fix: update geolocation fallback API and use static country code mapping

[ucodia]

Description

Fixes geolocation fallback API call

Issue 1

The fallback API will always fail when using the HTTPS endpoint

curl https://ip-api.com/json/
{"status":"fail","message":"SSL unavailable for this endpoint, order a key at https://members.ip-api.com/"}%

Using the HTTP endpoint is not great because this could expose the application location to eavesdropper and be vulnerable to man-in-the-middle.

Issue 2

If the fallback API call fails, it reports on the wrong URL on the exception: https://github.com/ucodia/codecarbon/blob/ebc8d1a3f5dcae9013dcbe6883d05e711ad6ce28/codecarbon/external/geography.py#L136

Non-issue 1

The current fallback relies on another API to map the 2 letters country code to a 3 letters code. This is static data that almost never changes, by switching to a static mapping provider we can make the fallback more resilient.

Related Issue

Fixes #1101

Motivation and Context

Certain IP ranges are not covered by the primary API and therefore a fallback is necessary otherwise CodeCarbon client will estimate on the wrong energy mix.

How Has This Been Tested?

• Unit testing
• Tested in production environment

Screenshots (if appropriate):

N/A

Types of changes

What types of changes does your code introduce? Put an x in all the boxes that apply:

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

Go over all the following points, and put an x in all the boxes that apply.

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING.md document.
• I have added tests to cover my changes.
• All new and existing tests passed.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.25%. Comparing base (b655866) to head (89e5a62).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1104      +/-   ##
==========================================
+ Coverage   78.21%   78.25%   +0.04%     
==========================================
  Files          38       38              
  Lines        3631     3638       +7     
==========================================
+ Hits         2840     2847       +7     
  Misses        791      791              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[benoit-cty]

Thanks Lionel for testing your previous PR in a production environment.

Can you run uv run pre-commit run --all-files to fix the error.

I will prefer to avoid adding a dependency. But it is not really related to your PR. I think we could handle this later. Maybe @SaboniAmine could add the 2 letters code in the carbon intensity dataset from https://github.com/pycountry/pycountry/blob/main/src/pycountry/databases/iso3166-1.json in his work on https://github.com/mlco2/codecarbon-hardware-data . So we will remove the dependency later.

[ucodia]

@benoit-cty ran the pre-commit script and fixed the code formatting issue.

I understand we want to keep dependency to the minimum, though highlighting we really traded a weak network dependency for a direct code dependency, which should improves the library resiliency.

[benoit-cty]

Thanks, I will open an issue to remove this dependency later.

[SaboniAmine]

Hi, thanks for this contribution!
Indeed, I guess adding temporarily pycountry is a good trade-off, as the work to integrate the new dataset still needs 1-2 weeks to be complete.

[benoit-cty]

Hi, thanks for this contribution! Indeed, I guess adding temporarily pycountry is a good trade-off, as the work to integrate the new dataset still needs 1-2 weeks to be complete.

Issue created : https://github.com/mlco2/codecarbon-hardware-data/issues/3