# BreakPilot Compliance SDK - Legal Corpus

Pre-indexed legal documents for the RAG system.

## EU Regulations

| Document | Chunks | Description |
|----------|--------|-------------|
| DSGVO (GDPR) | ~99 | EU General Data Protection Regulation |
| AI Act | ~85 | EU Artificial Intelligence Act |
| NIS2 | ~46 | Network and Information Security Directive |
| ePrivacy | ~32 | ePrivacy Directive (Cookie Directive) |
| CRA | ~41 | Cyber Resilience Act |
| EUCSA | ~28 | EU Cybersecurity Act |
| Data Act | ~35 | EU Data Act |
| DGA | ~25 | Data Governance Act |
| DSA | ~52 | Digital Services Act |
| DMA | ~38 | Digital Markets Act |
| EAA | ~22 | European Accessibility Act |
| SCC | ~18 | Standard Contractual Clauses |
| DPF | ~15 | EU-US Data Privacy Framework |

## German Regulations

| Document | Chunks | Description |
|----------|--------|-------------|
| TDDDG | ~28 | Telekommunikation-Digitale-Dienste-Datenschutz-Gesetz |
| TTDSG | ~24 | Telekommunikation-Telemedien-Datenschutz-Gesetz |
| BDSG | ~45 | Bundesdatenschutzgesetz |
| IT-SiG | ~32 | IT-Sicherheitsgesetz |
| BSI-KritisV | ~28 | BSI-Kritisverordnung |

## Directory Structure

```
legal-corpus/
├── eu/
│   ├── dsgvo/
│   │   ├── articles/
│   │   ├── recitals/
│   │   └── metadata.json
│   ├── ai-act/
│   ├── nis2/
│   ├── eprivacy/
│   ├── cra/
│   ├── eucsa/
│   ├── data-act/
│   ├── dga/
│   ├── dsa/
│   ├── dma/
│   ├── eaa/
│   ├── scc/
│   └── dpf/
├── de/
│   ├── tdddg/
│   ├── ttdsg/
│   ├── bdsg/
│   ├── it-sig/
│   └── bsi-kritisv/
└── embeddings/
    └── (generated vector embeddings)
```

## Indexing

Documents are automatically indexed on first startup of the RAG service.

To manually re-index:

```bash
# Via CLI
breakpilot-cli index --all

# Via API
POST /api/v1/rag/index
```

## Adding Custom Documents

Organizations can add their own internal documents:

```bash
# Upload via CLI
breakpilot-cli upload --file policy.pdf --category internal

# Via API
POST /api/v1/rag/documents
Content-Type: multipart/form-data
```

## Embedding Model

Default: `bge-m3` via Ollama

Supports:
- German legal terminology
- Multi-lingual (DE/EN)
- High-quality semantic search