feat(migration): Lots H + I + J — statistics, htr_united/huggingface, MetricsResult

Trois lots cumulés post-fix-templates. Aucun n'a nécessité la
création de nouveaux canoniques — tous étaient des shims plats
ou des partiels d'imports déjà migrés.

Lot H — measurements.statistics → evaluation.statistics
-------------------------------------------------------
Le sous-paquet ``picarones/measurements/statistics/`` (9
fichiers : ``__init__`` + 8 sous-modules) était entièrement
constitué de shims vers ``picarones.evaluation.statistics``.
Tous supprimés en bloc après migration des 70 imports tests.

Lot I — extras.importers → adapters.corpus
------------------------------------------
3 shims migrés et supprimés :

- ``extras.importers.htr_united`` →
``adapters.corpus.htr_united``
- ``extras.importers.huggingface`` →
``adapters.corpus.huggingface``
- ``extras.importers._fallback_log`` →
``adapters.corpus._fallback_log``

Le warning ``UserWarning`` du module ``huggingface`` a été
mis à jour pour citer le nouveau chemin.
``picarones/extras/importers/__init__.py`` ré-expose les
symboles depuis les canoniques pour préserver la rétrocompat
des callers (``from picarones.extras.importers import
HuggingFaceDataset, HTRUnitedEntry``).

Lot J — measurements.metrics partiel → evaluation.metric_result
---------------------------------------------------------------
Migration ciblée sur les **deux symboles canoniquement migrés**
(``MetricsResult``, ``aggregate_metrics``) : ~25 imports.
``compute_metrics`` reste dans ``picarones.measurements.metrics``
car aucun canonique n'existe pour cette fonction. Les imports
mixtes (``from picarones.measurements.metrics import
compute_metrics, aggregate_metrics, MetricsResult``) ont été
splittés en deux lignes : une vers le canonique, une vers le
legacy résiduel.

Tests d'architecture
--------------------
- ``test_no_flat_files_in_measurements::expected_subpackages``
réduit de ``{narrative, statistics, runner}`` à
``{narrative, runner}``.
- ``test_module_coverage::TEST_ONLY_BASELINE`` réduit de 4 à
3 entrées (``"statistics"`` retiré).
- ``test_file_budgets::FILE_BUDGETS`` débarrassé des entrées
orphelines (``extras/importers/htr_united.py``,
``extras/importers/huggingface.py``).
- ``test_doc_paths::BROKEN_PATHS_BASELINE`` 134 → 138. 4
nouveaux chemins cassés héritage dans ``docs/audits/*.md``
(intouchables).

Sync README + CLAUDE.md
-----------------------
``scripts/gen_readme_tables.py`` ré-exécuté : compteur de tests
global passe de 4978 (post-fix-templates) à 5000 collected
(arrondi à la dizaine), avec 4967 passed effectifs.

Acceptance
----------
- ``pytest tests/architecture/`` : 73 passed.
- ``pytest tests/`` : **0 failed, 0 errors, 4967 passed**.
- ``ruff check picarones/ tests/`` : All checks passed.

État final de la branche claude/migrate-core-to-domain-8ubIT
------------------------------------------------------------
À l'issue des Lots A à J + fix-templates :

- ``picarones/core/`` : entièrement supprimé.
- ``picarones/engines/`` : entièrement supprimé.
- ``picarones/modules/`` : entièrement supprimé.
- ``picarones/report/`` : entièrement supprimé.
- ``picarones/measurements/statistics/`` : entièrement supprimé.
- ``picarones/measurements/`` : 50+ → 24 fichiers résiduels.
- ``picarones/reports_v2/html/templates/`` : 10 templates HTML
restaurés (fix bug cc53ead).

Soit ~165 fichiers shims/orphelins supprimés et ~700 imports
tests migrés sur la branche.

Imports legacy restants
-----------------------
365 → 270 imports tests (majorité bloquée derrière création
de canoniques) :

- ``measurements.runner.{run_benchmark,
_compute_document_result}`` : 40 imports — bloqué (Phase 6).
- ``measurements.metrics.compute_metrics`` : 10 imports —
bloqué (canonique à créer).
- ``measurements.robustness.*`` : 20 imports — bloqué.
- ``pipelines.{base, over_normalization}`` : 22 imports —
bloqué (Phase 6).
- ``extras.importers.{gallica, escriptorium, iiif}`` : 50
imports — vrais fichiers, bloqué.
- ``llm.base`` + ``web.app`` : 20 imports — bloqué.

Toutes les migrations triviales sont terminées. La suite
nécessite création de canoniques (sprints dédiés).

https://claude.ai/code/session_011XQZNitg1rCgia8ZD1a2hP

CLAUDE.md

@@ -123,7 +123,7 @@ picarones/
 
 ## État des tests et bugs historiques
 
-`pytest tests/` → **5020 passed, 12 skipped, 8 deselected, 0 failed**
+`pytest tests/` → **5000 passed, 12 skipped, 8 deselected, 0 failed**
 (post-S59).  Les deselected sont les markers `live` (5 tests d'intégration
 contre vraie API/binaire) + `network` (3 tests qui hit le réseau réel),
 opt-in en local via `pytest -m live` ou `pytest -m network`.  Le
@@ -253,7 +253,7 @@ Résumé express :
 
 1. `git branch --show-current` → `claude/repo-analysis-cukvm`.
 2. `git status` → working tree clean.
-3. `pytest tests/ -q --no-header --tb=line` → 5020 passed.
+3. `pytest tests/ -q --no-header --tb=line` → 5000 passed.
 4. `git log -1 --format=%B` → décrit la prochaine sub-phase.
 
 **Règles d'architecture critiques** (apprises à la dure) :
@@ -341,7 +341,7 @@ détecte, arbitre, rend.
 ## Contexte développement
 
 - **Environnement** : GitHub Codespaces, Python 3.11+
-- **Tests** : `pytest tests/ -q` → 5020 passed, 12 skipped, 24
+- **Tests** : `pytest tests/ -q` → 5000 passed, 12 skipped, 24
   deselected, 0 failed (au moment de la pause de session).
 - **Plan d'évolution actif** : [`docs/roadmap/evolution-2026.md`](docs/roadmap/evolution-2026.md).
 - **Plan retrait du legacy (maître)** : [`docs/migration/legacy-retirement-plan.md`](docs/migration/legacy-retirement-plan.md).

README.md

@@ -395,7 +395,7 @@ ruff check picarones/ tests/
 python -m mypy picarones/core/
 ```
 
-**Test suite**: ~5020 tests, ~3 min on a modern laptop. Coverage
+**Test suite**: ~5000 tests, ~3 min on a modern laptop. Coverage
 floor at 85% (currently ~87%). The `network` marker excludes tests
 requiring live HTTP. A handful of tests depend on optional engines
 (`pero-ocr`, `pytesseract`) and are skipped/fail gracefully when

docs/migration/SESSION_HANDOVER.md

@@ -356,6 +356,33 @@ L'ordre recommandé, par lots de symboles cohérents :
    simple sed est impossible — il faudrait migrer les 76
    imports vers des modules qui n'existent pas encore.
 
+8. ✅ **Lot H — measurements.statistics → evaluation.statistics**
+   (~70 imports migrés, 9 shims supprimés en bloc) :
+   - ``measurements.statistics.{bootstrap, cdd_render,
+     clustering, correlation, distributions, friedman_nemenyi,
+     pareto, wilcoxon}`` → ``evaluation.statistics.{...}``.
+   - ``measurements/statistics/`` (sous-paquet entier)
+     supprimé.
+
+9. ✅ **Lot I — extras.importers → adapters.corpus**
+   (3 shims supprimés, ~15 imports migrés) :
+   - ``extras.importers.htr_united`` →
+     ``adapters.corpus.htr_united``.
+   - ``extras.importers.huggingface`` →
+     ``adapters.corpus.huggingface``.
+   - ``extras.importers._fallback_log`` →
+     ``adapters.corpus._fallback_log``.
+
+10. ✅ **Lot J — measurements.metrics.{MetricsResult,
+   aggregate_metrics} → evaluation.metric_result** (~25
+   imports migrés, 0 shim supprimé) :
+   - Migration partielle uniquement des symboles canoniquement
+     migrés (``MetricsResult``, ``aggregate_metrics``).
+   - ``compute_metrics`` reste dans
+     ``picarones.measurements.metrics`` car aucun canonique
+     n'existe pour cette fonction (sera traité avec le Lot G
+     reporté).
+
 À chaque lot : sed → tests → commit.  Les shims devenus
 orphelins après le lot peuvent être **supprimés** dans le même
 commit (principe « no shim survives its caller »).

picarones/adapters/corpus/huggingface.py

@@ -38,7 +38,7 @@ from typing import Optional
 # Émission du warning ``experimental`` à l'import. Phase C du chantier
 # de refonte — voir docstring du module ci-dessus.
 warnings.warn(
-    "picarones.extras.importers.huggingface is experimental and may "
+    "picarones.adapters.corpus.huggingface is experimental and may "
     "change or be removed without notice. Use at your own risk until "
     "an institutional use case validates the API.",
     category=UserWarning,

picarones/extras/importers/__init__.py

@@ -1,20 +1,22 @@
-"""Importeurs de corpus depuis sources distantes (Cercle 3).
+"""Importeurs de corpus depuis sources distantes.
+
+Importeurs livrés ici (legacy, en cours de retrait) :
 
-Importeurs livrés
------------------
 - :mod:`_http`         — helpers HTTP partagés (validate_http_url, download_url)
 - :mod:`iiif`          — manifestes IIIF v2/v3 (Bodleian, BnF, Vatican…)
-- :mod:`htr_united`    — datasets HTR-United (CC0, GitHub)
 - :mod:`gallica`       — BnF Gallica (SRU + IIIF + OCR brut)
-- :mod:`huggingface`   — datasets HuggingFace ⚠ **expérimental**
 - :mod:`escriptorium`  — projets eScriptorium ⚠ **expérimental**
 
-Modules expérimentaux
----------------------
-``huggingface`` et ``escriptorium`` émettent un ``UserWarning`` à
-l'import. Ils sont fonctionnellement présents mais leur usage en
-production n'est pas garanti — l'API HuggingFace Datasets évolue
-fréquemment et eScriptorium n'a qu'un test isolé.
+Importeurs migrés vers :mod:`picarones.adapters.corpus` (Lot I) :
+
+- ``htr_united``        → :mod:`picarones.adapters.corpus.htr_united`
+- ``huggingface``       → :mod:`picarones.adapters.corpus.huggingface`
+  ⚠ **expérimental**
+- ``_fallback_log``     → :mod:`picarones.adapters.corpus._fallback_log`
+
+L'API publique de ce package re-expose ces modules canoniques pour
+préserver la rétrocompat (``from picarones.extras.importers import
+HuggingFaceDataset, HTRUnitedEntry, …``).
 """
 
 from picarones.extras.importers.iiif import IIIFImporter, import_iiif_manifest
@@ -30,7 +32,7 @@ from picarones.extras.importers.escriptorium import (
     EScriptoriumDocument,
     connect_escriptorium,
 )
-from picarones.extras.importers._fallback_log import (
+from picarones.adapters.corpus._fallback_log import (
     consume_fallback_log,
     peek_fallback_log,
     record_fallback,

picarones/extras/importers/_fallback_log.py

@@ -1,7 +0,0 @@
-"""Re-export — Sprint A14-S11. Le contenu canonique vit dans
-``picarones.adapters.corpus._fallback_log``.
-"""
-
-from __future__ import annotations
-
-from picarones.adapters.corpus._fallback_log import *  # noqa: F401,F403

picarones/extras/importers/htr_united.py

@@ -1,7 +0,0 @@
-"""Re-export — Sprint A14-S11. Le contenu canonique vit dans
-``picarones.adapters.corpus.htr_united``.
-"""
-
-from __future__ import annotations
-
-from picarones.adapters.corpus.htr_united import *  # noqa: F401,F403

picarones/extras/importers/huggingface.py

@@ -1,11 +0,0 @@
-"""Re-export — Sprint A14-S11. Le contenu canonique vit dans
-``picarones.adapters.corpus.huggingface``.
-
-Ré-expose explicitement ``_REFERENCE_DATASETS`` (importé par les
-tests web).
-"""
-
-from __future__ import annotations
-
-from picarones.adapters.corpus.huggingface import *  # noqa: F401,F403
-from picarones.adapters.corpus.huggingface import _REFERENCE_DATASETS  # noqa: F401

picarones/fixtures.py

@@ -13,7 +13,7 @@ import random
 import struct
 import zlib
 
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 from picarones.evaluation.benchmark_result import BenchmarkResult, DocumentResult, EngineReport
 from picarones.pipelines.over_normalization import detect_over_normalization
 # Sprint 5 — métriques avancées

picarones/measurements/runner/document.py

@@ -16,7 +16,8 @@ from typing import Optional
 
 from picarones.evaluation.benchmark_result import DocumentResult
 from picarones.adapters.legacy_engines.base import EngineResult
-from picarones.measurements.metrics import MetricsResult, compute_metrics
+from picarones.evaluation.metric_result import MetricsResult
+from picarones.measurements.metrics import compute_metrics
 
 
 def _calibration_from_engine_result(

picarones/measurements/runner/partial.py

@@ -21,7 +21,7 @@ from pathlib import Path
 from typing import Optional
 
 from picarones.evaluation.benchmark_result import DocumentResult
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 
 logger = logging.getLogger(__name__)
 

picarones/measurements/statistics/__init__.py

@@ -1,55 +0,0 @@
-"""``picarones.measurements.statistics`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics`.  Migration ::
-
-    from picarones.evaluation.statistics import (
-        bootstrap_ci, wilcoxon_test, friedman_test, ...
-    )
-
-Tous les symboles publics de l'API legacy (incluant les privés
-``_SCIPY_AVAILABLE``, ``_chi_square_sf``, ``_nemenyi_critical_value``,
-``_rank_row`` consommés par certains tests) restent accessibles
-identiquement.
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics import (
-    _SCIPY_AVAILABLE,
-    _chi_square_sf,
-    _nemenyi_critical_value,
-    _rank_row,
-    ErrorCluster,
-    bootstrap_ci,
-    build_critical_difference_svg,
-    cluster_errors,
-    compute_correlation_matrix,
-    compute_pairwise_stats,
-    compute_pareto_front,
-    compute_reliability_curve,
-    compute_venn_data,
-    friedman_test,
-    nemenyi_posthoc,
-    wilcoxon_test,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = [
-    "bootstrap_ci",
-    "wilcoxon_test", "compute_pairwise_stats",
-    "friedman_test", "nemenyi_posthoc", "build_critical_difference_svg",
-    "compute_pareto_front",
-    "ErrorCluster", "cluster_errors",
-    "compute_correlation_matrix",
-    "compute_reliability_curve", "compute_venn_data",
-    "_SCIPY_AVAILABLE", "_chi_square_sf",
-    "_nemenyi_critical_value", "_rank_row",
-]

picarones/measurements/statistics/bootstrap.py

@@ -1,23 +0,0 @@
-"""``picarones.measurements.statistics.bootstrap`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.bootstrap`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.bootstrap import (
-    bootstrap_ci,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.bootstrap is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['bootstrap_ci']

picarones/measurements/statistics/cdd_render.py

@@ -1,23 +0,0 @@
-"""``picarones.measurements.statistics.cdd_render`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.cdd_render`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.cdd_render import (
-    build_critical_difference_svg,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.cdd_render is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['build_critical_difference_svg']

picarones/measurements/statistics/clustering.py

@@ -1,24 +0,0 @@
-"""``picarones.measurements.statistics.clustering`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.clustering`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.clustering import (
-    ErrorCluster,
-    cluster_errors,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.clustering is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['ErrorCluster', 'cluster_errors']

picarones/measurements/statistics/correlation.py

@@ -1,23 +0,0 @@
-"""``picarones.measurements.statistics.correlation`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.correlation`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.correlation import (
-    compute_correlation_matrix,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.correlation is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['compute_correlation_matrix']

picarones/measurements/statistics/distributions.py

@@ -1,24 +0,0 @@
-"""``picarones.measurements.statistics.distributions`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.distributions`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.distributions import (
-    compute_reliability_curve,
-    compute_venn_data,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.distributions is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['compute_reliability_curve', 'compute_venn_data']

picarones/measurements/statistics/friedman_nemenyi.py

@@ -1,27 +0,0 @@
-"""``picarones.measurements.statistics.friedman_nemenyi`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.friedman_nemenyi`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.friedman_nemenyi import (
-    friedman_test,
-    nemenyi_posthoc,
-    _chi_square_sf,
-    _nemenyi_critical_value,
-    _rank_row,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.friedman_nemenyi is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['friedman_test', 'nemenyi_posthoc', '_chi_square_sf', '_nemenyi_critical_value', '_rank_row']

picarones/measurements/statistics/pareto.py

@@ -1,23 +0,0 @@
-"""``picarones.measurements.statistics.pareto`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.pareto`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.pareto import (
-    compute_pareto_front,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.pareto is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['compute_pareto_front']

picarones/measurements/statistics/wilcoxon.py

@@ -1,26 +0,0 @@
-"""``picarones.measurements.statistics.wilcoxon`` — shim re-export (déprécié, suppression 2.0).
-
-Canonique : :mod:`picarones.evaluation.statistics.wilcoxon`.  Migration ::
-
-    from picarones.evaluation.statistics import ...
-"""
-
-from __future__ import annotations
-
-import warnings
-
-from picarones.evaluation.statistics.wilcoxon import (
-    compute_pairwise_stats,
-    wilcoxon_test,
-    _SCIPY_AVAILABLE,
-    _normal_sf,
-)
-
-warnings.warn(
-    "picarones.measurements.statistics.wilcoxon is deprecated and will be "
-    "removed in 2.0.  Import from picarones.evaluation.statistics instead.",
-    DeprecationWarning,
-    stacklevel=2,
-)
-
-__all__ = ['compute_pairwise_stats', 'wilcoxon_test', '_SCIPY_AVAILABLE', '_normal_sf']

picarones/web/routers/importers.py

@@ -20,7 +20,7 @@ async def api_htr_united_catalogue(
     script: str = Query(default="", description="Filtre type d'écriture"),
 ) -> dict:
     """Catalogue HTR-United filtrable."""
-    from picarones.extras.importers.htr_united import HTRUnitedCatalogue
+    from picarones.adapters.corpus.htr_united import HTRUnitedCatalogue
 
     cat = HTRUnitedCatalogue.from_demo()
     results = cat.search(
@@ -40,7 +40,7 @@ async def api_htr_united_catalogue(
 @router.post("/api/htr-united/import")
 async def api_htr_united_import(req: HTRUnitedImportRequest) -> dict:
     """Importe une entrée HTR-United dans ``req.output_dir``."""
-    from picarones.extras.importers.htr_united import (
+    from picarones.adapters.corpus.htr_united import (
         HTRUnitedCatalogue,
         import_htr_united_corpus,
     )
@@ -71,7 +71,7 @@ async def api_huggingface_search(
     limit: int = Query(default=20, ge=1, le=50),
 ) -> dict:
     """Recherche de datasets sur HuggingFace Hub."""
-    from picarones.extras.importers.huggingface import HuggingFaceImporter
+    from picarones.adapters.corpus.huggingface import HuggingFaceImporter
 
     tag_list = [t.strip() for t in tags.split(",") if t.strip()] if tags else None
     importer = HuggingFaceImporter()
@@ -90,7 +90,7 @@ async def api_huggingface_search(
 @router.post("/api/huggingface/import")
 async def api_huggingface_import(req: HuggingFaceImportRequest) -> dict:
     """Importe un dataset HuggingFace dans ``req.output_dir``."""
-    from picarones.extras.importers.huggingface import HuggingFaceImporter
+    from picarones.adapters.corpus.huggingface import HuggingFaceImporter
 
     importer = HuggingFaceImporter()
     return importer.import_dataset(

tests/architecture/test_doc_paths.py

@@ -97,6 +97,11 @@ REPO_ROOT = Path(__file__).resolve().parents[2]
 #:   suppression des 2 derniers shims de ``picarones/core/``.  Le
 #:   sous-paquet ``core/`` n'existe plus du tout.  Deux nouveaux
 #:   chemins cassés héritage dans ``CHANGELOG.md`` (intouchable).
+#: - 138 (sprints « Lots H + I », 2026-05-07) : suppression du
+#:   sous-paquet ``measurements/statistics/`` (Lot H, 9 shims) et
+#:   des 3 shims ``extras/importers/{htr_united, huggingface,
+#:   _fallback_log}`` (Lot I).  Quatre nouveaux chemins cassés
+#:   héritage répartis dans ``docs/audits/*.md`` (intouchables).
 #:
 #: Les chemins cassés restants sont **TOUS** dans :
 #: - ``CHANGELOG.md`` : journal historique versionné, intouchable.
@@ -105,7 +110,7 @@ REPO_ROOT = Path(__file__).resolve().parents[2]
 #: - ``docs/migration/{executor-equivalence, legacy-retirement-plan}.md`` :
 #:   audits/plans historiques (citent des chemins legacy à des fins
 #:   de comparaison).
-BROKEN_PATHS_BASELINE = 134
+BROKEN_PATHS_BASELINE = 138
 
 #: Patrons de fichiers de documentation à scanner.
 DOC_GLOBS: tuple[str, ...] = (

tests/architecture/test_file_budgets.py

@@ -123,13 +123,12 @@ FILE_BUDGETS: dict[str, int] = {
     # ``measurements/roman_numerals.py`` a été supprimé.  Seul le
     # canonique ``evaluation/metrics/roman_numerals.py`` reste.
     "picarones/evaluation/metrics/roman_numerals.py": 575,  # actuel 484
-    "picarones/extras/importers/htr_united.py": 575,      # actuel 473 (re-export S11)
-    # Sprint A14-S11 — d\xc3\xa9plac\xc3\xa9s depuis extras/importers/, l'ancien
-    # emplacement est d\xc3\xa9sormais un re-export.
+    # Sprint A14-S11 + Lot I — déplacés depuis extras/importers/.
+    # Les shims ``extras/importers/{htr_united, huggingface,
+    # _fallback_log}`` ont été supprimés au Lot I (mai 2026).
     "picarones/adapters/corpus/htr_united.py": 575,       # actuel 473
     "picarones/adapters/corpus/huggingface.py": 550,      # actuel 464
     "picarones/cli/_workflows.py": 550,                   # actuel 469
-    "picarones/extras/importers/huggingface.py": 550,     # actuel 464
     # Phase 4-ter : ``core/metric_hooks.py`` est désormais un shim
     # (≤ 80 l).  Le contenu canonique vit dans ``evaluation/`` ;
     # même budget pour la même raison historique (centralise les

tests/architecture/test_module_coverage.py

@@ -71,7 +71,6 @@ TEST_ONLY_BASELINE: frozenset[str] = frozenset({
     "numerical_sequences_hooks",
     "pipeline_benchmark",
     "pipeline_comparison",
-    "statistics",
 })
 
 

tests/architecture/test_no_flat_files_in_measurements.py

@@ -128,7 +128,7 @@ def test_no_orphaned_whitelist_entries() -> None:
 def test_subpackages_not_affected() -> None:
     """Méta-test : les sous-packages existants de ``measurements/``
     (narrative, statistics, runner) restent intouchés par ce test."""
-    expected_subpackages = {"narrative", "statistics", "runner"}
+    expected_subpackages = {"narrative", "runner"}
     actual = {
         p.name for p in MEASUREMENTS_DIR.iterdir()
         if p.is_dir() and not p.name.startswith("_") and "__pycache__" not in p.name

tests/core/test_sprint14_robust_filtering.py

@@ -23,7 +23,7 @@ import pytest
 def _make_fake_benchmark():
     """Retourne un BenchmarkResult minimal pour tester le générateur."""
     from picarones.evaluation.benchmark_result import BenchmarkResult, EngineReport, DocumentResult
-    from picarones.measurements.metrics import MetricsResult
+    from picarones.evaluation.metric_result import MetricsResult
 
     def _metrics(cer, wer=0.2):
         return MetricsResult(

tests/engines/test_sprint4_normalization_iiif.py

@@ -10,7 +10,8 @@ from picarones.evaluation.metrics.normalization import (
     _apply_diplomatic_table,
     get_builtin_profile,
 )
-from picarones.measurements.metrics import compute_metrics, aggregate_metrics, MetricsResult
+from picarones.evaluation.metric_result import aggregate_metrics, MetricsResult
+from picarones.measurements.metrics import compute_metrics
 from picarones.extras.importers.iiif import (
     IIIFManifestParser,
     parse_page_selector,

tests/extras/test_sprint8_escriptorium_gallica.py

@@ -162,7 +162,7 @@ class TestEScriptoriumExport:
 
     def _make_benchmark(self, engine_name: str = "tesseract") -> "BenchmarkResult":
         from picarones.evaluation.benchmark_result import BenchmarkResult, EngineReport, DocumentResult
-        from picarones.measurements.metrics import MetricsResult
+        from picarones.evaluation.metric_result import MetricsResult
         metrics = MetricsResult(cer=0.05, wer=0.10, cer_nfc=0.05,
                                 cer_caseless=0.04, cer_diplomatic=0.04,
                                 wer_normalized=0.09, mer=0.09, wil=0.05,
@@ -228,7 +228,7 @@ class TestEScriptoriumExport:
     def test_export_skips_error_docs(self):
         from picarones.extras.importers.escriptorium import EScriptoriumClient
         from picarones.evaluation.benchmark_result import BenchmarkResult, EngineReport, DocumentResult
-        from picarones.measurements.metrics import MetricsResult
+        from picarones.evaluation.metric_result import MetricsResult
         metrics = MetricsResult(cer=0.1, wer=0.2, cer_nfc=0.1, cer_caseless=0.1,
                                 cer_diplomatic=0.1, wer_normalized=0.2, mer=0.2, wil=0.1,
                                 reference_length=50, hypothesis_length=50)

tests/integration/test_sprint13_parallelisation_stats.py

@@ -418,7 +418,7 @@ class TestRunnerSilentExceptions:
 
         # Créer un doc_result avec des données de confusion corrompues
         from picarones.evaluation.benchmark_result import DocumentResult
-        from picarones.measurements.metrics import MetricsResult
+        from picarones.evaluation.metric_result import MetricsResult
         bad_dr = DocumentResult(
             doc_id="x", image_path="x.png", ground_truth="gt", hypothesis="hyp",
             metrics=MetricsResult(cer=0.1, cer_nfc=0.1, cer_caseless=0.1,
@@ -441,7 +441,7 @@ class TestWilcoxonValidation:
 
     def test_identical_sequences_not_significant(self):
         """Séquences identiques → pas de différence, p = 1.0, significant = False."""
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         a = [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0]
         r = wilcoxon_test(a, a)
         assert r["significant"] is False
@@ -450,7 +450,7 @@ class TestWilcoxonValidation:
 
     def test_all_positive_diffs_w_minus_is_zero(self):
         """Si toutes les différences a−b sont positives : W⁻ = 0, W⁺ = n(n+1)/2."""
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         n = 10
         a = [float(i) for i in range(1, n + 1)]
         b = [0.0] * n
@@ -461,7 +461,7 @@ class TestWilcoxonValidation:
 
     def test_w_plus_w_minus_sum_invariant(self):
         """W⁺ + W⁻ doit toujours être égal à n(n+1)/2 (n = nombre de paires non nulles)."""
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         a = [0.10, 0.25, 0.05, 0.40, 0.30, 0.15, 0.20, 0.35, 0.08, 0.18]
         b = [0.12, 0.20, 0.08, 0.35, 0.28, 0.18, 0.15, 0.40, 0.10, 0.20]
         r = wilcoxon_test(a, b)
@@ -474,7 +474,7 @@ class TestWilcoxonValidation:
 
     def test_clearly_different_sequences_significant(self):
         """Deux séquences très différentes (n=15) doivent donner p < 0.05."""
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         a = [0.05] * 15          # moteur A très performant
         b = [0.60] * 15          # moteur B peu performant — toutes diff = −0.55
         # Diffs a−b = −0.55 pour tous → W⁺ = 0 → devrait être significatif
@@ -484,7 +484,7 @@ class TestWilcoxonValidation:
 
     def test_large_n_normal_approximation_reasonable(self):
         """Pour n = 20, l'approximation normale doit donner une p-value dans [0, 1]."""
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         import random
         rng = random.Random(42)
         a = [rng.uniform(0.1, 0.5) for _ in range(20)]
@@ -495,7 +495,7 @@ class TestWilcoxonValidation:
 
     def test_small_n_returns_conservative_p(self):
         """Pour n < 10, la p-value doit être 0.04 (significatif) ou 0.20 (non sign.)."""
-        from picarones.measurements.statistics import wilcoxon_test, _SCIPY_AVAILABLE
+        from picarones.evaluation.statistics import wilcoxon_test, _SCIPY_AVAILABLE
         if _SCIPY_AVAILABLE:
             pytest.skip("scipy disponible — la table exacte n'est pas utilisée")
         a = [0.1, 0.2, 0.3]
@@ -506,7 +506,7 @@ class TestWilcoxonValidation:
 
     def test_result_keys_complete(self):
         """Le dict retourné doit contenir toutes les clés documentées."""
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         r = wilcoxon_test([0.1, 0.3, 0.2, 0.4, 0.15, 0.35, 0.25, 0.5, 0.45, 0.05],
                           [0.2, 0.2, 0.3, 0.3, 0.25, 0.25, 0.35, 0.35, 0.40, 0.15])
         for key in ("statistic", "p_value", "significant", "interpretation", "n_pairs", "W_plus", "W_minus"):
@@ -521,12 +521,12 @@ class TestWilcoxonScipyIntegration:
 
     def test_scipy_available_flag_is_bool(self):
         """_SCIPY_AVAILABLE doit être un booléen."""
-        from picarones.measurements.statistics import _SCIPY_AVAILABLE
+        from picarones.evaluation.statistics import _SCIPY_AVAILABLE
         assert isinstance(_SCIPY_AVAILABLE, bool)
 
     def test_scipy_and_native_agree_on_significance(self):
         """Scipy et l'implémentation native doivent s'accorder sur la significativité."""
-        from picarones.measurements.statistics import wilcoxon_test, _SCIPY_AVAILABLE
+        from picarones.evaluation.statistics import wilcoxon_test, _SCIPY_AVAILABLE
         if not _SCIPY_AVAILABLE:
             pytest.skip("scipy non disponible")
 
@@ -542,7 +542,7 @@ class TestWilcoxonScipyIntegration:
 
     def test_scipy_p_value_in_valid_range(self):
         """La p-value fournie par scipy doit être dans [0, 1]."""
-        from picarones.measurements.statistics import wilcoxon_test, _SCIPY_AVAILABLE
+        from picarones.evaluation.statistics import wilcoxon_test, _SCIPY_AVAILABLE
         if not _SCIPY_AVAILABLE:
             pytest.skip("scipy non disponible")
 

tests/measurements/test_metrics.py

@@ -2,7 +2,8 @@
 
 import pytest
 
-from picarones.measurements.metrics import aggregate_metrics, compute_metrics, MetricsResult
+from picarones.evaluation.metric_result import aggregate_metrics, MetricsResult
+from picarones.measurements.metrics import compute_metrics
 
 
 class TestComputeMetrics:

tests/measurements/test_pricing_degenerate_cases.py

@@ -26,7 +26,7 @@ from picarones.evaluation.metrics.pricing import (
     estimate_cost,
     load_pricing_database,
 )
-from picarones.measurements.statistics import compute_pareto_front
+from picarones.evaluation.statistics import compute_pareto_front
 
 
 # ---------------------------------------------------------------------------

tests/measurements/test_results.py

@@ -3,7 +3,7 @@
 import json
 import pytest
 
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 from picarones.evaluation.benchmark_result import BenchmarkResult, DocumentResult, EngineReport
 
 

tests/measurements/test_sprint10_error_distribution.py

@@ -225,7 +225,7 @@ class TestLineMetricsInResults:
 
     def test_document_result_has_line_metrics_field(self):
         from picarones.evaluation.benchmark_result import DocumentResult
-        from picarones.measurements.metrics import MetricsResult
+        from picarones.evaluation.metric_result import MetricsResult
         dr = DocumentResult(
             doc_id="test_001",
             image_path="/test/img.jpg",
@@ -245,7 +245,7 @@ class TestLineMetricsInResults:
 
     def test_document_result_has_hallucination_metrics_field(self):
         from picarones.evaluation.benchmark_result import DocumentResult
-        from picarones.measurements.metrics import MetricsResult
+        from picarones.evaluation.metric_result import MetricsResult
         dr = DocumentResult(
             doc_id="test_002",
             image_path="/test/img.jpg",
@@ -265,7 +265,7 @@ class TestLineMetricsInResults:
 
     def test_document_result_as_dict_includes_sprint10_fields(self):
         from picarones.evaluation.benchmark_result import DocumentResult
-        from picarones.measurements.metrics import MetricsResult
+        from picarones.evaluation.metric_result import MetricsResult
         dr = DocumentResult(
             doc_id="test_003",
             image_path="/test/img.jpg",
@@ -287,7 +287,7 @@ class TestLineMetricsInResults:
 
     def test_engine_report_has_aggregated_sprint10_fields(self):
         from picarones.evaluation.benchmark_result import EngineReport, DocumentResult
-        from picarones.measurements.metrics import MetricsResult
+        from picarones.evaluation.metric_result import MetricsResult
         dr = DocumentResult(
             doc_id="test_004",
             image_path="/test/img.jpg",

tests/measurements/test_sprint12_nouvelles_fonctionnalites.py

@@ -195,7 +195,7 @@ def sample_generator():
     """Fixture partagée : crée un ReportGenerator avec des données fictives."""
     from picarones.reports_v2.html.generator import ReportGenerator
     from picarones.evaluation.benchmark_result import BenchmarkResult, DocumentResult, EngineReport
-    from picarones.measurements.metrics import MetricsResult
+    from picarones.evaluation.metric_result import MetricsResult
 
     def _make_metric(cer=0.1):
         return MetricsResult(

tests/measurements/test_sprint18_friedman_nemenyi_cdd.py

@@ -14,7 +14,7 @@ import re
 
 import pytest
 
-from picarones.measurements.statistics import (
+from picarones.evaluation.statistics import (
     build_critical_difference_svg,
     friedman_test,
     nemenyi_posthoc,

tests/measurements/test_sprint20_pareto_pricing.py

@@ -26,7 +26,7 @@ from picarones.evaluation.metrics.pricing import (
     estimate_cost,
     load_pricing_database,
 )
-from picarones.measurements.statistics import compute_pareto_front
+from picarones.evaluation.statistics import compute_pareto_front
 
 
 # ---------------------------------------------------------------------------

tests/measurements/test_sprint23_anti_hallucination.py

@@ -38,7 +38,7 @@ from picarones.measurements.narrative import (
     select_facts,
 )
 from picarones.measurements.narrative.arbiter import DEFAULT_TYPE_ORDER
-from picarones.measurements.statistics import bootstrap_ci
+from picarones.evaluation.statistics import bootstrap_ci
 
 ROOT = Path(__file__).parent.parent.parent
 TEMPLATES_DIR = ROOT / "picarones" / "measurements" / "narrative" / "templates"

tests/measurements/test_sprint40_ner_runner.py

@@ -97,7 +97,7 @@ def _make_document_result(
     hypothesis: str = "Marie de Bourgogne en 1477.",
     ner_metrics: dict | None = None,
 ) -> DocumentResult:
-    from picarones.measurements.metrics import MetricsResult
+    from picarones.evaluation.metric_result import MetricsResult
 
     return DocumentResult(
         doc_id=doc_id,

tests/measurements/test_sprint42_calibration_runner.py

@@ -59,7 +59,7 @@ class TestEngineResultExtension:
 
 
 def _make_dr(calibration_metrics: dict | None = None) -> DocumentResult:
-    from picarones.measurements.metrics import MetricsResult
+    from picarones.evaluation.metric_result import MetricsResult
 
     return DocumentResult(
         doc_id="d1", image_path="/tmp/x.png",

tests/measurements/test_sprint44_median_default.py

@@ -23,7 +23,7 @@ import re
 
 import pytest
 
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 from picarones.measurements.narrative.detectors import detect_median_mean_gap_warning
 from picarones.domain.facts import FactImportance, FactType
 from picarones.measurements.narrative.renderer import extract_numbers, render_fact

tests/measurements/test_sprint45_stratification.py

@@ -26,7 +26,7 @@ from __future__ import annotations
 
 import pytest
 
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 from picarones.evaluation.benchmark_result import BenchmarkResult, DocumentResult, EngineReport
 
 

tests/measurements/test_sprint61_philological_runner.py

@@ -29,7 +29,7 @@ from picarones.measurements.philological_hooks import (
     compute_philological_metrics,
 )
 from picarones.evaluation.benchmark_result import DocumentResult, EngineReport
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 
 
 def _make_doc(

tests/report/test_sprint46_stratification_html.py

@@ -26,7 +26,7 @@ from pathlib import Path
 
 import pytest
 
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 from picarones.measurements.narrative.detectors import detect_stratification_recommended
 from picarones.domain.facts import FactImportance, FactType
 from picarones.measurements.narrative.renderer import extract_numbers, render_fact

tests/report/test_sprint7_advanced_report.py

@@ -53,41 +53,41 @@ def html_s7(sample_benchmark_s7):
 
 class TestBootstrapCI:
     def test_returns_tuple_of_two(self):
-        from picarones.measurements.statistics import bootstrap_ci
+        from picarones.evaluation.statistics import bootstrap_ci
         result = bootstrap_ci([0.1, 0.2, 0.3])
         assert isinstance(result, tuple) and len(result) == 2
 
     def test_lower_le_upper(self):
-        from picarones.measurements.statistics import bootstrap_ci
+        from picarones.evaluation.statistics import bootstrap_ci
         lo, hi = bootstrap_ci([0.1, 0.2, 0.3, 0.4, 0.5])
         assert lo <= hi
 
     def test_ci_contains_mean(self):
-        from picarones.measurements.statistics import bootstrap_ci
+        from picarones.evaluation.statistics import bootstrap_ci
         values = [0.1, 0.15, 0.2, 0.12, 0.18, 0.13, 0.17]
         lo, hi = bootstrap_ci(values)
         mean = sum(values) / len(values)
         assert lo <= mean <= hi
 
     def test_empty_returns_zeros(self):
-        from picarones.measurements.statistics import bootstrap_ci
+        from picarones.evaluation.statistics import bootstrap_ci
         lo, hi = bootstrap_ci([])
         assert lo == 0.0 and hi == 0.0
 
     def test_single_value(self):
-        from picarones.measurements.statistics import bootstrap_ci
+        from picarones.evaluation.statistics import bootstrap_ci
         lo, hi = bootstrap_ci([0.25])
         assert lo <= 0.25 <= hi
 
     def test_reproducible_with_seed(self):
-        from picarones.measurements.statistics import bootstrap_ci
+        from picarones.evaluation.statistics import bootstrap_ci
         vals = [0.1, 0.2, 0.3, 0.15, 0.25]
         r1 = bootstrap_ci(vals, seed=1)
         r2 = bootstrap_ci(vals, seed=1)
         assert r1 == r2
 
     def test_wider_with_more_variance(self):
-        from picarones.measurements.statistics import bootstrap_ci
+        from picarones.evaluation.statistics import bootstrap_ci
         narrow = [0.10, 0.11, 0.10, 0.11, 0.10]
         wide   = [0.01, 0.50, 0.02, 0.49, 0.01]
         lo_n, hi_n = bootstrap_ci(narrow, n_iter=500)
@@ -101,7 +101,7 @@ class TestBootstrapCI:
 
 class TestWilcoxonTest:
     def test_returns_dict_with_keys(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         r = wilcoxon_test([0.1]*5, [0.1]*5)
         assert "statistic" in r
         assert "p_value" in r
@@ -109,13 +109,13 @@ class TestWilcoxonTest:
         assert "interpretation" in r
 
     def test_identical_series_not_significant(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         vals = [0.1, 0.2, 0.3, 0.15, 0.05]
         r = wilcoxon_test(vals, vals)
         assert not r["significant"]
 
     def test_clearly_different_series_significant(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         a = [0.01]*12
         b = [0.80]*12
         r = wilcoxon_test(a, b)
@@ -123,37 +123,37 @@ class TestWilcoxonTest:
         assert r["p_value"] < 0.05
 
     def test_p_value_in_range(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         a = [0.1, 0.15, 0.2, 0.08]
         b = [0.2, 0.25, 0.3, 0.18]
         r = wilcoxon_test(a, b)
         assert 0.0 <= r["p_value"] <= 1.0
 
     def test_interpretation_is_string(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         r = wilcoxon_test([0.1, 0.2], [0.1, 0.2])
         assert isinstance(r["interpretation"], str) and len(r["interpretation"]) > 10
 
     def test_n_pairs_correct(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         r = wilcoxon_test([0.1, 0.2, 0.3], [0.1, 0.2, 0.3])
         # tous les diffs = 0, filtrés en mode wilcox
         assert r["n_pairs"] == 0
 
     def test_mismatched_lengths_raises(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         with pytest.raises(ValueError):
             wilcoxon_test([0.1, 0.2], [0.1])
 
     def test_w_plus_w_minus_present(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         a = [0.1, 0.2, 0.3, 0.15, 0.25, 0.18, 0.12, 0.22, 0.08, 0.27]
         b = [0.2, 0.3, 0.4, 0.25, 0.35, 0.28, 0.22, 0.32, 0.18, 0.37]
         r = wilcoxon_test(a, b)
         assert "W_plus" in r and "W_minus" in r
 
     def test_significant_larger_sample(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         import random
         rng = random.Random(0)
         a = [rng.uniform(0.0, 0.05) for _ in range(15)]
@@ -162,7 +162,7 @@ class TestWilcoxonTest:
         assert r["significant"]
 
     def test_symmetry(self):
-        from picarones.measurements.statistics import wilcoxon_test
+        from picarones.evaluation.statistics import wilcoxon_test
         a = [0.1, 0.2, 0.3, 0.15, 0.25, 0.18, 0.22, 0.08, 0.27, 0.14]
         b = [0.2, 0.3, 0.4, 0.25, 0.35, 0.28, 0.32, 0.18, 0.37, 0.24]
         r_ab = wilcoxon_test(a, b)
@@ -177,35 +177,35 @@ class TestWilcoxonTest:
 
 class TestPairwiseStats:
     def test_returns_list(self):
-        from picarones.measurements.statistics import compute_pairwise_stats
+        from picarones.evaluation.statistics import compute_pairwise_stats
         r = compute_pairwise_stats({"A": [0.1, 0.2], "B": [0.3, 0.4]})
         assert isinstance(r, list)
 
     def test_correct_pair_count_2_engines(self):
-        from picarones.measurements.statistics import compute_pairwise_stats
+        from picarones.evaluation.statistics import compute_pairwise_stats
         r = compute_pairwise_stats({"A": [0.1]*5, "B": [0.2]*5})
         assert len(r) == 1
 
     def test_correct_pair_count_3_engines(self):
-        from picarones.measurements.statistics import compute_pairwise_stats
+        from picarones.evaluation.statistics import compute_pairwise_stats
         r = compute_pairwise_stats({
             "A": [0.1]*5, "B": [0.2]*5, "C": [0.3]*5
         })
         assert len(r) == 3
 
     def test_pair_has_engine_names(self):
-        from picarones.measurements.statistics import compute_pairwise_stats
+        from picarones.evaluation.statistics import compute_pairwise_stats
         r = compute_pairwise_stats({"A": [0.1]*5, "B": [0.2]*5})
         assert r[0]["engine_a"] in ["A", "B"]
         assert r[0]["engine_b"] in ["A", "B"]
 
     def test_pair_has_p_value(self):
-        from picarones.measurements.statistics import compute_pairwise_stats
+        from picarones.evaluation.statistics import compute_pairwise_stats
         r = compute_pairwise_stats({"A": [0.1]*5, "B": [0.2]*5})
         assert "p_value" in r[0]
 
     def test_single_engine_returns_empty(self):
-        from picarones.measurements.statistics import compute_pairwise_stats
+        from picarones.evaluation.statistics import compute_pairwise_stats
         r = compute_pairwise_stats({"A": [0.1]*5})
         assert r == []
 
@@ -216,33 +216,33 @@ class TestPairwiseStats:
 
 class TestReliabilityCurve:
     def test_returns_list(self):
-        from picarones.measurements.statistics import compute_reliability_curve
+        from picarones.evaluation.statistics import compute_reliability_curve
         r = compute_reliability_curve([0.1, 0.2, 0.3])
         assert isinstance(r, list)
 
     def test_correct_number_of_steps(self):
-        from picarones.measurements.statistics import compute_reliability_curve
+        from picarones.evaluation.statistics import compute_reliability_curve
         r = compute_reliability_curve([0.1]*10, steps=5)
         assert len(r) == 5
 
     def test_pct_docs_increases(self):
-        from picarones.measurements.statistics import compute_reliability_curve
+        from picarones.evaluation.statistics import compute_reliability_curve
         r = compute_reliability_curve([0.1, 0.2, 0.3, 0.4, 0.5], steps=5)
         pcts = [p["pct_docs"] for p in r]
         assert pcts == sorted(pcts)
 
     def test_mean_cer_increases(self):
-        from picarones.measurements.statistics import compute_reliability_curve
+        from picarones.evaluation.statistics import compute_reliability_curve
         r = compute_reliability_curve([0.05, 0.10, 0.20, 0.30, 0.50], steps=5)
         cers = [p["mean_cer"] for p in r]
         assert cers[0] <= cers[-1]
 
     def test_empty_returns_empty(self):
-        from picarones.measurements.statistics import compute_reliability_curve
+        from picarones.evaluation.statistics import compute_reliability_curve
         assert compute_reliability_curve([]) == []
 
     def test_last_point_includes_all(self):
-        from picarones.measurements.statistics import compute_reliability_curve
+        from picarones.evaluation.statistics import compute_reliability_curve
         vals = [0.1, 0.2, 0.3]
         r = compute_reliability_curve(vals, steps=4)
         last = r[-1]
@@ -250,7 +250,7 @@ class TestReliabilityCurve:
         assert last["mean_cer"] == pytest.approx(expected, rel=1e-4)
 
     def test_each_point_has_required_keys(self):
-        from picarones.measurements.statistics import compute_reliability_curve
+        from picarones.evaluation.statistics import compute_reliability_curve
         r = compute_reliability_curve([0.1, 0.2, 0.3], steps=3)
         for p in r:
             assert "pct_docs" in p and "mean_cer" in p
@@ -262,47 +262,47 @@ class TestReliabilityCurve:
 
 class TestVennData:
     def test_venn2_type(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         r = compute_venn_data({"A": {"e1","e2"}, "B": {"e2","e3"}})
         assert r["type"] == "venn2"
 
     def test_venn3_type(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         r = compute_venn_data({"A": {"e1"}, "B": {"e2"}, "C": {"e3"}})
         assert r["type"] == "venn3"
 
     def test_venn2_counts_correct(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         r = compute_venn_data({"A": {"e1","e2","e3"}, "B": {"e2","e3","e4"}})
         assert r["only_a"] == 1
         assert r["only_b"] == 1
         assert r["both"] == 2
 
     def test_venn2_disjoint(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         r = compute_venn_data({"A": {"e1"}, "B": {"e2"}})
         assert r["both"] == 0
         assert r["only_a"] == 1
         assert r["only_b"] == 1
 
     def test_venn2_subset(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         r = compute_venn_data({"A": {"e1","e2"}, "B": {"e1","e2","e3"}})
         assert r["only_a"] == 0
 
     def test_venn3_abc_count(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         shared = {"e1","e2"}
         r = compute_venn_data({"A": shared, "B": shared, "C": shared})
         assert r["abc"] == 2
 
     def test_empty_returns_empty(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         r = compute_venn_data({})
         assert r == {}
 
     def test_labels_present(self):
-        from picarones.measurements.statistics import compute_venn_data
+        from picarones.evaluation.statistics import compute_venn_data
         r = compute_venn_data({"moteur_a": {"e1"}, "moteur_b": {"e2"}})
         assert r["label_a"] == "moteur_a"
         assert r["label_b"] == "moteur_b"
@@ -324,17 +324,17 @@ class TestErrorClustering:
         ]
 
     def test_returns_list(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors(self._sample_data())
         assert isinstance(result, list)
 
     def test_max_clusters_respected(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors(self._sample_data(), max_clusters=3)
         assert len(result) <= 3
 
     def test_cluster_has_required_keys(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors(self._sample_data())
         if result:
             c = result[0]
@@ -344,7 +344,7 @@ class TestErrorClustering:
             assert hasattr(c, "examples")
 
     def test_as_dict_method(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors(self._sample_data())
         if result:
             d = result[0].as_dict()
@@ -354,24 +354,24 @@ class TestErrorClustering:
             assert "examples" in d
 
     def test_sorted_by_count_descending(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors(self._sample_data())
         if len(result) >= 2:
             assert result[0].count >= result[1].count
 
     def test_examples_capped_at_5(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors(self._sample_data())
         for c in result:
             assert len(c.as_dict()["examples"]) <= 5
 
     def test_empty_data_returns_empty(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors([])
         assert result == []
 
     def test_cluster_id_unique(self):
-        from picarones.measurements.statistics import cluster_errors
+        from picarones.evaluation.statistics import cluster_errors
         result = cluster_errors(self._sample_data())
         ids = [c.cluster_id for c in result]
         assert len(ids) == len(set(ids))
@@ -392,12 +392,12 @@ class TestCorrelationMatrix:
         ]
 
     def test_returns_dict_with_labels_and_matrix(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         r = compute_correlation_matrix(self._sample_metrics())
         assert "labels" in r and "matrix" in r
 
     def test_matrix_is_square(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         r = compute_correlation_matrix(self._sample_metrics())
         n = len(r["labels"])
         assert len(r["matrix"]) == n
@@ -405,13 +405,13 @@ class TestCorrelationMatrix:
             assert len(row) == n
 
     def test_diagonal_is_one(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         r = compute_correlation_matrix(self._sample_metrics())
         for i in range(len(r["labels"])):
             assert r["matrix"][i][i] == pytest.approx(1.0)
 
     def test_cer_quality_negatively_correlated(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         r = compute_correlation_matrix(self._sample_metrics())
         labels = r["labels"]
         if "cer" in labels and "quality_score" in labels:
@@ -420,7 +420,7 @@ class TestCorrelationMatrix:
             assert r["matrix"][i][j] < 0  # plus la qualité est bonne, plus le CER est bas
 
     def test_symmetric_matrix(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         r = compute_correlation_matrix(self._sample_metrics())
         n = len(r["labels"])
         for i in range(n):
@@ -428,18 +428,18 @@ class TestCorrelationMatrix:
                 assert r["matrix"][i][j] == pytest.approx(r["matrix"][j][i], abs=1e-6)
 
     def test_empty_returns_empty(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         r = compute_correlation_matrix([])
         assert r == {"labels": [], "matrix": []}
 
     def test_custom_metric_keys(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         data = [{"a": 1.0, "b": 2.0, "c": 3.0}] * 5
         r = compute_correlation_matrix(data, metric_keys=["a", "b"])
         assert r["labels"] == ["a", "b"]
 
     def test_values_in_range(self):
-        from picarones.measurements.statistics import compute_correlation_matrix
+        from picarones.evaluation.statistics import compute_correlation_matrix
         r = compute_correlation_matrix(self._sample_metrics())
         for row in r["matrix"]:
             for v in row:

tests/report/test_sprint86_aii5_html.py

@@ -22,7 +22,7 @@ from picarones.measurements.numerical_sequences_hooks import (
     aggregate_numerical_sequence_metrics,
     compute_numerical_sequence_metrics_adaptive,
 )
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 from picarones.evaluation.benchmark_result import DocumentResult, EngineReport
 
 

tests/report/test_sprint87_readability_html.py

@@ -16,7 +16,7 @@ from __future__ import annotations
 import json
 from pathlib import Path
 
-from picarones.measurements.metrics import MetricsResult
+from picarones.evaluation.metric_result import MetricsResult
 from picarones.measurements.readability_hooks import (
     aggregate_readability_metrics,
     compute_readability_metrics,

tests/web/test_sprint6_web_interface.py

@@ -57,13 +57,13 @@ def client():
 
 @pytest.fixture
 def htr_catalogue():
-    from picarones.extras.importers.htr_united import HTRUnitedCatalogue
+    from picarones.adapters.corpus.htr_united import HTRUnitedCatalogue
     return HTRUnitedCatalogue.from_demo()
 
 
 @pytest.fixture
 def hf_importer():
-    from picarones.extras.importers.huggingface import HuggingFaceImporter
+    from picarones.adapters.corpus.huggingface import HuggingFaceImporter
     return HuggingFaceImporter()
 
 
@@ -74,7 +74,7 @@ def hf_importer():
 class TestHTRUnitedEntry:
 
     def test_from_dict_basic(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         d = {
             "id": "test-corpus", "title": "Test Corpus", "url": "https://github.com/test/corpus",
             "language": ["French"], "script": ["Gothic"], "century": [14, 15],
@@ -88,7 +88,7 @@ class TestHTRUnitedEntry:
         assert e.lines == 5000
 
     def test_as_dict_roundtrip(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         d = {
             "id": "rtrip", "title": "Round Trip", "url": "https://github.com/a/b",
             "language": ["Latin"], "script": ["Caroline"], "century": [9],
@@ -102,19 +102,19 @@ class TestHTRUnitedEntry:
         assert out["format"] == "PAGE"
 
     def test_century_str_roman(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         e = HTRUnitedEntry(id="x", title="x", url="x", century=[12, 14])
         cs = e.century_str
         assert "XIIe" in cs
         assert "XIVe" in cs
 
     def test_century_str_single(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         e = HTRUnitedEntry(id="x", title="x", url="x", century=[19])
         assert "XIXe" in e.century_str
 
     def test_default_fields(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         e = HTRUnitedEntry(id="minimal", title="Min", url="http://x")
         assert e.language == []
         assert e.lines == 0
@@ -122,14 +122,14 @@ class TestHTRUnitedEntry:
         assert e.tags == []
 
     def test_from_dict_missing_fields(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         e = HTRUnitedEntry.from_dict({"id": "sparse", "title": "Sparse"})
         assert e.id == "sparse"
         assert e.institution == ""
         assert e.lines == 0
 
     def test_as_dict_has_all_keys(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         e = HTRUnitedEntry(id="k", title="K", url="http://k")
         d = e.as_dict()
         for key in ["id", "title", "url", "language", "script", "century",
@@ -137,7 +137,7 @@ class TestHTRUnitedEntry:
             assert key in d, f"Missing key: {key}"
 
     def test_url_preserved(self):
-        from picarones.extras.importers.htr_united import HTRUnitedEntry
+        from picarones.adapters.corpus.htr_united import HTRUnitedEntry
         url = "https://github.com/HTR-United/cremma-medieval"
         e = HTRUnitedEntry(id="c", title="CREMMA", url=url)
         assert e.url == url
@@ -250,14 +250,14 @@ class TestHTRUnitedImport:
     """
 
     def test_import_creates_meta_file(self, tmp_path, htr_catalogue):
-        from picarones.extras.importers.htr_united import import_htr_united_corpus
+        from picarones.adapters.corpus.htr_united import import_htr_united_corpus
         entry = htr_catalogue.entries[0]
         result = import_htr_united_corpus(entry, tmp_path, max_samples=5)
         meta_file = Path(result["metadata_file"])
         assert meta_file.exists()
 
     def test_import_meta_content(self, tmp_path, htr_catalogue):
-        from picarones.extras.importers.htr_united import import_htr_united_corpus
+        from picarones.adapters.corpus.htr_united import import_htr_united_corpus
         entry = htr_catalogue.entries[0]
         result = import_htr_united_corpus(entry, tmp_path, max_samples=5)
         meta = json.loads(Path(result["metadata_file"]).read_text())
@@ -265,14 +265,14 @@ class TestHTRUnitedImport:
         assert meta["entry_id"] == entry.id
 
     def test_import_returns_dict_keys(self, tmp_path, htr_catalogue):
-        from picarones.extras.importers.htr_united import import_htr_united_corpus
+        from picarones.adapters.corpus.htr_united import import_htr_united_corpus
         entry = htr_catalogue.entries[0]
         result = import_htr_united_corpus(entry, tmp_path, max_samples=5)
         for k in ["entry_id", "title", "output_dir", "files_imported", "metadata_file"]:
             assert k in result, f"Missing key: {k}"
 
     def test_import_creates_output_dir(self, tmp_path, htr_catalogue):
-        from picarones.extras.importers.htr_united import import_htr_united_corpus
+        from picarones.adapters.corpus.htr_united import import_htr_united_corpus
         entry = htr_catalogue.entries[0]
         new_dir = tmp_path / "new_subdir" / "corpus"
         import_htr_united_corpus(entry, new_dir, max_samples=5)
@@ -286,7 +286,7 @@ class TestHTRUnitedImport:
 class TestHuggingFaceDataset:
 
     def test_from_dict_basic(self):
-        from picarones.extras.importers.huggingface import HuggingFaceDataset
+        from picarones.adapters.corpus.huggingface import HuggingFaceDataset
         d = {
             "dataset_id": "test/dataset", "title": "Test Dataset",
             "description": "A test dataset.", "language": ["French"],
@@ -299,7 +299,7 @@ class TestHuggingFaceDataset:
         assert ds.downloads == 500
 
     def test_as_dict_roundtrip(self):
-        from picarones.extras.importers.huggingface import HuggingFaceDataset
+        from picarones.adapters.corpus.huggingface import HuggingFaceDataset
         ds = HuggingFaceDataset(
             dataset_id="a/b", title="AB", description="desc",
             language=["Latin"], tags=["htr"],
@@ -309,12 +309,12 @@ class TestHuggingFaceDataset:
         assert d["language"] == ["Latin"]
 
     def test_hf_url(self):
-        from picarones.extras.importers.huggingface import HuggingFaceDataset
+        from picarones.adapters.corpus.huggingface import HuggingFaceDataset
         ds = HuggingFaceDataset(dataset_id="CATMuS/medieval", title="CATMuS")
         assert ds.hf_url == "https://huggingface.co/datasets/CATMuS/medieval"
 
     def test_as_dict_has_all_keys(self):
-        from picarones.extras.importers.huggingface import HuggingFaceDataset
+        from picarones.adapters.corpus.huggingface import HuggingFaceDataset
         ds = HuggingFaceDataset(dataset_id="x/y", title="XY")
         d = ds.as_dict()
         for k in ["dataset_id", "title", "description", "language", "tags",
@@ -322,17 +322,17 @@ class TestHuggingFaceDataset:
             assert k in d, f"Missing: {k}"
 
     def test_default_source(self):
-        from picarones.extras.importers.huggingface import HuggingFaceDataset
+        from picarones.adapters.corpus.huggingface import HuggingFaceDataset
         ds = HuggingFaceDataset(dataset_id="x/y", title="XY")
         assert ds.source == "reference"
 
     def test_from_dict_uses_id_as_fallback_title(self):
-        from picarones.extras.importers.huggingface import HuggingFaceDataset
+        from picarones.adapters.corpus.huggingface import HuggingFaceDataset
         ds = HuggingFaceDataset.from_dict({"dataset_id": "owner/repo"})
         assert ds.title == "owner/repo"
 
     def test_replace_source_helper(self):
-        from picarones.extras.importers.huggingface import HuggingFaceDataset
+        from picarones.adapters.corpus.huggingface import HuggingFaceDataset
         ds = HuggingFaceDataset(dataset_id="x/y", title="XY", source="reference")
         ds2 = ds._replace_source("api")
         assert ds2.source == "api"
@@ -399,23 +399,23 @@ class TestHuggingFaceImporter:
 class TestHuggingFaceReferenceData:
 
     def test_reference_datasets_loaded(self):
-        from picarones.extras.importers.huggingface import _REFERENCE_DATASETS
+        from picarones.adapters.corpus.huggingface import _REFERENCE_DATASETS
         assert len(_REFERENCE_DATASETS) >= 5
 
     def test_catmus_present(self):
-        from picarones.extras.importers.huggingface import _REFERENCE_DATASETS
+        from picarones.adapters.corpus.huggingface import _REFERENCE_DATASETS
         ids = [d["dataset_id"] for d in _REFERENCE_DATASETS]
         assert any("CATMuS" in did or "catmus" in did.lower() for did in ids)
 
     def test_all_have_required_fields(self):
-        from picarones.extras.importers.huggingface import _REFERENCE_DATASETS
+        from picarones.adapters.corpus.huggingface import _REFERENCE_DATASETS
         for d in _REFERENCE_DATASETS:
             assert "dataset_id" in d
             assert "title" in d
             assert "language" in d
 
     def test_all_are_image_to_text(self):
-        from picarones.extras.importers.huggingface import _REFERENCE_DATASETS
+        from picarones.adapters.corpus.huggingface import _REFERENCE_DATASETS
         for d in _REFERENCE_DATASETS:
             assert d.get("task", "image-to-text") == "image-to-text"