Environment

Tracarbon version: latest from pip

Environment:

• OS: kubuntu

Bug

What happened:

File "/usr/local/lib/python3.10/dist-packages/tracarbon/hardwares/sensors.py", line 63, in from_platform
tracarbon.exceptions.TracarbonException: This platform Linux is not yet implemented.

How to reproduce it:

Launch tracarbon on linux (or windows).

More details:

The (simple) fix is to add a return here and here (hardwares/sensors.py, lines 61 & 63).

Description

Python modules are not necessarily in path, so we may want to use python3 -m tracarbon run instead of tracarbon run.

This led to an error.

The actual way of running tracarbon with Python where installed modules are not on path is python3 -m tracarbon.cli.__init__.

The correct way of running a Python package is python3 -m [module_name] (see here an example on "why it's better" with pip).

If we create a __main__.py file at the root of the package, then Python will call it by default using only the name of the package.

Here's a small example:

$ mkdir a && nano a/__main__.py
$ cat a/__main__.py 
if __name__ == "__main__":
    print("Main!")
$ python3 a
Main!

Hello, thank you for providing such an excellent project and code. I don’t understand something and would like to ask you for advice. After configuring the api, I ran the command tracarbon run in the terminal but an error was displayed. I would like to ask you for advice on which one should be run. Does the program or what other parts of the program need to be changed before it can run?
error：
ValueError: could not convert string to float: ''

Dear Florian,
Would it possible to add functions that print at the total of energy consumed and co2 generated from the entire training phase of the algorithm? Furthermore, can you also add one that print the carbon intensity derived from the CO2 signal API?
Thank you,
Best Regards,
Max

Description

• Tracks the power consumption from a Linux device

Use Cases

• Reading RAPL energy measurements from Linux

Description

Create a file exporter for Tracarbon.

Use Case
When Tracarbon is launched, it should be able to write everything to local files.

Description

• Tracks the power consumption from a Windows device

Does tracarbon work in a cloud infrastructure? (private cloud) to measure my application's power consumption in my container?

I think the readme or the doc website is missing some explanation regarding the output of the program.

I think that tracarbon.hardware_memory_usage and tracarbon.hardware_cpu_usage are expressed in percentage?

What about tracarbon.co2_emission? What is the unit? Same for tracarbon.energy_consumption.

How to read the json file (or output to stdout)? Do we need to create a graph in order to see the consumption? Do we need to calculate a median or an average value?

Hello,
Is it possible for Tracarbon to track only my application's power consumption and carbon footprint or does it always measure the whole machine's power?
What if I have a container? Does that mean that if I have an application in that container, it will give me metrics only for my app?