Calibration Explorer

io.github.m-sanchezv0.2.0更新于 Oct 5, 2026

Check classifier confidence against accuracy, fit temperature scaling and save local reports.

已验证STDIO仅桌面AI & MLData & Analytics

概览

AI 生成的概览

让助手检查分类器保存的置信度是否与实际准确率相符、拟合温度缩放,并在本地生成 HTML 和 JSON 报告。

功能
这是一个用于 Calibration Explorer 的本地 stdio MCP 服务器。其工具可加载预测文件、评估保存的置信度与真实正确率是否一致、在校准行上拟合单一温度、在策略验证行上设定并锁定接受阈值、复核测试结果,并生成 HTML 与 JSON 评估报告。它复用与浏览器应用相同的解析器、数值库、流程防护和报告生成器,不训练也不运行分类器。
适用场景
当你已有保存好的分类器预测结果,想判断模型置信度是否可信、用温度缩放进行校准,或在查看测试数据前确定接受阈值时,适合使用。它面向项目目录内的本地、基于文件的校准工作。
运行要求
通过 npx 从 npm 包 @m-sanchez/calibration-explorer-mcp 以 stdio 在本地运行;需要 Node.js 20 或更高版本。--input-root 与 --output-root 参数为必填且可重复,必须是已存在的绝对路径目录。未声明任何账号、API 密钥、环境变量或请求头。已验证的客户端是 Claude Code 交互式终端;其他 MCP 宿主未声明为已验证。
安装前请注意
它会读取你选择的预测文件,且只把报告写入新文件夹;配置根目录之外的路径会被拒绝,已有文件绝不会被覆盖。在 review_test 通过客户端的征询表单收到明确决定之前,测试结果保持不打开。没有遥测或远程服务,但返回的摘要仍可能到达你的 MCP 客户端的模型提供方。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Calibration Explorer,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Desktop only,通过 STDIO。 STDIO 服务会启动本地进程,因此需要 SourceWeft 桌面宿主。

其他 MCP 客户端

参照 仓库 中的启动说明。

README

Calibration Explorer MCP

Local stdio MCP server for Calibration Explorer. It checks whether a classifier's saved confidence matches how often it was correct, fits one temperature on calibration rows, evaluates and locks an acceptance threshold on policy-validation rows, and writes HTML and JSON assessment reports into your project.

It reads prediction files you select, uses the same parser, numerical library, workflow guards and report generator as the browser app, and does not train or run a classifier. There is no telemetry or remote service. Returned summaries can still reach your MCP client's model provider.

Use with Claude Code

Save this configuration, replacing both directories with absolute paths that already exist:

json
{  "mcpServers": {    "calibration": {      "type": "stdio",      "command": "npx",      "args": [        "-y", "@m-sanchez/[email protected]",        "--input-root", "ABSOLUTE_PREDICTION_DIRECTORY",        "--output-root", "ABSOLUTE_REPORT_DIRECTORY"      ]    }  }}

Then launch claude --mcp-config calibration.mcp.json --strict-mcp-config and check /mcp. The verified client is the Claude Code interactive terminal, tested against the server run from a checkout; this npm route and the plugin have passed the clean-install test suite but not yet a full session in Claude Code. Other MCP hosts are not claimed as verified. Node.js 20 or newer is required.

--input-root and --output-root are required and repeatable. Paths outside them are refused, reports go into new folders only, and existing files are never overwritten. --help and --version print and exit.

Tools

load_predictions, assess_predictions, fit_temperature, set_policy, review_test and create_report. Test outcomes stay unopened until review_test receives an explicit decision through the client's elicitation form. Input formats, the complete workflow, disclosure rules and limits are documented in docs/mcp.md.

What the package contains

dist/cli.js is a readable, unminified bundle of the server, the shared app modules and @m-sanchez/calibrated from the exact Git commit the app pins. The official MCP TypeScript SDK and zod are exact-pinned runtime dependencies. The UCI digits reference ships under public/examples with its CC BY 4.0 attribution. Reports record the same app source and numerical distribution hashes as a checkout of the same commit; dist/NOTICE.txt lists them with the bundled licence.

MIT licence. Copyright (c) 2026 Miguel Sánchez Durán.

来源:packages/mcp/README.md,提交 1d2c32f

工具

0
工具元数据尚未被收录。

版本历史

1
  1. v0.2.0最新Oct 5, 2026