First Langfuse evaluations

[lotif]

Summary

Adding Langfuse integration code and adding an evaluation script to the report generation agent.

Clickup Ticket(s): NA

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 💥 Breaking change (fix or feature that would cause existing functionality to not work as expected)
• 📝 Documentation update
• 🔧 Refactoring (no functional changes)
• ⚡ Performance improvement
• 🧪 Test improvements
• 🔒 Security fix

Changes Made

• Small refactorings and fixes
• Adding common Langfuse code in the langfuse.py file
• Adding a ground truth dataset
• Adding a script to upload the dataset to Langfuse
• Adding an evaluation script to run an LLM-as-a-judge against the Report Generation Agent and the ground truth dataset
• Updating the instructions in the README.md file

For an example of how the evaluation results are looking like:
https://us.cloud.langfuse.com/project/cmkwsswke005dad07gxujnipq/datasets/cmkyev4nd000nad084ds2xm30/runs/27328bba-9843-4ccb-940f-6fe1b9e3b0ea

Testing

• Tests pass locally (uv run pytest tests/)
• Type checking passes (uv run mypy <src_dir>)
• Linting passes (uv run ruff check src_dir/)
• Manual testing performed (describe below)

Manual testing details:

Performed manual testing by following the instructions in the README.md file.

Checklist

• Code follows the project's style guidelines
• Self-review of code completed
• Documentation updated (if applicable)
• No sensitive information (API keys, credentials) exposed

[fcogidi]

A couple of comments:

• Will you switch to google-adk later since Amrit and I are using that? Asking 'cause the langfuse integration may be different for that.
• The langfuse_upload script may be general enough to be in the aieng-eval-agents package

[lotif]

@fcogidi

Will you switch to google-adk later since Amrit and I are using that? Asking 'cause the langfuse integration may be different for that.

Yes. I will have one more PR that I will put out on monday with the trajectory evals and next in line is the move to google-adk.

The langfuse_upload script may be general enough to be in the aieng-eval-agents package

Good point. I'm planning to move things around in follow up PRs as well and will keep that in mind.