@ossintel/scoring

@ossintel/scoring: Deterministic scoring engine for calculating OSS health, impact, risk, activity, and community metrics from normalized repository and ecosystem data.

@ossintel/scoring

Deterministic OSS scoring engine with modular architecture.

Purpose

Converts normalized repository and developer metrics into objective scores.

Responsibilities

Repository Scoring

• Health, Impact, Activity, Community, Risk, and Overall scores

Identity Scoring (OSSIQ)

Four-pillar reputation engine:

• Maintainer — Repository health weighted by popularity (GitHub-only)
• Contributor — Quality-weighted external PR scoring
• Organization — Active org leadership evaluation
• Influence — Downstream reach (stars, forks + additive npm/SO bonuses)

Capability Scoring

• Package Publishing (calculatePublishingScore) — npm downloads, active packages, verified publisher (extensible to NuGet, PyPI, crates.io, Go)
• Knowledge Sharing (calculateKnowledgeScore) — Stack Overflow reputation, answers, acceptance rate (extensible to Dev.to, Hashnode)

Supporting Modules

• Badges — Achievement detection from cross-platform activity
• Skills — Topic expertise radar aggregating GitHub, npm, and Stack Overflow signals
• Evidence & Factors — Human-readable explanations for each pillar

Philosophy

• Scores are deterministic: same inputs → same outputs
• Scores are monotonic: linking additional providers can only increase reputation
• GitHub defines the core OSS reputation (~80-85%)
• Additional providers contribute additive capability bonuses
• No AI participates in score calculation

Module Architecture

index.ts
├── repository-scoring.ts      # Repo-level: health, impact, activity, community, risk
├── identity-scoring.ts        # Orchestrator: computes overall OSSIQ score
├── maintainer-scoring.ts      # Pillar: GitHub repo health weighted average
├── contributor-scoring.ts     # Pillar: external PR quality weighting
├── organization-scoring.ts    # Pillar: org leadership evaluation
├── influence-scoring.ts       # Pillar: stars + forks + additive bonuses
├── publishing-scoring.ts      # Capability: Package Publishing (npm)
├── knowledge-scoring.ts       # Capability: Knowledge Sharing (Stack Overflow)
├── badges.ts                  # Achievement detection
├── skills.ts                  # Topic expertise aggregation
├── evidence.ts                # Evidence & factors generation
└── topic-mappings.ts          # Canonical topic → keyword mappings

Non-goals

• API calls
• AI
• Charts
• UI

Testing

Every scoring algorithm should have comprehensive unit tests.

No external dependencies should affect score calculation.

✨ Why @ossintel/scoring?

• 100% Deterministic & Transparent: Zero network requests or probabilistic AI models; calculations are mathematical and fully reproducible.
• Modular Pillar Architecture: Each scoring dimension (Maintainer, Contributor, Organization, Influence) is a self-contained module, easy to extend or override.
• Capability-Specific Scoring: First-class calculatePublishingScore and calculateKnowledgeScore exports for direct use by consumers.
• Monotonic Guarantees: Linking npm or Stack Overflow can only increase scores, never reduce them.
• Fully Tested: Every score calculation is validated by comprehensive unit tests to prevent regression.

🌍 Open Source Ecosystem Impact Engine

The Ecosystem Impact Engine measures developer impact beyond owned repositories, focusing on contributions to external projects.

Heuristics Classification Pipeline

[User Merged PR]
       │
       ├──► Contains "typo", "readme", "docs" in Title/Labels? ─────► [Docs / Typo Fix] (Weight: 0.4x)
       ├──► Contains "test", "spec", "qa" in Title/Labels? ────────► [Test Contribution] (Weight: 1.2x)
       ├──► Contains "chore", "dependencies" in Title/Labels? ──────► [Chore / Dependency] (Weight: 0.5x)
       └──► Default Code change ────────────────────────────────────► [Core Code Contribution] (Weight: 1.0x)

Weighting Matrix

Merged pull requests are weighted according to target repository popularity and quality multipliers:

| Target Repo Stargazers | Tier Category | Base Weight | Multiplier Applied | | ---------------------- | ----------------------- | ----------- | ---------------------------------------------------------------- | | ≥ 20,000 Stars | Tier 1 (Frameworks) | 5.0 | Quality Multiplier (Docs: 0.4, Test: 1.2, Chore: 0.5, Code: 1.0) | | ≥ 2,000 Stars | Tier 2 (Medium OSS) | 3.0 | Quality Multiplier | | < 2,000 Stars | Tier 3 (Small OSS) | 1.0 | Quality Multiplier |

The resulting Ecosystem Contribution Score ($S_{\text{eco}}$) is aggregated as:

$$ S_{\text{eco}} = \min\left(100, \sum (\text{BaseWeight} \times \text{QualityMultiplier})\right) $$

📦 Installation

[!TIP] This package (ossintel) is an official alias of @ossintel/scoring.

We provide this package to offer shorter import paths and improved discoverability. While both packages provide identical functionality, @ossintel/scoring is the primary source of truth.

• Use ossintel if you prefer the shorter name or specific branding or ESM Only.
• Use @ossintel/scoring for the most stable long-term reference and standard alignment.

| Feature | @ossintel/scoring | ossintel | | --- | --- | --- | | Source Code | ✅ Primary | 🔗 Proxy | | Updates | Immediate | Synchronized (Immediately) | | Bundle Size | 100% | 100% (Zero overhead) | | Format | ESM + CJS | ESM Only | | Maintenance | ✅ Primary | 🔗 Proxy (inherits) | | Security | ✅ Primary | 🔗 Proxy (inherits) |

Security: Security audits are performed on the canonical package; this alias inherits all security patches automatically.

$ pnpm add @ossintel/scoring

or

$ npm install @ossintel/scoring

or

$ yarn add @ossintel/scoring

License

This library is licensed under the MIT open-source license.