z3rno-evals

⚠️ Deprecated — full rewrite in progress

z3rno is being rewritten from the ground up as a single open-source Rust monorepo (Apache 2.0), following direct feedback from enterprise users that the current architecture is too complex.

• Maintenance mode only — bug fixes, no new features.
• Existing installs keep working. Nothing here is being yanked, archived, or removed.
• The new monorepo lands at the-ai-project-co/z3rno.
• Track progress: https://github.com/the-ai-project-co/z3rno-evals/issues/3

Eval framework for Z3rno. Measures the four metrics that gate every Phase E release:

• recall@k — fraction of golden questions where an expected Memo is in the top-k results
• MRR — mean reciprocal rank of the first correct Memo
• faithfulness — LLM-judge score: does the recalled context support the expected answer?
• latency — p50 / p95 / p99 of recall() calls against a live z3rno-server

A golden dataset is a JSON file of Q&A items with expected memory IDs, expected entities, and latency budgets. The CLI runs the dataset against a live server and emits both a results.json (machine-readable) and a report.md (PR-comment-ready).

Install

pip install z3rno-evals
# Optional — for the LLM-driven faithfulness judge
pip install 'z3rno-evals[llm-judge]'

Quickstart

# 1. Boot a local z3rno-server (see z3rno-server/Makefile)
make -C ../z3rno-server dev-up

# 2. Seed the memories the golden dataset expects
python -m z3rno_evals.seed --against http://localhost:8000 --api-key sk-... \
    --dataset datasets/golden_v1.json

# 3. Run the evals
z3rno-evals run \
    --against http://localhost:8000 \
    --api-key sk-... \
    --dataset datasets/golden_v1.json \
    --output results/

# 4. Look at the report
cat results/report.md

Metrics Table

Golden Dataset Format

{
  "version": 1,
  "name": "golden_v1",
  "items": [
    {
      "id": "q-001",
      "agent_id": "agent-1",
      "query": "What does the user prefer for notifications?",
      "expected_memory_ids": ["01HXY1...", "01HXY2..."],
      "expected_entities": ["dark mode", "weekly digest"],
      "expected_answer": "The user prefers dark mode and weekly digest emails.",
      "strategy": "AUTO",
      "top_k": 5,
      "latency_budget_ms": 500
    }
  ]
}

See datasets/golden_v1.json for the full sample.

Architecture

• src/z3rno_evals/dataset.py — Pydantic models + loader for the golden dataset format
• src/z3rno_evals/metrics.py — pure-function metric implementations
• src/z3rno_evals/runner.py — drives the eval against a live server via the z3rno SDK
• src/z3rno_evals/judges/ — FaithfulnessJudge ABC + LLM + stub implementations
• src/z3rno_evals/report.py — JSON + Markdown report renderers
• src/z3rno_evals/cli.py — argparse entry point (z3rno-evals run, z3rno-evals seed)

License

Apache 2.0.