Hacktoberfest 2026：メンテナが10月に向けて印を付けた、オープンで初心者向けの issue。 Hacktoberfest の issue を見る

[bot] Instrument Guardrails AI (30,625 weekly downloads)

オープン
  • 何か問題がありますか？
#807 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 1 日以内に返信

まだ誰も着手していません。

評価

難易度
4/5
見積もり時間
3〜5日
初心者へのやさしさ
48/100
issue の種類
機能追加
明瞭さ
おおむね明確
活発さ
活発
技術スタック
python
領域
observability

調査の方向性

Start in py/src/braintrust/integrations/ and inspect py/src/braintrust/wrappers/, then read the [tool.braintrust.matrix] section of py/pyproject.toml and py/noxfile.py for existing integration patterns. Done means the guarded call and parse entry points are instrumented, with the integration represented in the matrix and a nox session or other repository test coverage added.

索引モデルが issue の本文から書いたものです。

説明

new-integration

Summary

guardrails-ai (import name guardrails, by Guardrails AI Inc.) is a framework for adding structural, type, and quality validation directly around LLM calls. Its primary execution surfaces are Guard.for_pydantic(output_class=...) / Guard().use(<validators>) to construct a guard, and then the callable form guard(model=..., messages=[...]), which invokes the LLM directly and returns a response with .validated_output — i.e. the Guard itself owns the LLM call, not just post-hoc validation of already-generated text. guard.parse(...) is the sibling entrypoint for validating a given string against attached validators.

This is a distinct project/package from NVIDIA NeMo Guardrails (nemoguardrails, tracked separately in #768) — different maintainer, different PyPI package, different API shape (Guard/validators around a direct LLM call vs. NeMo's LLMRails rails pipeline). This repository has zero instrumentation for Guardrails AI anywhere: no py/src/braintrust/integrations/guardrails* directory, no wrapper, no nox session, no cassette directory, and no entry in py/pyproject.toml's [tool.braintrust.matrix].

What's missing

  • guard(model=..., messages=[...]) (and async variant) — the primary "guarded LLM call" entrypoint; since Guardrails AI drives the LLM call itself, an unwrapped call here means the underlying model call, validated output, and any validator retries are entirely invisible to Braintrust tracing.
  • guard.parse(...) / guard.parse_async(...) — validates a given string against attached validators, returning a result with .validation_passed.
  • Guard.for_pydantic(output_class=...) / Guard().use(<validators>) — guard construction with attached validators (useful for capturing validator configuration as metadata on the guarded-call span).

Weekly downloads

Weekly downloads: 30,625 (as of 2026-09-28; https://pypistats.org/api/packages/guardrails-ai/recent)

Braintrust docs status

not_found. Checked https://www.braintrust.dev/docs/guides/tracing and https://www.braintrust.dev/docs/guides/tracing/integrations (both checked 2026-09-28) — neither the tracing-integrations list ("OpenTelemetry, Temporal, Vercel AI SDK, OpenRouter SDK, LangChain, LangChain4j, LangSmith, LlamaIndex, Apollo GraphQL, Cloudflare Workers AI, DSPy, Google Discovery Engine, Instructor, LiteLLM, Ruby LLM, Spring AI, Traceloop, TrueFoundry") nor the agent-framework list mentions Guardrails AI or Guardrails.

Upstream sources

Braintrust docs sources checked

Local repo files inspected

  • py/src/braintrust/integrations/ — no guardrails/ directory (only unrelated py/src/braintrust/integrations/__init__.py and the separately-tracked nemoguardrails gap in #768)
  • py/src/braintrust/wrappers/ — no Guardrails AI wrapper
  • py/pyproject.toml [tool.braintrust.matrix] — no guardrails-ai entry
  • py/noxfile.py — no test_guardrails session
  • Repo-wide case-insensitive grep for guardrails.ai/guardrails_ai under py/src/braintrust/**/*.py — zero matches
主要言語
Python
スター
20
フォーク
18
平均マージ
21時間 8分
マージ済み PR（30日）
81

環境構築

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

braintrustdata/braintrust-sdk-python のほかの issue

braintrustdata/braintrust-sdk-python の issue をすべて見る

似ている issue

Python の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。