📌 Feature Summary
Write better benchmarking scripts (output destination and names)
💡 Motivation
Currently, our benchmarking scripts save generated CSV files and graphs directly in the same directory as the script. This clutters the folder and makes it harder to organize results, especially when running multiple benchmarks. File names are also inconsistent, making it difficult to distinguish between runs.
🧩 Suggested Implementation
- Add a configurable
output_dir parameter (default: benchmarks/outputs/).
- Create the directory if it does not exist.
- Define a naming convention, e.g.
{script_name}{date}{time}.{ext}
- Apply the same convention for graph images and other generated files.
🧠 Estimated Complexity
Medium (Changes multiple files, some model updates)
🔗 Related Issues / PRs
No response
🧠 Additional Notes
- Add appropriate logging where required.
- Consider adding an optional
--tag CLI argument to append a custom identifier to output filenames for easier tracking of experimental runs.
📌 Feature Summary
Write better benchmarking scripts (output destination and names)
💡 Motivation
Currently, our benchmarking scripts save generated CSV files and graphs directly in the same directory as the script. This clutters the folder and makes it harder to organize results, especially when running multiple benchmarks. File names are also inconsistent, making it difficult to distinguish between runs.
🧩 Suggested Implementation
output_dirparameter (default:benchmarks/outputs/).{script_name}{date}{time}.{ext}🧠 Estimated Complexity
Medium (Changes multiple files, some model updates)
🔗 Related Issues / PRs
No response
🧠 Additional Notes
--tagCLI argument to append a custom identifier to output filenames for easier tracking of experimental runs.