aboutsummaryrefslogtreecommitdiffstats
path: root/resources/tools
diff options
context:
space:
mode:
Diffstat (limited to 'resources/tools')
-rwxr-xr-xresources/tools/report_gen/run_improvments_tables.py212
-rw-r--r--resources/tools/report_gen/run_report.cfg1
-rwxr-xr-xresources/tools/report_gen/run_report.sh14
3 files changed, 227 insertions, 0 deletions
diff --git a/resources/tools/report_gen/run_improvments_tables.py b/resources/tools/report_gen/run_improvments_tables.py
new file mode 100755
index 0000000000..7bd6775a87
--- /dev/null
+++ b/resources/tools/report_gen/run_improvments_tables.py
@@ -0,0 +1,212 @@
+#!/usr/bin/python
+
+# Copyright (c) 2017 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.
+
+"""Generate csv files for the chapter "CSIT Release Notes" from json files
+generated by Jenkins' jobs.
+"""
+
+from sys import exit as sys_exit
+from os import walk
+from os.path import join
+from math import sqrt
+from argparse import ArgumentParser, RawDescriptionHelpFormatter
+from json import load
+
+
+EXT_JSON = ".json"
+EXT_TMPL = ".template"
+EXT_CSV = ".csv"
+
+
+def get_files(path, extension):
+ """Generates the list of files to process.
+
+ :param path: Path to files.
+ :param extension: Extension of files to process. If it is the empty string,
+ all files will be processed.
+ :type path: str
+ :type extension: str
+ :returns: List of files to process.
+ :rtype: list
+ """
+
+ file_list = list()
+ for root, _, files in walk(path):
+ for filename in files:
+ if extension:
+ if filename.endswith(extension):
+ file_list.append(join(root, filename))
+ else:
+ file_list.append(join(root, filename))
+
+ return file_list
+
+
+def parse_args():
+ """Parse arguments from cmd line.
+
+ :returns: Parsed arguments.
+ :rtype ArgumentParser
+ """
+
+ parser = ArgumentParser(description=__doc__,
+ formatter_class=RawDescriptionHelpFormatter)
+ parser.add_argument("-i", "--input",
+ required=True,
+ help="Input folder with data files.")
+ parser.add_argument("-o", "--output",
+ required=True,
+ help="Output folder with csv files and templates for "
+ "csv files.")
+ return parser.parse_args()
+
+
+def calculate_stats(data):
+ """Calculate statistics:
+ - average,
+ - standard deviation.
+
+ :param data: Data to process.
+ :type data: list
+ :returns: Average and standard deviation.
+ :rtype: tuple
+ """
+
+ if len(data) == 0:
+ return None, None
+
+ def average(items):
+ """Calculate average from the items.
+
+ :param items: Average is calculated from these items.
+ :type items: list
+ :returns: Average.
+ :rtype: float
+ """
+ return float(sum(items)) / len(items)
+
+ avg = average(data)
+ variance = [(x - avg) ** 2 for x in data]
+ stdev = sqrt(average(variance))
+
+ return avg, stdev
+
+
+def write_line_to_file(file_handler, item):
+ """Write a line to the csv file.
+
+ :param file_handler: File handler for the csv file. It must be open for
+ writing text.
+ :param item: Item to be written to the file.
+ :type file_handler: BinaryIO
+ :type item: dict
+ """
+
+ mean = "" if item["mean"] is None else "{:.1f}".format(item["mean"])
+ stdev = "" if item["stdev"] is None else "{:.1f}".format(item["stdev"])
+ change = "" if item["change"] is None else "{:.0f}%".format(item["change"])
+ file_handler.write("{},{},{},{}\n".format(item["old"], mean, stdev, change))
+
+
+def main():
+ """Main function to generate csv files for the chapter "CSIT Release Notes"
+ from json files generated by Jenkins' jobs.
+ """
+
+ args = parse_args()
+
+ json_files = get_files(args.input, EXT_JSON)
+ tmpl_files = get_files(args.output, EXT_TMPL)
+
+ if len(json_files) == 0:
+ print("No json data to process.")
+ exit(1)
+
+ if len(tmpl_files) == 0:
+ print("No template files to process.")
+ exit(1)
+
+ # Get information from template files
+ csv_data = list()
+ for tmpl_file in tmpl_files:
+ with open(tmpl_file, mode='r') as file_handler:
+ for line in file_handler:
+ line_list = line.split(',')
+ try:
+ csv_data.append({
+ "ID": line_list[0],
+ "type": line_list[0].rsplit("-", 1)[-1],
+ "old": ",".join(line_list[1:])[:-1],
+ "last_old": line_list[-1][:-1],
+ "rates": list(),
+ "mean": None,
+ "stdev": None,
+ "change": None})
+ except IndexError:
+ pass
+
+ # Update existing data with the new information from json files
+ for json_file in json_files:
+ with open(json_file) as file_handler:
+ tests_data = load(file_handler)
+ for item in csv_data:
+ try:
+ rate = tests_data["data"][item["ID"]]["throughput"]["value"]
+ item["rates"].append(rate)
+ except KeyError:
+ pass
+
+ # Add statistics
+ for item in csv_data:
+ mean, stdev = calculate_stats(item["rates"])
+ if mean is not None:
+ mean = mean / 1000000
+ old = float(item["last_old"])
+ item["mean"] = mean
+ item["change"] = ((mean - old) / old) * 100
+ item["stdev"] = stdev / 1000000
+
+ # Sort the list, key = change
+ csv_data.sort(key=lambda data: data["change"], reverse=True)
+
+ # Write csv files
+ for tmpl_file in tmpl_files:
+ csv_file = tmpl_file.replace(EXT_TMPL, EXT_CSV)
+ with open(csv_file, "w") as file_handler:
+ for item in csv_data:
+ if "pdr_" in csv_file \
+ and "_others" not in csv_file \
+ and item["type"] == "pdrdisc" \
+ and item["change"] >= 9.5:
+ write_line_to_file(file_handler, item)
+ elif "pdr_" in csv_file \
+ and "_others" in csv_file \
+ and item["type"] == "pdrdisc" \
+ and item["change"] < 9.5:
+ write_line_to_file(file_handler, item)
+ elif "ndr_" in csv_file \
+ and "_others" not in csv_file \
+ and item["type"] == "ndrdisc" \
+ and item["change"] >= 9.5:
+ write_line_to_file(file_handler, item)
+ elif "ndr_" in csv_file \
+ and "_others" in csv_file \
+ and item["type"] == "ndrdisc" \
+ and item["change"] < 9.5:
+ write_line_to_file(file_handler, item)
+
+
+if __name__ == "__main__":
+ sys_exit(main())
diff --git a/resources/tools/report_gen/run_report.cfg b/resources/tools/report_gen/run_report.cfg
index 70eb411fe9..642a3e3435 100644
--- a/resources/tools/report_gen/run_report.cfg
+++ b/resources/tools/report_gen/run_report.cfg
@@ -28,6 +28,7 @@ DIR[DTR,PERF,HC]=${DIR[DTR]}/honeycomb_performance_results
DIR[DTR,FUNC,VPP]=${DIR[DTR]}/vpp_functional_results
DIR[DTR,FUNC,HC]=${DIR[DTR]}/honeycomb_functional_results
DIR[DTR,FUNC,NSHSFC]=${DIR[DTR]}/nshsfc_functional_results
+DIR[DTR,PERF,VPP,IMPRV]=${DIR[RST]}/vpp_performance_tests/performance_improvements
DIR[DTC]=${DIR[RST]}/test_configuration
DIR[DTC,PERF,VPP]=${DIR[DTC]}/vpp_performance_configuration
diff --git a/resources/tools/report_gen/run_report.sh b/resources/tools/report_gen/run_report.sh
index 4c104928f3..db079773c8 100755
--- a/resources/tools/report_gen/run_report.sh
+++ b/resources/tools/report_gen/run_report.sh
@@ -143,6 +143,15 @@ then
--output ${DIR[DTO,PERF,VPP]}/vpp_performance_operational_data.rst \
--data "SH_RUN" --formatting rst --start 4 --level 2
+ blds=${JOB[PERF,VPP,BLD]}
+ for i in ${blds[@]}; do
+ unzip -o ${DIR[STATIC,ARCH]}/${JOB[PERF,VPP]}-${i}.zip -d ${DIR[WORKING]}/
+ ./run_robot_json_data.py \
+ --input ${DIR[WORKING]}/output.xml \
+ --output ${DIR[DTR,PERF,VPP,IMPRV]}/${JOB[PERF,VPP]}-${i}.json \
+ --vdevice ${i}
+ done
+
# DPDK PERF
unzip -o ${DIR[STATIC,ARCH]}/${JOB[PERF,DPDK]}-${JOB[PERF,DPDK,FBLD]}.zip -d ${DIR[WORKING]}/
python run_robot_data.py -i ${DIR[WORKING]}/robot-plugin/output.xml \
@@ -171,6 +180,11 @@ then
--formatting rst --start 5 --level 2
fi
+# Generate tables for performance improvements
+./run_improvments_tables.py \
+ --input ${DIR[DTR,PERF,VPP,IMPRV]} \
+ --output ${DIR[DTR,PERF,VPP,IMPRV]}
+
# Delete temporary json files
find ${DIR[RST]} -name "*.json" -type f -delete