Add offset and limit parameters to API endpoints

[arne-bdt]

Add comprehensive task documentation for implementing offset/limit pagination on three collection endpoints that currently return all results.

Audit Findings:

• ✅ GET /version/history - already has pagination (limit=100, max=1000)
• ✅ GET /version/blame - already has pagination (limit=100, max=1000)
• ❌ GET /version/branches - missing pagination (BranchController:73)
• ❌ GET /version/tags - missing pagination (TagController:64)
• ❌ GET /version/refs - missing pagination (RefsController:43)

Task Structure:

• README.md: Overview, status, implementation pattern
• 01-branch-pagination.md: BranchController (2-3 hours)
• 02-tag-pagination.md: TagController (2-3 hours)
• 03-refs-pagination.md: RefsController (2-3 hours)

Total Estimated Time: 6-8 hours

Benefits:

• Performance: Prevent memory exhaustion with large datasets
• Consistency: All collection endpoints behave the same
• Scalability: Support datasets with thousands of branches/tags
• UX: Incremental data loading for clients

Pattern: Follow HistoryController implementation

• Default: limit=100, offset=0, max=1000
• RFC 5988 Link headers for pagination
• PaginationMetadata in response DTOs

Protocol Notes:
While not mandated by SPARQL 1.2 Protocol, pagination is a REST best practice for collection endpoints and aligns with existing history/blame implementations.

🤖 Generated with Claude Code