnels", + "l2": "L2 Ethernet Switching", + "lb": "Load Balancer", + "srv6": "SRv6 Routing", + "vm_vhost": "VMs vhost-user", + "nfv_density.dcr_memif.chain_ipsec": "CNF Service Chains Routing IPSec", + "nfv_density.vm_vhost.chain_dot1qip4vxlan":"VNF Service Chains Tunnels", + "nfv_density.vm_vhost.chain": "VNF Service Chains Routing", + "nfv_density.dcr_memif.pipeline": "CNF Service Pipelines Routing", + "nfv_density.dcr_memif.chain": "CNF Service Chains Routing", + "hoststack": "Hoststack", + "flow": "Flow", + "l2bd": "L2 Bridge Domain", + "crypto.ethip4": "IPSec IPv4 Routing", + "crypto.ethip6": "IPSec IPv6 Routing", + "interfaces": "Interfaces", + "ip4_tunnels.lisp": "IPv4 Tunnels LISP", + "ip6_tunnels.lisp": "IPv6 Tunnels LISP", + "l2patch": "L2 Patch", + "l2xc": "L2 Cross Connect", + "vm_vhost.ethip4": "VMs vhost-user IPv4 Routing", + "vm_vhost.ethip6": "VMs vhost-user IPv6 Routing" + } + + # URL style. + URL_STYLE = { + "background-color": "#d2ebf5", + "border-color": "#bce1f1", + "color": "#135d7c" + } + + ############################################################################ + # General, normalization constants. + + NORM_FREQUENCY = 2.0 # [GHz] + FREQUENCY = { # [GHz] + "1n-aws": 3.400, + "2n-aws": 3.400, + "2n-c6in": 3.500, + "2n-clx": 2.300, + "2n-icx": 2.600, + "2n-spr": 2.800, + "2n-tx2": 2.500, + "2n-zn2": 2.900, + "3n-alt": 3.000, + "3n-icx": 2.600, + "3n-icxd": 2.000, + "3n-snr": 2.200, + "3n-tsh": 2.200, + "3na-spr": 2.800, + "3nb-spr": 2.800 + } + + ############################################################################ + # General, plots and tables constants. + + PLOT_COLORS = ( + "#1A1110", "#DA2647", "#214FC6", "#01786F", "#BD8260", "#FFD12A", + "#A6E7FF", "#738276", "#C95A49", "#FC5A8D", "#CEC8EF", "#391285", + "#6F2DA8", "#FF878D", "#45A27D", "#FFD0B9", "#FD5240", "#DB91EF", + "#44D7A8", "#4F86F7", "#84DE02", "#FFCFF1", "#614051" + ) + + # Trending, anomalies. + ANOMALY_COLOR = { + "regression": 0.0, + "normal": 0.5, + "progression": 1.0 + } + + COLORSCALE_TPUT = [ + [0.00, "red"], + [0.33, "red"], + [0.33, "white"], + [0.66, "white"], + [0.66, "green"], + [1.00, "green"] + ] + + TICK_TEXT_TPUT = ["Regression", "Normal", "Progression"] + + COLORSCALE_LAT = [ + [0.00, "green"], + [0.33, "green"], + [0.33, "white"], + [0.66, "white"], + [0.66, "red"], + [1.00, "red"] + ] + + TICK_TEXT_LAT = ["Progression", "Normal", "Regression"] + + # Access to the results. + VALUE = { + "mrr": "result_receive_rate_rate_avg", + "ndr": "result_ndr_lower_rate_value", + "pdr": "result_pdr_lower_rate_value", + "mrr-bandwidth": "result_receive_rate_bandwidth_avg", + "ndr-bandwidth": "result_ndr_lower_bandwidth_value", + "pdr-bandwidth": "result_pdr_lower_bandwidth_value", + "latency": "result_latency_forward_pdr_50_avg", + "hoststack-cps": "result_rate_value", + "hoststack-rps": "result_rate_value", + "hoststack-cps-bandwidth": "result_bandwidth_value", + "hoststack-rps-bandwidth": "result_bandwidth_value", + "hoststack-bps": "result_bandwidth_value", + "hoststack-latency": "result_latency_value", + "soak": "result_critical_rate_lower_rate_value", + "soak-bandwidth": "result_critical_rate_lower_bandwidth_value" + } + + VALUE_ITER = { + "mrr": "result_receive_rate_rate_values", + "ndr": "result_ndr_lower_rate_value", + "pdr": "result_pdr_lower_rate_value", + "mrr-bandwidth": "result_receive_rate_bandwidth_avg", + "ndr-bandwidth": "result_ndr_lower_bandwidth_value", + "pdr-bandwidth": "result_pdr_lower_bandwidth_value", + "latency": "result_latency_forward_pdr_50_avg", + "hoststack-cps": "result_rate_value", + "hoststack-rps": "result_rate_value", + "hoststack-cps-bandwidth": "result_bandwidth_value", + "hoststack-rps-bandwidth": "result_bandwidth_value", + "hoststack-bps": "result_bandwidth_value", + "hoststack-latency": "result_latency_value", + "soak": "result_critical_rate_lower_rate_value", + "soak-bandwidth": "result_critical_rate_lower_bandwidth_value" + } + + UNIT = { + "mrr": "result_receive_rate_rate_unit", + "ndr": "result_ndr_lower_rate_unit", + "pdr": "result_pdr_lower_rate_unit", + "mrr-bandwidth": "result_receive_rate_bandwidth_unit", + "ndr-bandwidth": "result_ndr_lower_bandwidth_unit", + "pdr-bandwidth": "result_pdr_lower_bandwidth_unit", + "latency": "result_latency_forward_pdr_50_unit", + "hoststack-cps": "result_rate_unit", + "hoststack-rps": "result_rate_unit", + "hoststack-cps-bandwidth": "result_bandwidth_unit", + "hoststack-rps-bandwidth": "result_bandwidth_unit", + "hoststack-bps": "result_bandwidth_unit", + "hoststack-latency": "result_latency_unit", + "soak": "result_critical_rate_lower_rate_unit", + "soak-bandwidth": "result_critical_rate_lower_bandwidth_unit" + } + + TESTS_WITH_BANDWIDTH = ( + "ndr", + "pdr", + "mrr", + "hoststack-cps", + "hoststack-rps", + "soak" + ) + TESTS_WITH_LATENCY = ( + "pdr", + "hoststack-cps", + "hoststack-rps" + ) + + # Latencies. + LAT_HDRH = ( # Do not change the order + "result_latency_forward_pdr_0_hdrh", + "result_latency_reverse_pdr_0_hdrh", + "result_latency_forward_pdr_10_hdrh", + "result_latency_reverse_pdr_10_hdrh", + "result_latency_forward_pdr_50_hdrh", + "result_latency_reverse_pdr_50_hdrh", + "result_latency_forward_pdr_90_hdrh", + "result_latency_reverse_pdr_90_hdrh", + ) + + # This value depends on latency stream rate (9001 pps) and duration (5s). + # Keep it slightly higher to ensure rounding errors to not remove tick mark. + PERCENTILE_MAX = 99.999501 + + GRAPH_LAT_HDRH_DESC = { + "result_latency_forward_pdr_0_hdrh": "No-load.", + "result_latency_reverse_pdr_0_hdrh": "No-load.", + "result_latency_forward_pdr_10_hdrh": "Low-load, 10% PDR.", + "result_latency_reverse_pdr_10_hdrh": "Low-load, 10% PDR.", + "result_latency_forward_pdr_50_hdrh": "Mid-load, 50% PDR.", + "result_latency_reverse_pdr_50_hdrh": "Mid-load, 50% PDR.", + "result_latency_forward_pdr_90_hdrh": "High-load, 90% PDR.", + "result_latency_reverse_pdr_90_hdrh": "High-load, 90% PDR." + } + + # Operators used to filter data in comparison tables. + OPERATORS = ( + ("contains ", ), + ("lt ", "<"), + ("gt ", ">"), + ("eq ", "="), + ("ge ", ">="), + ("le ", "<="), + ("ne ", "!="), + ("datestartswith ", ) + ) + + ############################################################################ + # News. + + # The title. + NEWS_TITLE = "Failures and Anomalies" + + # The pathname prefix for the application. + NEWS_ROUTES_PATHNAME_PREFIX = "/news/" + + # Time period for regressions and progressions. + NEWS_TIME_PERIOD = TIME_PERIOD # [days] + + # Time periods for summary tables. + NEWS_LAST = 1 # [days] + NEWS_SHORT = 7 # [days] + NEWS_LONG = NEWS_TIME_PERIOD # [days] + + ############################################################################ + # Report. + + # The title. + REPORT_TITLE = "Per Release Performance" + + # The pathname prefix for the application. + REPORT_ROUTES_PATHNAME_PREFIX = "/report/" + + # Layout of plot.ly graphs. + REPORT_GRAPH_LAYOUT_FILE = "cdash/report/layout.yaml" + + # Default name of downloaded file with selected data. + REPORT_DOWNLOAD_FILE_NAME = "iterative_data.csv" + + ############################################################################ + # Comparisons. + + # The title. + COMP_TITLE = "Per Release Performance Comparisons" + + # The pathname prefix for the application. + COMP_ROUTES_PATHNAME_PREFIX = "/comparisons/" + + # Default name of downloaded file with selected data. + COMP_DOWNLOAD_FILE_NAME = "comparison_data.csv" + + # This parameter specifies the method to use for estimating the percentile. + # Possible values: + # - inverted_cdf + # - averaged_inverted_cdf + # - closest_observation + # - interpolated_inverted_cdf + # - hazen + # - weibull + # - linear (default) + # - median_unbiased + # - normal_unbiased + COMP_PERCENTILE_METHOD = "linear" + + # Extreme or mild outlier? + OUTLIER_EXTREME = 3 + OUTLIER_MILD = 1.5 + COMP_OUTLIER_TYPE = OUTLIER_EXTREME + + ############################################################################ + # Statistics. + + # The title. + STATS_TITLE = "Test Job Statistics" + + # The pathname prefix for the application. + STATS_ROUTES_PATHNAME_PREFIX = "/stats/" + + # Layout of plot.ly graphs. + STATS_GRAPH_LAYOUT_FILE = "cdash/stats/layout.yaml" + + # The default job displayed when the page is loaded first time. + STATS_DEFAULT_JOB = "csit-vpp-perf-mrr-daily-master-2n-icx" + + # Default name of downloaded file with selected data. + STATS_DOWNLOAD_FILE_NAME = "stats.csv" + + # The width of the bar in the graph in miliseconds. + STATS_BAR_WIDTH_DAILY = 1000 * 3600 * 15 + STATS_BAR_WIDTH_WEEKLY = 1000 * 3600 * 24 + + ############################################################################ + # Trending. + + # The title. + TREND_TITLE = "Performance Trending" + + # The pathname prefix for the application. + TREND_ROUTES_PATHNAME_PREFIX = "/trending/" + + # Layout of plot.ly graphs. + TREND_GRAPH_LAYOUT_FILE = "cdash/trending/layout.yaml" + + # Default name of downloaded file with selected data. + TREND_DOWNLOAD_FILE_NAME = "trending_data.csv" + TELEMETRY_DOWNLOAD_FILE_NAME = "telemetry_data.csv" + + ############################################################################ + # Coverage data. + + # The title. + COVERAGE_TITLE = "Per Release Coverage Data" + + # The pathname prefix for the application. + COVERAGE_ROUTES_PATHNAME_PREFIX = "/coverage/" + + # Default name of downloaded file with selected data. + COVERAGE_DOWNLOAD_FILE_NAME = "coverage_data.csv" + + ############################################################################ + # Search tests. + + # The title. + SEARCH_TITLE = "Search Tests" + + # The pathname prefix for the application. + SEARCH_ROUTES_PATHNAME_PREFIX = "/search/" + + # Layout of plot.ly graphs. + SEARCH_GRAPH_LAYOUT_FILE = "cdash/search/layout.yaml" + + # Default name of downloaded file with selected data. + SEARCH_DOWNLOAD_FILE_NAME = "search_data.csv" + + ############################################################################ diff --git a/csit.infra.dash/app/cdash/utils/control_panel.py b/csit.infra.dash/app/cdash/utils/control_panel.py new file mode 100644 index 0000000000..3da44e3901 --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/control_panel.py @@ -0,0 +1,87 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""A module implementing the control panel data structure. +""" + +from copy import deepcopy +from typing import Any + +class ControlPanel: + """A class representing the control panel. + """ + + def __init__(self, params: dict, panel: dict) -> None: + """Initialisation of the control pannel by default values. If + particular values are provided (parameter "panel") they are set + afterwards. + + :param params: Default values to be set to the control panel. This + dictionary also defines the full set of the control panel's + parameters and their order. + :param panel: Custom values to be set to the control panel. + :type params: dict + :type panel: dict + """ + + if not params: + raise ValueError("The params must be defined.") + self._panel = deepcopy(params) + if panel: + for key in panel: + if key in self._panel: + self._panel[key] = panel[key] + else: + raise AttributeError( + f"The parameter {key} is not defined in the list of " + f"parameters." + ) + + @property + def panel(self) -> dict: + return self._panel + + @property + def values(self) -> tuple: + """Returns the values from the Control panel as a tuple. + + :returns: The values from the Control panel. + :rtype: tuple + """ + return tuple(self._panel.values()) + + def set(self, kwargs: dict=dict()) -> None: + """Set the values of the Control panel. + + :param kwargs: key - value pairs to be set. + :type kwargs: dict + :raises KeyError: If the key in kwargs is not present in the Control + panel. + """ + for key, val in kwargs.items(): + if key in self._panel: + self._panel[key] = val + else: + raise KeyError(f"The key {key} is not defined.") + + def get(self, key: str) -> Any: + """Returns the value of a key from the Control panel. + + :param key: The key which value should be returned. + :type key: str + :returns: The value of the key. + :rtype: any + :raises KeyError: If the key in kwargs is not present in the Control + panel. + """ + return self._panel[key] diff --git a/csit.infra.dash/app/cdash/utils/telemetry_data.py b/csit.infra.dash/app/cdash/utils/telemetry_data.py new file mode 100644 index 0000000000..9975874d96 --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/telemetry_data.py @@ -0,0 +1,362 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""A module implementing the parsing of OpenMetrics data and elementary +operations with it. +""" + + +import binascii +import zlib +import pandas as pd + +from ..trending.graphs import select_trending_data + + +class TelemetryData: + """A class to store and manipulate the telemetry data. + """ + + def __init__(self, tests: list=list()) -> None: + """Initialize the object. + + :param in_data: Input data. + :param tests: List of selected tests. + :type in_data: pandas.DataFrame + :type tests: list + """ + + self._tests = tests + self._data = None + self._unique_metrics = list() + self._unique_metrics_labels = pd.DataFrame() + self._selected_metrics_labels = pd.DataFrame() + + def from_dataframe(self, in_data: pd.DataFrame=pd.DataFrame()) -> None: + """Read the input from pandas DataFrame. + + This method must be called at the beginning to create all data + structures. + """ + + if in_data.empty: + return + + metrics = set() # A set of unique metrics + + # Create a dataframe with metrics for selected tests: + lst_items = list() + for itm in self._tests: + sel_data = select_trending_data(in_data, itm) + if sel_data is not None: + sel_data["test_name"] = itm["id"] + lst_items.append(sel_data) + df = pd.concat(lst_items, ignore_index=True, copy=False) + + # Use only neccessary data: + df = df[[ + "job", + "build", + "dut_type", + "dut_version", + "start_time", + "passed", + "test_name", + "test_type", + "result_receive_rate_rate_avg", + "result_receive_rate_rate_stdev", + "result_receive_rate_rate_unit", + "result_pdr_lower_rate_value", + "result_pdr_lower_rate_unit", + "result_ndr_lower_rate_value", + "result_ndr_lower_rate_unit", + "telemetry" + ]] + # Transform metrics from strings to dataframes: + lst_telemetry = list() + for _, row in df.iterrows(): + d_telemetry = { + "metric": list(), + "labels": list(), # list of tuple(label, value) + "value": list(), + "timestamp": list() + } + + # If there is no telemetry data, use empty dictionary + if row["telemetry"] is None or isinstance(row["telemetry"], float): + lst_telemetry.append(pd.DataFrame(data=d_telemetry)) + continue + + # Read telemetry data + # - list of uncompressed strings List[str, ...], or + # - list with only one compressed string List[str] + try: + tm_data = zlib.decompress( + binascii.a2b_base64(row["telemetry"][0].encode()) + ).decode().split("\n") + except (binascii.Error, zlib.error, AttributeError, IndexError): + tm_data = row["telemetry"] + + # Pre-process telemetry data + for itm in tm_data: + itm_lst = itm.replace("'", "").rsplit(" ", maxsplit=2) + metric, labels = itm_lst[0].split("{") + d_telemetry["metric"].append(metric) + d_telemetry["labels"].append( + [tuple(x.split("=")) for x in labels[:-1].split(",")] + ) + d_telemetry["value"].append(itm_lst[1]) + d_telemetry["timestamp"].append(itm_lst[2]) + + metrics.update(d_telemetry["metric"]) + lst_telemetry.append(pd.DataFrame(data=d_telemetry)) + df["telemetry"] = lst_telemetry + + self._data = df + self._unique_metrics = sorted(metrics) + + def from_json(self, in_data: dict) -> None: + """Read the input data from json. + """ + + df = pd.read_json(in_data) + lst_telemetry = list() + metrics = set() # A set of unique metrics + for _, row in df.iterrows(): + telemetry = pd.DataFrame(row["telemetry"]) + lst_telemetry.append(telemetry) + metrics.update(telemetry["metric"].to_list()) + df["telemetry"] = lst_telemetry + + self._data = df + self._unique_metrics = sorted(metrics) + + def from_metrics(self, in_data: set) -> None: + """Read only the metrics. + """ + self._unique_metrics = in_data + + def from_metrics_with_labels(self, in_data: dict) -> None: + """Read only metrics with labels. + """ + self._unique_metrics_labels = pd.DataFrame.from_dict(in_data) + + def to_json(self) -> str: + """Return the data transformed from dataframe to json. + + :returns: Telemetry data transformed to a json structure. + :rtype: dict + """ + return self._data.to_json() + + @property + def unique_metrics(self) -> list: + """Return a set of unique metrics. + + :returns: A set of unique metrics. + :rtype: set + """ + return self._unique_metrics + + @property + def unique_metrics_with_labels(self) -> dict: + """ + """ + return self._unique_metrics_labels.to_dict() + + def get_selected_labels(self, metrics: list) -> dict: + """Return a dictionary with labels (keys) and all their possible values + (values) for all selected 'metrics'. + + :param metrics: List of metrics we are interested in. + :type metrics: list + :returns: A dictionary with labels and all their possible values. + :rtype: dict + """ + + lst_labels = list() + tmp_labels = dict() + for _, row in self._data.iterrows(): + telemetry = row["telemetry"] + for itm in metrics: + df = telemetry.loc[(telemetry["metric"] == itm)] + lst_labels.append(df) + for _, tm in df.iterrows(): + for label in tm["labels"]: + if label[0] not in tmp_labels: + tmp_labels[label[0]] = set() + tmp_labels[label[0]].add(label[1]) + + df_labels = pd.concat(lst_labels, ignore_index=True, copy=False) + selected_labels = dict() + for key in sorted(tmp_labels): + selected_labels[key] = sorted(tmp_labels[key]) + + self._unique_metrics_labels = df_labels[["metric", "labels"]].\ + loc[df_labels[["metric", "labels"]].astype(str).\ + drop_duplicates().index] + + return selected_labels + + @property + def str_metrics(self) -> str: + """Returns all unique metrics as a string. + """ + return TelemetryData.metrics_to_str(self._unique_metrics_labels) + + @staticmethod + def metrics_to_str(in_data: pd.DataFrame) -> str: + """Convert metrics from pandas dataframe to string. Metrics in string + are separated by '\n'. + + :param in_data: Metrics to be converted to a string. + :type in_data: pandas.DataFrame + :returns: Metrics as a string. + :rtype: str + """ + metrics = str() + for _, row in in_data.iterrows(): + labels = ','.join([f"{itm[0]}='{itm[1]}'" for itm in row["labels"]]) + metrics += f"{row['metric']}{{{labels}}}\n" + return metrics[:-1] + + def search_unique_metrics(self, string: str) -> list: + """Return a list of metrics which name includes the given string. + + :param string: A string which must be in the name of metric. + :type string: str + :returns: A list of metrics which name includes the given string. + :rtype: list + """ + return [itm for itm in self._unique_metrics if string in itm] + + def filter_selected_metrics_by_labels( + self, + selection: dict + ) -> pd.DataFrame: + """Filter selected unique metrics by labels and their values. + + :param selection: Labels and their values specified by the user. + :type selection: dict + :returns: Pandas dataframe with filtered metrics. + :rtype: pandas.DataFrame + """ + + def _is_selected(labels: list, sel: dict) -> bool: + """Check if the provided 'labels' are selected by the user. + + :param labels: List of labels and their values from a metric. The + items in this lists are two-item-lists whre the first item is + the label and the second one is its value. + :param sel: User selection. The keys are the selected lables and the + values are lists with label values. + :type labels: list + :type sel: dict + :returns: True if the 'labels' are selected by the user. + :rtype: bool + """ + passed = list() + labels = dict(labels) + for key in sel.keys(): + if key in list(labels.keys()): + if sel[key]: + passed.append(labels[key] in sel[key]) + else: + passed.append(True) + else: + passed.append(False) + return bool(passed and all(passed)) + + self._selected_metrics_labels = pd.DataFrame() + lst_items = list() + for _, row in self._unique_metrics_labels.iterrows(): + if _is_selected(row["labels"], selection): + lst_items.append(row.to_frame().T) + self._selected_metrics_labels = \ + pd.concat(lst_items, ignore_index=True, axis=0, copy=False) + return self._selected_metrics_labels + + def select_tm_trending_data( + self, + selection: dict, + ignore_host: bool = False + ) -> pd.DataFrame: + """Select telemetry data for trending based on user's 'selection'. + + The output dataframe includes these columns: + - "job", + - "build", + - "dut_type", + - "dut_version", + - "start_time", + - "passed", + - "test_name", + - "test_id", + - "test_type", + - "result_receive_rate_rate_avg", + - "result_receive_rate_rate_stdev", + - "result_receive_rate_rate_unit", + - "result_pdr_lower_rate_value", + - "result_pdr_lower_rate_unit", + - "result_ndr_lower_rate_value", + - "result_ndr_lower_rate_unit", + - "tm_metric", + - "tm_value". + + :param selection: User's selection (metrics and labels). + :param ignore_host: Ignore 'hostname' and 'hook' labels in metrics. + :type selection: dict + :type ignore_host: bool + :returns: Dataframe with selected data. + :rtype: pandas.DataFrame + """ + + if self._data is None: + return pd.DataFrame() + if self._data.empty: + return pd.DataFrame() + if not selection: + return pd.DataFrame() + + df_sel = pd.DataFrame.from_dict(selection) + lst_rows = list() + for _, row in self._data.iterrows(): + tm_row = row["telemetry"] + for _, tm_sel in df_sel.iterrows(): + df_tmp = tm_row.loc[tm_row["metric"] == tm_sel["metric"]] + for _, tm in df_tmp.iterrows(): + do_it = False + if ignore_host: + if tm["labels"][2:] == tm_sel["labels"][2:]: + labels = ','.join( + [f"{i[0]}='{i[1]}'" for i in tm["labels"][2:]] + ) + do_it = True + else: + if tm["labels"] == tm_sel["labels"]: + labels = ','.join( + [f"{i[0]}='{i[1]}'" for i in tm["labels"]] + ) + do_it = True + if do_it: + row["tm_metric"] = f"{tm['metric']}{{{labels}}}" + row["tm_value"] = tm["value"] + lst_rows.append( + row.drop(labels=["telemetry", ]).to_frame().T + ) + if lst_rows: + return pd.concat( + lst_rows, ignore_index=True, axis=0, copy=False + ).drop_duplicates() + else: + return pd.DataFrame() diff --git a/csit.infra.dash/app/cdash/utils/tooltips.yaml b/csit.infra.dash/app/cdash/utils/tooltips.yaml new file mode 100644 index 0000000000..a51e9ffae4 --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/tooltips.yaml @@ -0,0 +1,51 @@ +help-area: + The area defines a DUT packet path and lookup type. +help-cadence: + The cadence of the Jenkins job which runs the tests. +help-cmp-parameter: + The parameter to be used for comparison. +help-cmp-value: + The value of parameter to be used for comparison. +help-cores: + Number of cores the DUT uses during the test. +help-csit-dut: + The version of CSIT (the part in front of the first hyphen) and the version of + Device under Test (the rest). +help-data-type: + The type of collected data. +help-download: + Download the selected data as a csv file. +help-dut: + Device Under Test (DUT) - In software networking, “device” denotes a specific + piece of software tasked with packet processing. Such device is surrounded + with other software components (such as operating system kernel). +help-dut-ver: + The version of the Device under Test. +help-framesize: + Frame size - size of an Ethernet Layer-2 frame on the wire, including any VLAN + tags (dot1q, dot1ad) and Ethernet FCS, but excluding Ethernet preamble and + inter-frame gap. Measured in Bytes. +help-infra: + Infrastructure is defined by the toplology (number of nodes), processor + architecture, NIC and driver. +help-measurement: + The measured quantity in interest. +help-normalize: + Normalize the results to CPU frequency 2GHz. The results from AWS environment + are not normalized as we do not know the exact value of CPU frequency. +help-release: + The CSIT release. +help-show-latency: + If selected, the latency is included in tables. +help-tbed: + The test bed is defined by toplology (number of nodes) and processor + architecture. +help-test: + The test specification consists of packet encapsulation, VPP packet processing + (packet forwarding mode and packet processing function(s)) and packet + forwarding path. +help-ttype: + Main measured variable. +help-url: + URL with current configuration. If there is no "Copy URL" button, use triple + click. diff --git a/csit.infra.dash/app/cdash/utils/trigger.py b/csit.infra.dash/app/cdash/utils/trigger.py new file mode 100644 index 0000000000..da0768b070 --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/trigger.py @@ -0,0 +1,65 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""A module implementing the processing of a trigger. +""" + +from typing import Any + +from json import loads, JSONDecodeError + + +class Trigger: + """ + """ + def __init__(self, trigger) -> None: + """ + """ + self._id = trigger[0]["prop_id"].split(".") + self._param = self._id[1] + try: + self._id = loads(self._id[0]) + except (JSONDecodeError, TypeError): + # It is a string + self._id = {"type": self._id[0], "index": None} + self._val = trigger[0]["value"] + + def __str__(self) -> str: + return ( + f"\nTrigger:\n" + f" ID: {self._id}\n" + f" Type: {self._id['type']}\n" + f" Index: {self._id['index']}\n" + f" Parameter: {self._param}\n" + f" Value: {self._val}\n" + ) + + @property + def id(self) -> dict: + return self._id + + @property + def type(self) -> str: + return self._id["type"] + + @property + def idx(self) -> Any: + return self._id["index"] + + @property + def parameter(self) -> str: + return self._param + + @property + def value(self) -> Any: + return self._val diff --git a/csit.infra.dash/app/cdash/utils/url_processing.py b/csit.infra.dash/app/cdash/utils/url_processing.py new file mode 100644 index 0000000000..c436ebc830 --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/url_processing.py @@ -0,0 +1,99 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""URL decoding and parsing and URL encoding. +""" + +import logging + +from base64 import urlsafe_b64encode, urlsafe_b64decode +from urllib.parse import urlencode, urlunparse, urlparse, parse_qs +from zlib import compress, decompress +from zlib import error as ZlibErr +from binascii import Error as BinasciiErr + + +def url_encode(params: dict) -> str: + """Encode the URL parameters and zip them and create the whole URL using + given data. + + :param params: All data necessary to create the URL: + - scheme, + - network location, + - path, + - query, + - parameters. + :type params: dict + :returns: Encoded URL. + :rtype: str + """ + + url_params = params.get("params", None) + if url_params: + encoded_params = urlsafe_b64encode( + compress(urlencode(url_params).encode("utf-8"), level=9) + ).rstrip(b"=").decode("utf-8") + else: + encoded_params = str() + + return urlunparse(( + params.get("scheme", "http"), + params.get("netloc", str()), + params.get("path", str()), + str(), # params + params.get("query", str()), + encoded_params + )) + + +def url_decode(url: str) -> dict: + """Parse the given URL and decode the parameters. + + :param url: URL to be parsed and decoded. + :type url: str + :returns: Paresed URL. + :rtype: dict + """ + + try: + parsed_url = urlparse(url) + except ValueError as err: + logging.warning(f"\nThe url {url} is not valid, ignoring.\n{repr(err)}") + return dict() + + if parsed_url.fragment: + try: + padding = b"=" * (4 - (len(parsed_url.fragment) % 4)) + params = parse_qs(decompress( + urlsafe_b64decode( + (parsed_url.fragment.encode("utf-8") + padding) + )).decode("utf-8") + ) + except (BinasciiErr, UnicodeDecodeError, ZlibErr) as err: + logging.warning( + f"\nNot possible to decode the parameters from url: {url}" + f"\nEncoded parameters: '{parsed_url.fragment}'" + f"\n{repr(err)}" + ) + return dict() + else: + params = None + + return { + "scheme": parsed_url.scheme, + "netloc": parsed_url.netloc, + "path": parsed_url.path, + "query": parsed_url.query, + "fragment": parsed_url.fragment, + "params": params + } diff --git a/csit.infra.dash/app/cdash/utils/utils.py b/csit.infra.dash/app/cdash/utils/utils.py new file mode 100644 index 0000000000..306b4f60d1 --- /dev/null +++ b/csit.infra.dash/app/cdash/utils/utils.py @@ -0,0 +1,895 @@ +# Copyright (c) 2024 Cisco and/or its affiliates. +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""Functions used by Dash applications. +""" + +import pandas as pd +import plotly.graph_objects as go +import dash_bootstrap_components as dbc + +import hdrh.histogram +import hdrh.codec + +from math import sqrt +from dash import dcc, no_update, html +from datetime import datetime + +from ..utils.constants import Constants as C +from ..utils.url_processing import url_encode +from ..utils.trigger import Trigger + + +def get_color(idx: int) -> str: + """Returns a color from the list defined in Constants.PLOT_COLORS defined by + its index. + + :param idx: Index of the color. + :type idx: int + :returns: Color defined by hex code. + :trype: str + """ + return C.PLOT_COLORS[idx % len(C.PLOT_COLORS)] + + +def show_tooltip(tooltips:dict, id: str, title: str, + clipboard_id: str=None) -> list: + """Generate list of elements to display a text (e.g. a title) with a + tooltip and optionaly with Copy&Paste icon and the clipboard + functionality enabled. + + :param tooltips: Dictionary with tooltips. + :param id: Tooltip ID. + :param title: A text for which the tooltip will be displayed. + :param clipboard_id: If defined, a Copy&Paste icon is displayed and the + clipboard functionality is enabled. + :type tooltips: dict + :type id: str + :type title: str + :type clipboard_id: str + :returns: List of elements to display a text with a tooltip and + optionaly with Copy&Paste icon. + :rtype: list + """ + + return [ + dcc.Clipboard(target_id=clipboard_id, title="Copy URL") \ + if clipboard_id else str(), + f"{title} ", + dbc.Badge( + id=id, + children="?", + pill=True, + color="white", + text_color="info", + class_name="border ms-1", + ), + dbc.Tooltip( + children=tooltips.get(id, str()), + target=id, + placement="auto" + ) + ] + + +def label(key: str) -> str: + """Returns a label for input elements (dropdowns, ...). + + If the label is not defined, the function returns the provided key. + + :param key: The key to the label defined in Constants.LABELS. + :type key: str + :returns: Label. + :rtype: str + """ + return C.LABELS.get(key, key) + + +def sync_checklists(options: list, sel: list, all: list, id: str) -> tuple: + """Synchronize a checklist with defined "options" with its "All" checklist. + + :param options: List of options for the cheklist. + :param sel: List of selected options. + :param all: List of selected option from "All" checklist. + :param id: ID of a checklist to be used for synchronization. + :returns: Tuple of lists with otions for both checklists. + :rtype: tuple of lists + """ + opts = {v["value"] for v in options} + if id =="all": + sel = list(opts) if all else list() + else: + all = ["all", ] if set(sel) == opts else list() + return sel, all + + +def list_tests(selection: dict) -> list: + """Transform list of tests to a list of dictionaries usable by checkboxes. + + :param selection: List of tests to be displayed in "Selected tests" window. + :type selection: list + :returns: List of dictionaries with "label", "value" pairs for a checkbox. + :rtype: list + """ + if selection: + return [{"label": v["id"], "value": v["id"]} for v in selection] + else: + return list() + + +def get_date(s_date: str) -> datetime: + """Transform string reprezentation of date to datetime.datetime data type. + + :param s_date: String reprezentation of date. + :type s_date: str + :returns: Date as datetime.datetime. + :rtype: datetime.datetime + """ + return datetime(int(s_date[0:4]), int(s_date[5:7]), int(s_date[8:10])) + + +def gen_new_url(url_components: dict, params: dict) -> str: + """Generate a new URL with encoded parameters. + + :param url_components: Dictionary with URL elements. It should contain + "scheme", "netloc" and "path". + :param url_components: URL parameters to be encoded to the URL. + :type parsed_url: dict + :type params: dict + :returns Encoded URL with parameters. + :rtype: str + """ + + if url_components: + return url_encode( + { + "scheme": url_components.get("scheme", ""), + "netloc": url_components.get("netloc", ""), + "path": url_components.get("path", ""), + "params": params + } + ) + else: + return str() + + +def get_duts(df: pd.DataFrame) -> list: + """Get the list of DUTs from the pre-processed information about jobs. + + :param df: DataFrame with information about jobs. + :type df: pandas.DataFrame + :returns: Alphabeticaly sorted list of DUTs. + :rtype: list + """ + return sorted(list(df["dut"].unique())) + + +def get_ttypes(df: pd.DataFrame, dut: str) -> list: + """Get the list of test types from the pre-processed information about + jobs. + + :param df: DataFrame with information about jobs. + :param dut: The DUT for which the list of test types will be populated. + :type df: pandas.DataFrame + :type dut: str + :returns: Alphabeticaly sorted list of test types. + :rtype: list + """ + return sorted(list(df.loc[(df["dut"] == dut)]["ttype"].unique())) + + +def get_cadences(df: pd.DataFrame, dut: str, ttype: str) -> list: + """Get the list of cadences from the pre-processed information about + jobs. + + :param df: DataFrame with information about jobs. + :param dut: The DUT for which the list of cadences will be populated. + :param ttype: The test type for which the list of cadences will be + populated. + :type df: pandas.DataFrame + :type dut: str + :type ttype: str + :returns: Alphabeticaly sorted list of cadences. + :rtype: list + """ + return sorted(list(df.loc[( + (df["dut"] == dut) & + (df["ttype"] == ttype) + )]["cadence"].unique())) + + +def get_test_beds(df: pd.DataFrame, dut: str, ttype: str, cadence: str) -> list: + """Get the list of test beds from the pre-processed information about + jobs. + + :param df: DataFrame with information about jobs. + :param dut: The DUT for which the list of test beds will be populated. + :param ttype: The test type for which the list of test beds will be + populated. + :param cadence: The cadence for which the list of test beds will be + populated. + :type df: pandas.DataFrame + :type dut: str + :type ttype: str + :type cadence: str + :returns: Alphabeticaly sorted list of test beds. + :rtype: list + """ + return sorted(list(df.loc[( + (df["dut"] == dut) & + (df["ttype"] == ttype) & + (df["cadence"] == cadence) + )]["tbed"].unique())) + + +def get_job(df: pd.DataFrame, dut, ttype, cadence, testbed): + """Get the name of a job defined by dut, ttype, cadence, test bed. + Input information comes from the control panel. + + :param df: DataFrame with information about jobs. + :param dut: The DUT for which the job name will be created. + :param ttype: The test type for which the job name will be created. + :param cadence: The cadence for which the job name will be created. + :param testbed: The test bed for which the job name will be created. + :type df: pandas.DataFrame + :type dut: str + :type ttype: str + :type cadence: str + :type testbed: str + :returns: Job name. + :rtype: str + """ + return df.loc[( + (df["dut"] == dut) & + (df["ttype"] == ttype) & + (df["cadence"] == cadence) & + (df["tbed"] == testbed) + )]["job"].item() + + +def generate_options(opts: list, sort: bool=True) -> list: + """Return list of options for radio items in control panel. The items in + the list are dictionaries with keys "label" and "value". + + :params opts: List of options (str) to be used for the generated list. + :type opts: list + :returns: List of options (dict). + :rtype: list + """ + if sort: + opts = sorted(opts) + return [{"label": i, "value": i} for i in opts] + + +def set_job_params(df: pd.DataFrame, job: str) -> dict: + """Create a dictionary with all options and values for (and from) the + given job. + + :param df: DataFrame with information about jobs. + :params job: The name of job for and from which the dictionary will be + created. + :type df: pandas.DataFrame + :type job: str + :returns: Dictionary with all options and values for (and from) the + given job. + :rtype: dict + """ + + l_job = job.split("-") + return { + "job": job, + "dut": l_job[1], + "ttype": l_job[3], + "cadence": l_job[4], + "tbed": "-".join(l_job[-2:]), + "duts": generate_options(get_duts(df)), + "ttypes": generate_options(get_ttypes(df, l_job[1])), + "cadences": generate_options(get_cadences(df, l_job[1], l_job[3])), + "tbeds": generate_options( + get_test_beds(df, l_job[1], l_job[3], l_job[4])) + } + + +def get_list_group_items( + items: list, + type: str, + colorize: bool=True, + add_index: bool=False + ) -> list: + """Generate list of ListGroupItems with checkboxes with selected items. + + :param items: List of items to be displayed in the ListGroup. + :param type: The type part of an element ID. + :param colorize: If True, the color of labels is set, otherwise the default + color is used. + :param add_index: Add index to the list items. + :type items: list + :type type: str + :type colorize: bool + :type add_index: bool + :returns: List of ListGroupItems with checkboxes with selected items. + :rtype: list + """ + + children = list() + for i, l in enumerate(items): + idx = f"{i + 1}. " if add_index else str() + label = f"{idx}{l['id']}" if isinstance(l, dict) else f"{idx}{l}" + children.append( + dbc.ListGroupItem( + children=[ + dbc.Checkbox( + id={"type": type, "index": i}, + label=label, + value=False, + label_class_name="m-0 p-0", + label_style={ + "font-size": ".875em", + "color": get_color(i) if colorize else "#55595c" + }, + class_name="info" + ) + ], + class_name="p-0" + ) + ) + + return children + + +def relative_change_stdev(mean1, mean2, std1, std2): + """Compute relative standard deviation of change of two values. + + The "1" values are the base for comparison. + Results are returned as percentage (and percentual points for stdev). + Linearized theory is used, so results are wrong for relatively large stdev. + + :param mean1: Mean of the first number. + :param mean2: Mean of the second number. + :param std1: Standard deviation estimate of the first number. + :param std2: Standard deviation estimate of the second number. + :type mean1: float + :type mean2: float + :type std1: float + :type std2: float + :returns: Relative change and its stdev. + :rtype: float + """ + mean1, mean2 = float(mean1), float(mean2) + quotient = mean2 / mean1 + first = std1 / mean1 + second = std2 / mean2 + std = quotient * sqrt(first * first + second * second) + return (quotient - 1) * 100, std * 100 + + +def get_hdrh_latencies(row: pd.Series, name: str) -> dict: + """Get the HDRH latencies from the test data. + + :param row: A row fron the data frame with test data. + :param name: The test name to be displayed as the graph title. + :type row: pandas.Series + :type name: str + :returns: Dictionary with HDRH latencies. + :rtype: dict + """ + + latencies = {"name": name} + for key in C.LAT_HDRH: + try: + latencies[key] = row[key] + except KeyError: + return None + + return latencies + + +def graph_hdrh_latency(data: dict, layout: dict) -> go.Figure: + """Generate HDR Latency histogram graphs. + + :param data: HDRH data. + :param layout: Layout of plot.ly graph. + :type data: dict + :type layout: dict + :returns: HDR latency Histogram. + :rtype: plotly.graph_objects.Figure + """ + + fig = None + + traces = list() + for idx, (lat_name, lat_hdrh) in enumerate(data.items()): + try: + decoded = hdrh.histogram.HdrHistogram.decode(lat_hdrh) + except (hdrh.codec.HdrLengthException, TypeError): + continue + previous_x = 0.0 + prev_perc = 0.0 + xaxis = list() + yaxis = list() + hovertext = list() + for item in decoded.get_recorded_iterator(): + # The real value is "percentile". + # For 100%, we cut that down to "x_perc" to avoid + # infinity. + percentile = item.percentile_level_iterated_to + x_perc = min(percentile, C.PERCENTILE_MAX) + xaxis.append(previous_x) + yaxis.append(item.value_iterated_to) + hovertext.append( + f"<b>{C.GRAPH_LAT_HDRH_DESC[lat_name]}</b><br>" + f"Direction: {('W-E', 'E-W')[idx % 2]}<br>" + f"Percentile: {prev_perc:.5f}-{percentile:.5f}%<br>" + f"Latency: {item.value_iterated_to}uSec" + ) + next_x = 100.0 / (100.0 - x_perc) + xaxis.append(next_x) + yaxis.append(item.value_iterated_to) + hovertext.append( + f"<b>{C.GRAPH_LAT_HDRH_DESC[lat_name]}</b><br>" + f"Direction: {('W-E', 'E-W')[idx % 2]}<br>" + f"Percentile: {prev_perc:.5f}-{percentile:.5f}%<br>" + f"Latency: {item.value_iterated_to}uSec" + ) + previous_x = next_x + prev_perc = percentile + + traces.append( + go.Scatter( + x=xaxis, + y=yaxis, + name=C.GRAPH_LAT_HDRH_DESC[lat_name], + mode="lines", + legendgroup=C.GRAPH_LAT_HDRH_DESC[lat_name], + showlegend=bool(idx % 2), + line=dict( + color=get_color(int(idx/2)), + dash="solid", + width=1 if idx % 2 else 2 + ), + hovertext=hovertext, + hoverinfo="text" + ) + ) + if traces: + fig = go.Figure() + fig.add_traces(traces) + layout_hdrh = layout.get("plot-hdrh-latency", None) + if lat_hdrh: + fig.update_layout(layout_hdrh) + + return fig + + +def navbar_trending(active: tuple): + """Add nav element with navigation panel. It is placed on the top. + + :param active: Tuple of boolean values defining the active items in the + navbar. True == active + :type active: tuple + :returns: Navigation bar. + :rtype: dbc.NavbarSimple + """ + return dbc.NavbarSimple( + children=[ + dbc.NavItem(dbc.NavLink( + C.TREND_TITLE, + active=active[0], + external_link=True, + href="/trending" + )), + dbc.NavItem(dbc.NavLink( + C.NEWS_TITLE, + active=active[1], + external_link=True, + href="/news" + )), + dbc.NavItem(dbc.NavLink( + C.STATS_TITLE, + active=active[2], + external_link=True, + href="/stats" + )), + dbc.NavItem(dbc.NavLink( + C.SEARCH_TITLE, + active=active[3], + external_link=True, + href="/search" + )), + dbc.NavItem(dbc.NavLink( + "Documentation", + id="btn-documentation", + )) + ], + id="navbarsimple-main", + brand=C.BRAND, + brand_href="/", + brand_external_link=True, + class_name="p-2", + fluid=True + ) + + +def navbar_report(active: tuple): + """Add nav element with navigation panel. It is placed on the top. + + :param active: Tuple of boolean values defining the active items in the + navbar. True == active + :type active: tuple + :returns: Navigation bar. + :rtype: dbc.NavbarSimple + """ + return dbc.NavbarSimple( + id="navbarsimple-main", + children=[ + dbc.NavItem(dbc.NavLink( + C.REPORT_TITLE, + active=active[0], + external_link=True, + href="/report" + )), + dbc.NavItem(dbc.NavLink( + "Comparisons", + active=active[1], + external_link=True, + href="/comparisons" + )), + dbc.NavItem(dbc.NavLink( + "Coverage Data", + active=active[2], + external_link=True, + href="/coverage" + )), + dbc.NavItem(dbc.NavLink( + C.SEARCH_TITLE, + active=active[3], + external_link=True, + href="/search" + )), + dbc.NavItem(dbc.NavLink( + "Documentation", + id="btn-documentation", + )) + ], + brand=C.BRAND, + brand_href="/", + brand_external_link=True, + class_name="p-2", + fluid=True + ) + + +def filter_table_data( + store_table_data: list, + table_filter: str + ) -> list: + """Filter table data using user specified filter. + + :param store_table_data: Table data represented as a list of records. + :param table_filter: User specified filter. + :type store_table_data: list + :type table_filter: str + :returns: A new table created by filtering of table data represented as + a list of records. + :rtype: list + """ + + # Checks: + if not any((table_filter, store_table_data, )): + return store_table_data + + def _split_filter_part(filter_part: str) -> tuple: + """Split a part of filter into column name, operator and value. + A "part of filter" is a sting berween "&&" operator. + + :param filter_part: A part of filter. + :type filter_part: str + :returns: Column name, operator, value + :rtype: tuple[str, str, str|float] + """ + for operator_type in C.OPERATORS: + for operator in operator_type: + if operator in filter_part: + name_p, val_p = filter_part.split(operator, 1) + name = name_p[name_p.find("{") + 1 : name_p.rfind("}")] + val_p = val_p.strip() + if (val_p[0] == val_p[-1] and val_p[0] in ("'", '"', '`')): + value = val_p[1:-1].replace("\\" + val_p[0], val_p[0]) + else: + try: + value = float(val_p) + except ValueError: + value = val_p + + return name, operator_type[0].strip(), value + return (None, None, None) + + df = pd.DataFrame.from_records(store_table_data) + for filter_part in table_filter.split(" && "): + col_name, operator, filter_value = _split_filter_part(filter_part) + if operator == "contains": + df = df.loc[df[col_name].str.contains(filter_value, regex=True)] + elif operator in ("eq", "ne", "lt", "le", "gt", "ge"): + # These operators match pandas series operator method names. + df = df.loc[getattr(df[col_name], operator)(filter_value)] + elif operator == "datestartswith": + # This is a simplification of the front-end filtering logic, + # only works with complete fields in standard format. + # Currently not used in comparison tables. + df = df.loc[df[col_name].str.startswith(filter_value)] + + return df.to_dict("records") + + +def sort_table_data( + store_table_data: list, + sort_by: list + ) -> list: + """Sort table data using user specified order. + + :param store_table_data: Table data represented as a list of records. + :param sort_by: User specified sorting order (multicolumn). + :type store_table_data: list + :type sort_by: list + :returns: A new table created by sorting the table data represented as + a list of records. + :rtype: list + """ + + # Checks: + if not any((sort_by, store_table_data, )): + return store_table_data + + df = pd.DataFrame.from_records(store_table_data) + if len(sort_by): + dff = df.sort_values( + [col["column_id"] for col in sort_by], + ascending=[col["direction"] == "asc" for col in sort_by], + inplace=False + ) + else: + # No sort is applied + dff = df + + return dff.to_dict("records") + + +def show_trending_graph_data( + trigger: Trigger, + data: dict, + graph_layout: dict + ) -> tuple: + """Generates the data for the offcanvas displayed when a particular point in + a trending graph (daily data) is clicked on. + + :param trigger: The information from trigger when the data point is clicked + on. + :param graph: The data from the clicked point in the graph. + :param graph_layout: The layout of the HDRH latency graph. + :type trigger: Trigger + :type graph: dict + :type graph_layout: dict + :returns: The data to be displayed on the offcanvas and the information to + show the offcanvas. + :rtype: tuple(list, list, bool) + """ + + if trigger.idx == "tput": + idx = 0 + elif trigger.idx == "bandwidth": + idx = 1 + elif trigger.idx == "lat": + idx = len(data) - 1 + else: + return list(), list(), False + try: + data = data[idx]["points"][0] + except (IndexError, KeyError, ValueError, TypeError): + return list(), list(), False + + metadata = no_update + graph = list() + + list_group_items = list() + for itm in data.get("text", None).split("<br>"): + if not itm: + continue + lst_itm = itm.split(": ") + if lst_itm[0] == "csit-ref": + list_group_item = dbc.ListGroupItem([ + dbc.Badge(lst_itm[0]), + html.A( + lst_itm[1], + href=f"{C.URL_LOGS}{lst_itm[1]}", + target="_blank" + ) + ]) + else: + list_group_item = dbc.ListGroupItem([ + dbc.Badge(lst_itm[0]), + lst_itm[1] + ]) + list_group_items.append(list_group_item) + + if trigger.idx == "tput": + title = "Throughput" + elif trigger.idx == "bandwidth": + title = "Bandwidth" + elif trigger.idx == "lat": + title = "Latency" + hdrh_data = data.get("customdata", None) + if hdrh_data: + graph = [dbc.Card( + class_name="gy-2 p-0", + children=[ + dbc.CardHeader(hdrh_data.pop("name")), + dbc.CardBody( + dcc.Graph( + id="hdrh-latency-graph", + figure=graph_hdrh_latency(hdrh_data, graph_layout) + ) + ) + ]) + ] + + metadata = [ + dbc.Card( + class_name="gy-2 p-0", + children=[ + dbc.CardHeader(children=[ + dcc.Clipboard( + target_id="tput-lat-metadata", + title="Copy", + style={"display": "inline-block"} + ), + title + ]), + dbc.CardBody( + dbc.ListGroup(list_group_items, flush=True), + id="tput-lat-metadata", + class_name="p-0", + ) + ] + ) + ] + + return metadata, graph, True + + +def show_iterative_graph_data( + trigger: Trigger, + data: dict, + graph_layout: dict + ) -> tuple: + """Generates the data for the offcanvas displayed when a particular point in + a box graph (iterative data) is clicked on. + + :param trigger: The information from trigger when the data point is clicked + on. + :param graph: The data from the clicked point in the graph. + :param graph_layout: The layout of the HDRH latency graph. + :type trigger: Trigger + :type graph: dict + :type graph_layout: dict + :returns: The data to be displayed on the offcanvas and the information to + show the offcanvas. + :rtype: tuple(list, list, bool) + """ + + if trigger.idx == "tput": + idx = 0 + elif trigger.idx == "bandwidth": + idx = 1 + elif trigger.idx == "lat": + idx = len(data) - 1 + else: + return list(), list(), False + + try: + data = data[idx]["points"] + except (IndexError, KeyError, ValueError, TypeError): + return list(), list(), False + + def _process_stats(data: list, param: str) -> list: + """Process statistical data provided by plot.ly box graph. + + :param data: Statistical data provided by plot.ly box graph. + :param param: Parameter saying if the data come from "tput" or + "lat" graph. + :type data: list + :type param: str + :returns: Listo of tuples where the first value is the + statistic's name and the secont one it's value. + :rtype: list + """ + if len(data) == 7: + stats = ("max", "upper fence", "q3", "median", "q1", + "lower fence", "min") + elif len(data) == 9: + stats = ("outlier", "max", "upper fence", "q3", "median", + "q1", "lower fence", "min", "outlier") + elif len(data) == 1: + if param == "lat": + stats = ("average latency at 50% PDR", ) + elif param == "bandwidth": + stats = ("bandwidth", ) + else: + stats = ("throughput", ) + else: + return list() + unit = " [us]" if param == "lat" else str() + return [(f"{stat}{unit}", f"{value['y']:,.0f}") + for stat, value in zip(stats, data)] + + customdata = data[0].get("customdata", dict()) + datapoint = customdata.get("metadata", dict()) + hdrh_data = customdata.get("hdrh", dict()) + + list_group_items = list() + for k, v in datapoint.items(): + if k == "csit-ref": + if len(data) > 1: + continue + list_group_item = dbc.ListGroupItem([ + dbc.Badge(k), + html.A(v, href=f"{C.URL_LOGS}{v}", target="_blank") + ]) + else: + list_group_item = dbc.ListGroupItem([dbc.Badge(k), v]) + list_group_items.append(list_group_item) + + graph = list() + if trigger.idx == "tput": + title = "Throughput" + elif trigger.idx == "bandwidth": + title = "Bandwidth" + elif trigger.idx == "lat": + title = "Latency" + if len(data) == 1: + if hdrh_data: + graph = [dbc.Card( + class_name="gy-2 p-0", + children=[ + dbc.CardHeader(hdrh_data.pop("name")), + dbc.CardBody(dcc.Graph( + id="hdrh-latency-graph", + figure=graph_hdrh_latency(hdrh_data, graph_layout) + )) + ]) + ] + + for k, v in _process_stats(data, trigger.idx): + list_group_items.append(dbc.ListGroupItem([dbc.Badge(k), v])) + + metadata = [ + dbc.Card( + class_name="gy-2 p-0", + children=[ + dbc.CardHeader(children=[ + dcc.Clipboard( + target_id="tput-lat-metadata", + title="Copy", + style={"display": "inline-block"} + ), + title + ]), + dbc.CardBody( + dbc.ListGroup(list_group_items, flush=True), + id="tput-lat-metadata", + class_name="p-0" + ) + ] + ) + ] + + return metadata, graph, True |