feat: academicCatalogsRelations

[caviri]

Note

Introduce an academic catalog enrichment agent with new models and Infoscience fixes, integrate it across repo/user/org pipelines, and correct token estimation and EPFL assessment config.

• Academic Catalog Enrichment
  • Add agent (src/agents/academic_catalog_enrichment.py) and prompts with structured output keyed by exact names.
  • Integrate into repository, user, and organization analysis flows for direct assignment of academicCatalogRelations.
• Data Models
  • New academic_catalog.py with AcademicCatalogRelation and AcademicCatalogEnrichmentResult.
  • Replace legacy infoscienceEntities with academicCatalogRelations in Person, Organization, SoftwareSourceCode, GitHubUser, etc.; update exports and model rebuilds.
• Infoscience Client
  • Fix searches to use configuration=person|orgunit|researchoutputs; add parsers for person/orgunit and get_entity_by_uuid(); improve markdown outputs (incl. UUIDs).
• Analysis Pipeline
  • Add run_academic_catalog_enrichment() to repo/user/org; assign relations via direct dict lookup; accumulate token usage.
  • Add final EPFL assessment token estimation and correct config key (run_epfl_assessment).
• LLM Config
  • Extend model_config.py with configs/env mappings for run_academic_catalog_enrichment and run_epfl_assessment.
• Token Tracking
  • Fix estimated token keys and implement estimation where missing across agents.
• Docs & Rules
  • Add detailed docs (Option B implementation, refactor summary, API findings, token tracking fix) and update Cursor rules for agents, architecture, deployment.

^{Written by Cursor Bugbot for commit c69405f. This will update automatically on new commits. Configure here.}