Feat links validator

[caviri]

Note

Adds ROR/Infoscience URL validation agents and integrates them into enrichment, updates models (use orcid, unify relatedToOrganizations), normalizes/validates links, and optimizes prompts with token breakdown logging.

• Validations:
  • Add URL validation agent (src/agents/url_validation.py) with ROR API v2 and Infoscience HTML normalization; new ValidationResult model and utilities (validation_utils.py).
  • Integrate validation into organization and academic catalog enrichment (normalize URLs, drop invalid, adjust confidence).
• Organization Enrichment:
  • Pre-search ROR data and include in prompts; add validate_ror_organization_tool; simplify instructions; add detailed token breakdown logging.
• Data Models:
  • Person: replace orcidId with orcid; email supports Union[str, List[str]].
  • Repository: merge relatedToOrganizationsROR into relatedToOrganizations as List[Union[str, Organization]]; validators/mappings updated.
  • AcademicCatalogRelation: remove externalId; fix matchedOn typing.
  • Infoscience entities: add type discriminators and strict URL validators; normalized URLs.
  • Conversion/exports updated to reflect new fields and JSON-LD mappings.
• Agents/Prompts:
  • Academic catalog: validate Infoscience relations; structured outputs assigned directly.
  • Repository/User/Org prompts: enforce single-search policy on 0 results; optimize wording; add token breakdown logs.
  • Analysis flows updated to handle mixed org strings/objects and new orcid field.
• Context/Utils:
  • Infoscience parsers now emit normalized entity URLs; “STOP searching” messaging in empty results.
  • URL utilities tolerate empty strings and update ORCID checks.
• Dependencies:
  • Add markdownify for HTML→markdown conversion.

^{Written by Cursor Bugbot for commit f944e6f. This will update automatically on new commits. Configure here.}