Skip to main content

MARS

MARS 接受 Pandas 或 Polars 宽表,提供数据画像、分箱评估、特征筛选、建模、监控、结构化 report、 Excel/HTML 导出和评分卡能力。

安装

MARS 0.0.26 支持 Python 3.8–3.12。Python 3.8 使用冻结兼容依赖栈;可选的建模、 调参、Notebook 和文档工具要求 Python 3.10+。

pip install mars-risk==0.0.26

从 GitHub 源码安装当前 main 分支:

pip install "git+https://github.com/leeesq/mars-risk.git"

也可以先克隆仓库,再从本地源码安装:

git clone https://github.com/leeesq/mars-risk.git
cd mars-risk
pip install .

建模与调参需要可选依赖:

pip install "mars-risk[ml,tuning]==0.0.26"

0.0.26 正式发布前,请从源码安装进行文档预览验收。Python 3.8 已停止官方安全维护; MARS 的运行兼容不代表解释器仍有安全支持。

最小风险评估

import polars as pl

from mars.analysis import profile_risk

df = pl.DataFrame(
    {
        "income": [3200, 3600, 5200, 6100, 3400, 4300, 5800, 6800],
        "utilization": [0.72, 0.61, 0.29, 0.18, 0.66, 0.48, 0.24, 0.12],
        "target": [1, 1, 0, 0, 1, 1, 0, 0],
    }
)

risk_profile = profile_risk(
    df,
    target="target",
    features=["income", "utilization"],
    method="quantile",
    n_bins=4,
)

summary = risk_profile.report.summary_table
binner = risk_profile.binner

完整的日期、分组、趋势和报告示例见 10 分钟 Quickstart

从任务开始

目标 文档
检查缺失、分布和 PSI 数据画像
使用基准规则评估当前数据 分箱与风险评估
统计、线性或重要性筛选 特征筛选
切分、调参、replay 与 Pipeline Modeling / Pipeline
分布、模型分和表现覆盖率监控 特征与模型监控
Excel、HTML、评分卡与 SQL 报告与评分卡
精确签名、默认值和异常 API Reference

稳定性

Analysis、Feature、Reporting 是当前 Stable 模块。Monitoring、Modeling、Pipeline、Scoring 为 Experimental;受控生产流程应固定精确版本,并为 report 字段、报警结果、评分映射、生成 SQL、 step 契约、replay 和 artifact 路径增加契约回归。

开发检查

python -m ruff check src tests scripts docs/snippets
python -m mypy src/mars
pydoclint src/mars
python scripts/check_private_docstrings.py src/mars
python -m pytest -q
python -m mkdocs build --strict

贡献要求见 CONTRIBUTING.md,许可证见 LICENSE

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mars_risk-0.0.26.tar.gz (460.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

mars_risk-0.0.26-py3-none-any.whl (464.9 kB view details)

Uploaded Python 3

File details

Details for the file mars_risk-0.0.26.tar.gz.

File metadata

  • Download URL: mars_risk-0.0.26.tar.gz
  • Upload date:
  • Size: 460.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for mars_risk-0.0.26.tar.gz
Algorithm Hash digest
SHA256 24445cfc691fb767fb38358c6fadf7774c6b39fd37dfb0f72ad56566742e7b7c
MD5 fbe11557cda800430c1481cd4c8b5d7c
BLAKE2b-256 1392b43ee18146a3661d9bf745bd8f7ccc0dc09c6439efa446d6b3ccca88805b

See more details on using hashes here.

Provenance

The following attestation bundles were made for mars_risk-0.0.26.tar.gz:

Publisher: publish.yml on leeesq/mars-risk

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file mars_risk-0.0.26-py3-none-any.whl.

File metadata

  • Download URL: mars_risk-0.0.26-py3-none-any.whl
  • Upload date:
  • Size: 464.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for mars_risk-0.0.26-py3-none-any.whl
Algorithm Hash digest
SHA256 aa8217f649c9ae023eea81a6930a6d3a82cc31d8a2819107d7be171201c0f55a
MD5 feb1e2d1d420d313288560bfac0e6cbf
BLAKE2b-256 5ec3af1dca360a58b96daf3bb13fcfbad3a91080d4bce2c91a0207484980afed

See more details on using hashes here.

Provenance

The following attestation bundles were made for mars_risk-0.0.26-py3-none-any.whl:

Publisher: publish.yml on leeesq/mars-risk

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

0.0.27

2 files

This release

0.0.26 This release

2 files

0.0.25

2 files

0.0.24

2 files

0.0.23

2 files

0.0.21

2 files

0.0.20

2 files

0.0.19

2 files

0.0.18

2 files

0.0.17

2 files

0.0.16

2 files

0.0.15

2 files

0.0.14

2 files

0.0.13

2 files

0.0.12

2 files

0.0.11

2 files

0.0.10

2 files

0.0.9

2 files

0.0.8

2 files

0.0.7

2 files

0.0.6

2 files

0.0.5

2 files

0.0.3

2 files

0.0.2

2 files

0.0.1

2 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page