IGUANA does output its test results as RDF (e.g. result.nt), this format is not particularly easy to manipulate for a data analyst.
We should write a small tool that based on a IGUANA test output is able to generate a tabular dataset with the following columns:
dataset_name | query_id | endpoint | total | success | errors | timeouts | totalTime | QPS | ...(more to add) |
random_sample_1 | 1 | split1 | 123 | 122 | 0 | 1 | 2837239 | 12 | ... |
random_sample_1 | 1 | split2 | 124 | 124 | 0 | 0 | 2837283 | 13 | ... |
We might consider loading the result into a sparql engine like jena and use a SPARQL query to extract all this.
AC:
- Decide and document the set of metrics we want to extract from IGUANA
- A standard IGUANA suite yaml file is created (will be used as a reference to create the actual test suite config)
- A tool is available to convert the result file into a CSV file