feat(v1): add researcher, journal and research paper models

[rdahis]

Purpose

Backend for the new Invited Researchers initiative (website page /invited-researchers), and a first model for research papers so datasets can be linked to the papers they come from or are used in.

Description

New models in v1 (single migration 0056_researchers_and_research_papers):

• Researcher: name, slug, picture, position, affiliations (M2M to Organization), phd_institution (FK to Organization), PhD year, description (blurb), public email, website, LinkedIn, Google Scholar, Lattes, ORCID, themes (research fields), optional OneToOne to Account.
  • position and description are translated (pt/en/es).
  • Properties exposed in GraphQL: is_invited_researcher (has an active term) and is_invited_researcher_alumni (had terms, all finished).
• InvitedResearcherTerm: one row per 2-year, indefinitely renewable term. cohort is a YYYY.S label (e.g. 2027.1, validated). end_at defaults to start_at + 2 years.
• Journal: name, abbreviation, ISSN/E-ISSN, publisher, website.
• ResearchPaper: title, citation author string, M2M to registered Researchers, M2M to Dataset (reverse Dataset.research_papers), journal FK, publication status (published / forthcoming / working paper), year, volume/issue/pages, DOI (normalized on clean; doi_url property), URL, Google Scholar link, abstract.

Admin:

• Pages for all four models.
• Researcher page has inlines for terms and papers.
• Dataset admin gets a Research Papers inline.

All models are publicly readable through the auto-generated GraphQL schema, with the default staff-only mutations.

Also applies ruff format to 3 pre-existing blocks in models.py, which the pre-commit hook requires once the file is touched.

Checklist

• I have reviewed the code changes.
• I have tested the changes locally.
• I have updated the documentation if needed.
• I have added/modified tests to ensure the changes are valid.

Testing and evidence

• New backend/apps/api/v1/tests/test_researchers.py (10 tests) covers:
  • the 2-year term default, including leap days
  • date validation
  • active/alumni logic
  • cohort label validation
  • DOI normalization and the paper–researcher–dataset links
  • anonymous GraphQL access
• The full suite has 3 failures and 24 errors, identical on main and on this branch. Some come from the shared dataset_dados_mestres fixture, which passes an organization field that Dataset no longer has.
• Tested end-to-end against the website branch with seeded data, in pt/en/es.

Next steps

• Deploy this before the website PR: the page queries these models.
• Researcher emails entered in admin are public via GraphQL.
• When adding researchers, fill all three language tabs. MODELTRANSLATION_AUTO_POPULATE copies the first value into the others.