chore: fix evals/runledger #18

	name: runledger
	on:
	workflow_dispatch:
	pull_request:
	paths:
	- "openchatbi/**"

	jobs:
	runledger:
	if: github.event_name == 'workflow_dispatch' \|\| contains(github.event.pull_request.labels.*.name, 'runledger')
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v4
	- uses: actions/setup-python@v5
	with:
	python-version: "3.11"
	- name: Install dependencies
	run: \|
	python -m pip install --upgrade pip
	python -m pip install runledger
	python -m pip install .
	- name: Run deterministic evals (replay)
	run: \|
	runledger run evals/runledger --mode replay --baseline baselines/runledger-openchatbi.json
	- name: Upload artifacts
	uses: actions/upload-artifact@v4
	with:
	name: runledger-artifacts
	path: runledger_out/**

	judge:
	# Non-gating LLM-as-Judge run (manual only). Never blocks PR merges:
	# only runs on explicit dispatch, and uploads a report artifact.
	if: github.event_name == 'workflow_dispatch'
	runs-on: ubuntu-latest
	continue-on-error: true
	steps:
	- uses: actions/checkout@v4
	- uses: actions/setup-python@v5
	with:
	python-version: "3.11"
	- name: Install dependencies
	run: \|
	python -m pip install --upgrade pip
	python -m pip install ".[eval]"
	- name: Run LLM-as-Judge over gold corpus
	env:
	OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
	run: \|
	python -m evals.judge.run_judge --cases evals/judge/cases --out judge_out/report.json
	- name: Upload judge report
	uses: actions/upload-artifact@v4
	with:
	name: judge-report
	path: judge_out/**

Navigation Menu