API Emissions Tracker

A Python library that estimates the carbon emissions linked to cloud-based API services, such as AI cloud services, for demonstration purposes.

Installation

To install the package, you can use pip, the Python package manager. Open a command line or terminal and run the following command:

pip install api_emissions_tracker

Usage

Once the package is installed, you can include and use it in your Python code.

Sample code:

from openai import AzureOpenAI
from api_emissions_tracker import APIEmissionsTracker

AZURE_ENDPOINT = "[YOUR_AZURE_ENDPOINT]"
OPENAI_MODEL = "gpt-35-turbo"   # specify a supported model from the list
OPENAI_API_KEY = "[YOUR_OPENAI_API_KEY]"

tracker = APIEmissionsTracker()
tracker.start()

client = AzureOpenAI(azure_endpoint=AZURE_ENDPOINT, api_key=OPENAI_API_KEY)

response = client.chat.completions.create(model=OPENAI_MODEL, messages=[{"role": "system", "content": "What is the origin of the Olympic Games?"}])

print(response.choices)

tracker.stop()

Sample output:

[Choice(finish_reason='stop', index=0, logprobs=None, message=ChatCompletionMessage(content='The Olympic Games originated in ancient Greece around the 8th century BCE. They were held in Olympia, a small town in the western region of the Peloponnese. The Games were a religious festival dedicated to the Greek god Zeus and were held every four years. The first recorded Olympic Games took place in 776 BCE, and they continued for nearly 12 centuries until they were abolished in 393 CE by the Christian Byzantine Emperor Theodosius I.', role='assistant', function_call=None, tool_calls=None), content_filter_results={'hate': {'filtered': False, 'severity': 'safe'}, 'self_harm': {'filtered': False, 'severity': 'safe'}, 'sexual': {'filtered': False, 'severity': 'safe'}, 'violence': {'filtered': False, 'severity': 'safe'}})]
Total emissions: 0.26814 g CO2e

For demonstration purposes you can also uses the Python library Stub AI which mimics the API requests. Simply replace the following lines:

from stub_ai import StubAzureOpenAI
...
client = StubAzureOpenAI(azure_endpoint=AZURE_ENDPOINT, api_key=OPENAI_API_KEY)

Logging

A log file with more details can be written to the hard disk using the following parameter. By default, this feature is disabled.

tracker = APIEmissionsTracker(write_log_file=True)

Customization

The emission factors can be checked, changed and extended in emission_factors.json. They can also get overwritten and enhanced as follows:

custom_emission_factors = {
  "gpt-35-turbo": {
    "per_completion_token": 0.001,
    "per_prompt_token": 0.002,
    "reference_url": "https://my.own.source"
  },
  "my-model": {
    "per_completion_token": 0.00051,
    "per_prompt_token": 0.00051,
    "reference_url": "https://my.own.source2"
  }
}

tracker = APIEmissionsTracker(emission_factors=custom_emission_factors)