chore: unit test different power models and get results

[rootfs]

Checklist for PR Author

---

In addition to approval, the author must confirm the following check list:

• Run the following command to format your code:

  make fmt

• Create issues for unresolved comments and link them to this PR. Use one of the following labels:

  • must-fix: The logic appears incorrect and must be addressed.
  • minor: Typos, minor issues, or potential refactoring for better readability.
  • nit: Trivial issues like extra spaces, commas, etc.

[rootfs]

@sunya-ch can you take a look at this error? I think the model server doesn't have predict for log regressor

[sunya-ch]

@rootfs
It depends on which model db is connected to the estimator. If we want to test all available models (trainers), we must first make sure that the model exists at the estimator.
For curvefit-class models, it is possible that it cannot find fitting model parameters regarding the data input.
I define SGDRegressor here since it is the simplest trainer that should always exist.

If you run estimator standalone, check the configuration file.
If you run estimator with model server, check the db that connected to the model-server.

[rootfs]

indeed! with a more realistic feature values, the predicator now works

tests/estimator_power_request_test.py::test_estimator_power_request {"trainer_name": "SGDRegressorTrainer_0", "metrics": ["bpf_cpu_time_ms", "bpf_page_cache_hit"], "system_features": ["node_info", "cpu_scaling_frequency_hertz"], "system_values": ["1", "1GHz"], "values": [[1000.0, 1000.0], [1000.0, 1000.0]], "output_type": "AbsPower", "source": "acpi"}
{'powers': {'platform': [79.01245846529595, 79.01245846529595]}, 'msg': '', 'core_ratio': 4.0}
{"trainer_name": "LogarithmicRegressionTrainer_0.json", "metrics": ["bpf_cpu_time_ms", "bpf_page_cache_hit"], "system_features": ["node_info", "cpu_scaling_frequency_hertz"], "system_values": ["1", "1GHz"], "values": [[1000.0, 1000.0], [1000.0, 1000.0]], "output_type": "AbsPower", "source": "acpi"}
{'powers': {'platform': [76.55652313028382, 76.55652313028382]}, 'msg': '', 'core_ratio': 1}
{"trainer_name": "SGDRegressorTrainer_0", "metrics": ["bpf_cpu_time_ms", "bpf_page_cache_hit"], "system_features": ["node_info", "cpu_scaling_frequency_hertz"], "system_values": ["1", "1GHz"], "values": [[1000.0, 1000.0], [1000.0, 1000.0]], "output_type": "AbsPower", "source": "rapl-sysfs"}
{'powers': {'package': [146.99476788632222, 146.99476788632222], 'core': [0.0, 0.0], 'uncore': [0.0, 0.0], 'dram': [18.704254540352803, 18.704254540352803]}, 'msg': '', 'core_ratio': 0.16666666666666666}
{"trainer_name": "LogarithmicRegressionTrainer_0.json", "metrics": ["bpf_cpu_time_ms", "bpf_page_cache_hit"], "system_features": ["node_info", "cpu_scaling_frequency_hertz"], "system_values": ["1", "1GHz"], "values": [[1000.0, 1000.0], [1000.0, 1000.0]], "output_type": "AbsPower", "source": "rapl-sysfs"}
{'powers': {'package': [59.31611863686947, 59.31611863686947], 'core': [0.0, 0.0], 'uncore': [0.0, 0.0], 'dram': [14.265734715496643, 14.265734715496643]}, 'msg': '', 'core_ratio': 1}

[sunya-ch]

@rootfs May I have an update of this PR?