SDKs: a hook called on each exposure, to send assignments to your own analytics
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 35/100
- Issue 类型
- 功能
- 描述清晰度
- 描述清楚
- 活跃度
- 活跃
- 技术栈
- csharp, go, java, javascript, python, react, react-native
调研方向
The issue spans the JavaScript, Python, Go, React, React Native, Java and .NET SDKs, so there is no single file to jump into; each client constructor needs an optional callback plus the plumbing that records exposures server-side. Start by finding where each SDK evaluates a flag and caches the answer, so the hook fires once per exposure rather than per evaluation, and check the React server-render/hydration path for duplicate calls. Done means the callback receives experiment key, variant and user id once per exposure, errors from it are swallowed, the holdout/targeting case is decided and documented, and each SDK's docs show one example per analytics tool. A newcomer could instead pick one SDK and add the docs example for a single analytics tool as a scoped first step.
由索引模型根据 Issue 内容生成。
描述
The SDKs record exposures on the server, but nothing lets an application see them. A team that analyses experiments
in its own analytics tool or warehouse (Segment, GA4, Amplitude, Mixpanel, RudderStack) has no way to send each
assignment there.
Expected:
- An optional callback, set when the client is created, called once per exposure with the experiment key, the
variant and the user id. Whether it is called when the user was not assigned (a holdout, a mutual exclusion group or
targeting, where the server records no exposure) is decided and documented. - Once per exposure, not once per evaluation: repeated and cached answers, and React's server render followed by
hydration, do not call it twice. - An exception thrown in the callback never reaches the caller or changes the answer.
- In the JavaScript, Python, Go, React, React Native, Java and .NET SDKs, with one example per common analytics tool
in the docs. (The OpenFeature providers use OpenFeature's own hooks.)
- 主要语言
- Python
- 星标
- 0
- 派生
- 4
- 平均合并
- 5 小时 36 分钟
- 30 天内合并 PR
- 509
环境准备
- 提供 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue，再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库，在一个分支上完成修改。
- 提交 Pull Request，并在描述里引用这个 Issue 编号。
Controlvar/controlvar 的其他 Issue
-
Dashboard: an archived flag shows Off with an on/off switch the API refuses可能已有人在做 关联的 PR 仍在进行中或已合并。 未关闭bug post-launch
难度 2/5 1-3 小时 新手友好度 72/100
Controlvar/controlvar#1748 ·
维护者通常 1 天内回复
-
Dashboard: the sequential panel's "Ready to stop" banner shows no sample-ratio-mismatch caution可能已有人在做 关联的 PR 仍在进行中或已合并。 未关闭bug post-launch
难度 2/5 1-3 小时 新手友好度 68/100
Controlvar/controlvar#1737 ·
维护者通常 1 天内回复
-
dashboard: a split URL draft's variant edit shows generic copy when the API refuses it可能已有人在做 关联的 PR 仍在进行中或已合并。 未关闭bug post-launch
难度 2/5 1-3 小时 新手友好度 70/100
Controlvar/controlvar#1562 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 90/100
Controlvar/controlvar#1447 · 2 条评论 ·
维护者通常 1 天内回复
-
Experiment setup: emptying a traffic share box and typing a number shows a leading zero ("034")可能已有人在做 关联的 PR 仍在进行中或已合并。 未关闭bug post-launch
难度 2/5 1-3 小时 新手友好度 72/100
Controlvar/controlvar#1428 ·
维护者通常 1 天内回复
查看 Controlvar/controlvar 的全部 Issue
相似的 Issue
-
难度 1/5 1-3 小时 新手友好度 88/100
aws-samples/sample-per-model-bedrock#782 ·
维护者通常 1 天内回复
-
Google text documents are base64-encoded twice可能已有人在做 @feiiiiii5 今天认领。 未关闭
难度 2/5 1-3 小时 新手友好度 84/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 75/100
维护者通常 1 天内回复
-
bug
难度 2/5 1-3 小时 新手友好度 76/100
coveragepy/coveragepy#2364 ·
维护者通常 1 天内回复
-
难度 1/5 1 小时以内 新手友好度 72/100