Hacktoberfest 2026:维护者为十月标记出来的 issue,仍然开放、适合新手。 浏览 Hacktoberfest issue

Numerical Instability in Tests

未关闭
#365 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

还没有人认领这个 Issue。

评估

难度
3/5
预计耗时
1-2 天
新手友好度
25/100
Issue 类型
缺陷
描述清晰度
需要澄清
活跃度
停滞
技术栈
python
领域
testing-qa

调研方向

Start with TestLaplaceDistribution.test_check_statistics_for_geo_unit_values and reproduce the failure in CI or the reported test environment. Investigate the sampling statistics and assertion tolerance; done means the test no longer flakes or fails while preserving meaningful checks for the Laplace distribution.

由索引模型根据 Issue 内容生成。

描述

Type: Bug :bug:

Description

I have noticed one of the tests flaps and sometimes fails, I guess its due to some numerical instability?
https://github.com/OpenMined/PyDP/runs/2679233108?check_suite_focus=true

Screen Shot 2021-05-27 at 12 21 47 pm
______ TestLaplaceDistribution.test_check_statistics_for_geo_unit_values _______
[gw1] linux -- Python 3.6.13 /home/runner/.cache/pypoetry/virtualenvs/pydp-KC8mieZ4-py3.6/bin/python

self = <test_distributions.TestLaplaceDistribution object at 0x7fe191d630b8>

    def test_check_statistics_for_geo_unit_values(self):
    
        ld = LaplaceDistribution(epsilon=1.0, sensitivity=1.0)
        samples = [ld.sample(scale=1.0) for _ in range(k_num_geometric_samples)]
        mean = dp.util.mean(samples)
        var = dp.util.variance(samples)
    
        assert expect_near(0.0, mean, 0.01)
        assert expect_near(2.0, var, 0.1)
        assert expect_near(0.0, skew(samples, mean, math.sqrt(var)), 0.1)
>       assert expect_near(3.0, kurtosis(samples, mean, var), 0.1)
E       assert False
E        +  where False = expect_near(3.0, 3.1097853317836956, 0.1)
E        +    where 3.1097853317836956 = kurtosis([-0.18739762318364228, 0.19336953977835947, 2.2439366761800557, -0.120824026083028, 0.3653907802163303, 2.181137632465834, ...], 0.00018447984155104587, 2.0099539175529073)

How to Reproduce

See CI

Expected Behavior

Not failing.

Screenshots

See above

System Information

See Link

Additional Context

See above link

主要语言
Python
星标
551
派生
142
PR 合并指标
30 天内没有已合并 PR

环境准备

  • 提供 Dockerfile 或 Docker Compose 文件
  • 没有 Pull Request 模板
  • 阅读贡献指南

从这里开始

  1. 先读完整个 Issue,再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库,在一个分支上完成修改。
  4. 提交 Pull Request,并在描述里引用这个 Issue 编号。

OpenMined/PyDP 的其他 Issue

查看 OpenMined/PyDP 的全部 Issue

相似的 Issue

更多 Python Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。