Add more metadata to the json report

[ev-br]

Looking at https://github.com/ev-br/numpy/blob/main/.github/workflows/array-api-report.yml, there's a bit of metadata which needs to be manually added to the JSON report from the test suite:

          library_version="$(python -c "import importlib.metadata; print(importlib.metadata.version('${{ env.ARRAY_LIBRARY }}'))")"
          platform="$(python -c 'import platform; print(platform.platform())')"
          system="$(python -c 'import platform; print(platform.system())')"
          release="$(python -c 'import platform; print(platform.release())')"
          machine="$(python -c 'import platform; print(platform.machine())')"
          python_version="$(python -c 'import platform; print(platform.python_version())')"
          api_version="$(python -c 'import ${{ env.ARRAY_API_TESTS_MODULE }}; print(${{ env.ARRAY_API_TESTS_MODULE }}.__array_api_version__)')"
          test_suite="$(git rev-parse HEAD)"
          jq \
            --arg schema 'v1' \
            --arg name '${{ env.ARRAY_LIBRARY }}-fork' \
            --arg version "${library_version}" \
            --arg api_version "${api_version}" \
            --arg timestamp "$(date -u +'%Y-%m-%dT%H:%M:%SZ')" \
            --arg platform "${platform}" \
            --arg system "${system}" \
            --arg release "${release}" \
            --arg machine "${machine}" \
            --arg execution_target 'cpu' \
            --arg python_version "${python_version}" \
            --arg test_suite "${test_suite}" \

Some of this data we either have in the test suite output itself or can add fairly easily:

• name : this will have to stay in the workflow YML (because -fork)
• version, api version: can add
• timestamp: .report.json has a created field
• platform, system, release, machine, python_version: can add

The question is whether it's worth adding since you already did the work to set up the current version @kgryte

[kgryte]

If the test suite is able to customize the output of --json-report, I see no issue with moving some of the meta data into the raw report for later extraction. It won't change the schema. It will just change how the raw report is transformed into the report format expected by the harvester.