dawidniezgodka/evaluate benchmark

Evaluate benchmark results

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
folder_with_bench_dataThe file serves as the repository for both historical benchmark data and the location where new benchmark data will be appended. Currently, only .js files are supported. nobench_data
file_with_bench_dataThe name of the file that contains the benchmark data. Currently, only .js files are supported. nodata.json
folder_with_current_benchmark_resultsA path to the folder where benchmark results are stored. Can be a single file or multiple files. yes
bench_group_nameThe name of a specific benchmark group. This name will be used to identify the benchmark ...noBenchmark
result_files_merge_strategy_for_each_metricThe strategy to use when merging the the benchmark results from the downloaded folder. This should be a comma separated list of strategies. Each strategy corresponds to a metric. The possible values are: sum, average, max, min, median. no
metrics_to_evaluateA comma-separated list of metrics to evaluate for the current benchmark. This must be a subset of metrics from the file pointed to by current_bench_res_file. If not specified, all metrics will be evaluated. no
github_tokenGitHub token to pull/push to repo, add benchmark info to the repo, etc.no
comment_to_commitAdd commit comment with the detailed benchmark info. Possible values are on, off, if_failed. noif_failed
alert_users_if_bench_failedCommas-separated list of users to alert if the benchmark failed. no
action_page_job_summaryLeave a short summary on the workflow page. Possible values are on, off, if_failednoon
save_curr_bench_resSave the benchmark data pointed to by current_bench_res_file to a file located in benchmark_data_file' notrue
evaluation_methodIt refers to a reference point by which we determine whether the benchmark is successful or not. The possible values are 'previous', 'previous_successful', 'threshold', 'jump_detection', 'trend_detection_moving_ave', 'trend_detection_deltas' yes
benchmark_group_to_compareThe name of the benchmark to compare to. This input is only used when the reference is set to 'previous' or 'previous_successful.' If not specified, the benchmark with the same name as the current benchmark is used. no
threshold_valuesComma-separated list of threshold values for comparison.no
comparison_operatorsComma-separated list of operators for threshold comparison.no
comparison_marginsDescribes by how much percentage a given value can deviate from the desired value (threshold or prev) to still qualify the benchmark as successful. If you have selected range as one of the modes in the above set of modes, and it is not the only function, please provide values of -1 for the remaining functions. For example, if you previously entered: smaller, smaller, range, and your range is 20%, this input should look as follows: "-1, -1, 20"
threshold_upperComma-separated list of upper threshold values for range comparison.no
threshold_lowerComma-separated list of lower threshold values for range comparison.no
jump_detection_thresholdsThe threshold values for jump detection.no
trend_thresholdsThe threshold value for both deltas trend detection and moving average.no
moving_ave_window_sizeThe window size for moving average trend detection.no
trend_det_no_sufficient_data_strategyThe strategy to use when there is not enough data to perform trend detection. The possible values are use_available and fail. nouse_available
trend_det_successful_release_branchThe name of the branch that contains the successful benchmark result in a release branch. If not specified, main branch will be used. nomain
failing_conditionSpecifies when the benchmark should fail. The possible values are any, all, and none. noany
link_to_templated_gh_page_with_resultsThe link to the GitHub page with the benchmark results. The link should be templated with the benchmark name. For example, if the benchmark name is 'Benchmark', the link should look as follows: https://dawidniezgodka.github.io/Bench1-Final/ no
github_contextThe GitHub context. This is used to get the current branch name. If not specified, the default GitHub context is used. no
namedescription
should_failIndicates whether the benchmark should fail or not. The possible values are true and false.