From 6e1a15acc7bd3685d71f9d34238181c681dcd4c0 Mon Sep 17 00:00:00 2001 From: Tibor Frank Date: Wed, 25 Apr 2018 15:28:14 +0200 Subject: CSIT-1041: Trending dashboard Change-Id: I1d6aae1839a9d6d44407c90ff257bc37495f0cfa Signed-off-by: Tibor Frank --- resources/tools/presentation/generator_tables.py | 205 +++++++---------------- 1 file changed, 63 insertions(+), 142 deletions(-) (limited to 'resources/tools/presentation') diff --git a/resources/tools/presentation/generator_tables.py b/resources/tools/presentation/generator_tables.py index c41c6de004..c2007a1a49 100644 --- a/resources/tools/presentation/generator_tables.py +++ b/resources/tools/presentation/generator_tables.py @@ -22,6 +22,7 @@ import pandas as pd from string import replace from math import isnan +from numpy import nan from xml.etree import ElementTree as ET from errors import PresentationError @@ -688,12 +689,12 @@ def table_performance_trending_dashboard(table, input_data): # Prepare the header of the tables header = ["Test Case", - "Throughput Trend [Mpps]", - "Long Trend Compliance", - "Trend Compliance", - "Top Anomaly [Mpps]", - "Change [%]", - "Outliers [Number]" + "Trend [Mpps]", + "Short-Term Change [%]", + "Long-Term Change [%]", + "Regressions [#]", + "Progressions [#]", + "Outliers [#]" ] header_str = ",".join(header) + "\n" @@ -719,154 +720,81 @@ def table_performance_trending_dashboard(table, input_data): if len(tbl_dict[tst_name]["data"]) > 2: pd_data = pd.Series(tbl_dict[tst_name]["data"]) + last_key = pd_data.keys()[-1] win_size = min(pd_data.size, table["window"]) - # Test name: - name = tbl_dict[tst_name]["name"] + key_14 = pd_data.keys()[-(pd_data.size - win_size)] + long_win_size = min(pd_data.size, table["long-trend-window"]) + + data_t, _ = split_outliers(pd_data, outlier_const=1.5, + window=win_size) - median = pd_data.rolling(window=win_size, min_periods=2).median() - median_idx = pd_data.size - table["long-trend-window"] - median_idx = 0 if median_idx < 0 else median_idx + median_t = data_t.rolling(window=win_size, min_periods=2).median() + stdev_t = data_t.rolling(window=win_size, min_periods=2).std() + median_idx = pd_data.size - long_win_size try: - max_median = max([x for x in median.values[median_idx:] + max_median = max([x for x in median_t.values[median_idx:] if not isnan(x)]) except ValueError: - max_median = None - trimmed_data, _ = split_outliers(pd_data, outlier_const=1.5, - window=win_size) - stdev_t = pd_data.rolling(window=win_size, min_periods=2).std() - - rel_change_lst = [None, ] - classification_lst = [None, ] - median_lst = [None, ] - sample_lst = [None, ] - first = True + max_median = nan + try: + last_median_t = median_t[last_key] + except KeyError: + last_median_t = nan + try: + median_t_14 = median_t[key_14] + except KeyError: + median_t_14 = nan + + # Test name: + name = tbl_dict[tst_name]["name"] + + # Classification list: + classification_lst = list() for build_nr, value in pd_data.iteritems(): - if first: - first = False - continue - # Relative changes list: - if not isnan(value) \ - and not isnan(median[build_nr]) \ - and median[build_nr] != 0: - rel_change_lst.append(round( - relative_change(float(median[build_nr]), float(value)), - 2)) - else: - rel_change_lst.append(None) - # Classification list: - if isnan(trimmed_data[build_nr]) \ - or isnan(median[build_nr]) \ + if isnan(data_t[build_nr]) \ + or isnan(median_t[build_nr]) \ or isnan(stdev_t[build_nr]) \ or isnan(value): classification_lst.append("outlier") - elif value < (median[build_nr] - 3 * stdev_t[build_nr]): + elif value < (median_t[build_nr] - 3 * stdev_t[build_nr]): classification_lst.append("regression") - elif value > (median[build_nr] + 3 * stdev_t[build_nr]): + elif value > (median_t[build_nr] + 3 * stdev_t[build_nr]): classification_lst.append("progression") else: classification_lst.append("normal") - sample_lst.append(value) - median_lst.append(median[build_nr]) - - last_idx = len(classification_lst) - 1 - first_idx = last_idx - int(table["evaluated-window"]) - if first_idx < 0: - first_idx = 0 - - nr_outliers = 0 - consecutive_outliers = 0 - failure = False - for item in classification_lst[first_idx:]: - if item == "outlier": - nr_outliers += 1 - consecutive_outliers += 1 - if consecutive_outliers == 3: - failure = True - else: - consecutive_outliers = 0 - - if failure: - classification = "failure" - elif "regression" in classification_lst[first_idx:]: - classification = "regression" - elif "progression" in classification_lst[first_idx:]: - classification = "progression" + + if isnan(last_median_t) or isnan(median_t_14) or median_t_14 == 0: + rel_change_last = nan else: - classification = "normal" + rel_change_last = round( + (last_median_t - median_t_14) / median_t_14, 2) - if classification == "normal": - index = len(classification_lst) - 1 + if isnan(max_median) or isnan(last_median_t) or max_median == 0: + rel_change_long = nan else: - tmp_classification = "outlier" if classification == "failure" \ - else classification - index = None - for idx in range(first_idx, len(classification_lst)): - if classification_lst[idx] == tmp_classification: - if rel_change_lst[idx]: - index = idx - break - if index is None: - continue - for idx in range(index+1, len(classification_lst)): - if classification_lst[idx] == tmp_classification: - if rel_change_lst[idx]: - if (abs(rel_change_lst[idx]) > - abs(rel_change_lst[index])): - index = idx - - logging.debug("{}".format(name)) - logging.debug("sample_lst: {} - {}". - format(len(sample_lst), sample_lst)) - logging.debug("median_lst: {} - {}". - format(len(median_lst), median_lst)) - logging.debug("rel_change: {} - {}". - format(len(rel_change_lst), rel_change_lst)) - logging.debug("classn_lst: {} - {}". - format(len(classification_lst), classification_lst)) - logging.debug("index: {}".format(index)) - logging.debug("classifica: {}".format(classification)) + rel_change_long = round( + (last_median_t - max_median) / max_median, 2) + + tbl_lst.append([name, + '-' if isnan(last_median_t) else + round(last_median_t / 1000000, 2), + '-' if isnan(rel_change_last) else rel_change_last, + '-' if isnan(rel_change_long) else rel_change_long, + classification_lst[win_size:].count("regression"), + classification_lst[win_size:].count("progression"), + classification_lst[win_size:].count("outlier")]) + + tbl_lst.sort(key=lambda rel: rel[0]) - try: - trend = round(float(median_lst[-1]) / 1000000, 2) \ - if not isnan(median_lst[-1]) else '-' - sample = round(float(sample_lst[index]) / 1000000, 2) \ - if not isnan(sample_lst[index]) else '-' - rel_change = rel_change_lst[index] \ - if rel_change_lst[index] is not None else '-' - if max_median is not None: - if not isnan(sample_lst[index]): - long_trend_threshold = \ - max_median * (table["long-trend-threshold"] / 100) - if sample_lst[index] < long_trend_threshold: - long_trend_classification = "failure" - else: - long_trend_classification = 'normal' - else: - long_trend_classification = "failure" - else: - long_trend_classification = '-' - tbl_lst.append([name, - trend, - long_trend_classification, - classification, - '-' if classification == "normal" else sample, - '-' if classification == "normal" else - rel_change, - nr_outliers]) - except IndexError as err: - logging.error("{}".format(err)) - continue - - # Sort the table according to the classification tbl_sorted = list() - for long_trend_class in ("failure", 'normal', '-'): - tbl_long = [item for item in tbl_lst if item[2] == long_trend_class] - for classification in \ - ("failure", "regression", "progression", "normal"): - tbl_tmp = [item for item in tbl_long if item[3] == classification] - tbl_tmp.sort(key=lambda rel: rel[0]) - tbl_sorted.extend(tbl_tmp) + for nrr in range(table["window"], -1, -1): + tbl_reg = [item for item in tbl_lst if item[4] == nrr] + for nrp in range(table["window"], -1, -1): + tbl_pro = [item for item in tbl_reg if item[5] == nrp] + for nro in range(table["window"], -1, -1): + tbl_out = [item for item in tbl_pro if item[5] == nro] + tbl_sorted.extend(tbl_out) file_name = "{0}{1}".format(table["output-file"], table["output-file-ext"]) @@ -1002,13 +930,6 @@ def table_performance_trending_dashboard_html(table, input_data): ref = ET.SubElement(td, "a", attrib=dict(href=url)) ref.text = item - if c_idx == 3: - if item == "regression": - td.set("bgcolor", "#eca1a6") - elif item == "failure": - td.set("bgcolor", "#d6cbd3") - elif item == "progression": - td.set("bgcolor", "#bdcebe") if c_idx > 0: td.text = item -- cgit 1.2.3-korg