recherche-theses-allison-fiorentino

What it does

Specialized backend skill for searching, mapping, and analyzing French doctoral theses in law, including in-preparation theses since 2006. It supports multi-criteria search (subject, directors, juries, institutions, doctoral schools, period, status), academic cartography (who directs on which subject, jury memberships, co-direction networks, productive doctoral schools), and temporal analyses (topic evolution, saturation detection, upcoming defenses). It also filters by sub-discipline and instructs to trigger when user mentions theses or related terms, and to use a doctrine-focused search for doctrinal works.

How it works

• Uses theses.fr API endpoints under /api/v1/theses/ and /api/v1/personnes/ to retrieve data about theses, persons, and their roles. It documents how details differ between endpoints (theses/recherche vs theses/these vs personnes/personne) and notes normalization needs when aggregating data.
• Provides a workflow for building queries with Lucene-style filters (e.g., titrePrincipal, discipline, dateSoutenance, status) and illustrates how to combine them with AND/OR/NOT and multi-criteria examples.
• Recommends a two-step cartography approach: extract unique PPNs from directeurs/rapporteurs/examinateurs, then query /personnes/personne/{PPN} to obtain aggregated roles; use this to map who leads or chairs on topics and to profile doctoral schools or establishments.
• Describes temporal analyses like computing year-based counts, calendars of upcoming defenses, and delays between inscription and defense, with guidance on how to compute these from date fields.
• Details filtering strategies by sub-discipline (droit*) and how to refine with sujetsLibelle and full-text queries, plus caveats about non-requêtable fields and field name differences between endpoints.

When to use it

Use when the user requests object-thesis data, academic cartography, or temporal analyses related to French law theses (soutenue or en cours) and when sub-discipline filtering is desired. Trigger when user mentions thèse, doctorat, directeur de thèse, jury de thèse, école doctorale, soutenance, doctorant, or asks for ecosystem analysis. For doctrinal articles or books, use the doctrine search instead.

What it can touch

• Endpoints: /api/v1/theses/recherche/, /api/v1/theses/these/{NNT}, /api/v1/personnes/recherche/, /api/v1/personnes/personne/{PPN} (as described in the guidance).
• Data fields: titrePrincipal, discipline, dateSoutenance, status, accessible, langues, etabSoutenanceN, sujetsLibelle, directeurs, rapporteurs, examinateurs, president, ecolesDoctorales, theses, personnes. Note that field names differ between endpoints and may require normalization on the client side.

Caveats

• The skill relies on the theses.fr API and HAL TEL endpoints; access may depend on public vs restricted data (accessible field must be oui to fetch full text).
• Endpoint structure changed in mai 2026; ensure to use /api/v1/theses/... routes described here and handle potential breakage with the provided recovery procedures.
• When combining data from multiple endpoints, normalize field names on the client side after retrieval to avoid mismatches in the cartography and analysis outputs.
• SomeLucene-implied fields are non-requêtables (e.g., directeurs/rapporteurs/membresJury.nom); use the workflow PPN or full-text filtering as recommended.