Add pure correction and pedagogy logic with full test coverage

scripts/correction/: AnswerValidator (multiple choice, sentence
correction via tap-the-faulty-word + retype) and DictationAnalyzer
build the CorrectionResult structure from cahier des charges §6.
Dictation analysis aligns words with WordAligner (Wagner-Fischer edit
distance) before any character comparison, so an omitted or added
word doesn't cascade into false errors on every following word.
TextNormalizer handles non-breaking spaces, typographic apostrophes,
and decomposed-vs-composed French accented characters via an
explicit table (GDScript has no built-in NFC/NFD normalization).
Error classification follows the data annotations only (rule_id
present -> grammatical, absent -> unclassified), never a runtime
linguistic guess.

scripts/pedagogy/: DifficultyManager (per-skill difficulty from
recent results), AttemptFlow (3-attempt graduated aid: hint, then
explanation, then reveal, matching exercise-lifecycle.md's second
gap), and ReviewScheduler (the five mastery states from §9, with
session-counted spaced review intervals [1, 3, 6, 12] rather than
wall-clock dates, since a family device's clock isn't trustworthy).

All of it is pure: no scene tree, no clock or RNG read directly,
every effectful input (session number, attempt outcome) is a
parameter. 65 headless tests across both layers, including the
word-omission-doesn't-cascade regression and the full
decouverte -> acquise -> a_reviser -> acquise path.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

scripts/pedagogy/difficulty_manager.gd

@@ -0,0 +1,35 @@
+class_name DifficultyManager
+extends RefCounted
+## Suggere le niveau de difficulte (1 a 3) du prochain exercice pour une
+## competence, a partir de l'historique recent de reussites. Code pur : ne lit
+## ni horloge ni hasard, ne modifie aucun etat (ProgressManager est responsable
+## du stockage).
+
+const MIN_DIFFICULTY := 1
+const MAX_DIFFICULTY := 3
+const WINDOW := 3
+
+
+## recent_results : resultats des tentatives les plus recentes, la derniere en
+## dernier. Monte d'un niveau apres WINDOW reussites consecutives ; descend
+## d'un niveau apres WINDOW echecs consecutifs ; ne bouge pas sinon. Jamais en
+## dehors de [MIN_DIFFICULTY, MAX_DIFFICULTY] : la difficulte evolue sans
+## jamais bloquer durablement l'enfant (cahier des charges §2).
+static func next_difficulty(current_difficulty: int, recent_results: Array[bool]) -> int:
+    if recent_results.size() < WINDOW:
+        return clampi(current_difficulty, MIN_DIFFICULTY, MAX_DIFFICULTY)
+
+    var last_window: Array[bool] = recent_results.slice(recent_results.size() - WINDOW)
+    var all_success := true
+    var all_failure := true
+    for result in last_window:
+        if result:
+            all_failure = false
+        else:
+            all_success = false
+
+    if all_success:
+        return clampi(current_difficulty + 1, MIN_DIFFICULTY, MAX_DIFFICULTY)
+    if all_failure:
+        return clampi(current_difficulty - 1, MIN_DIFFICULTY, MAX_DIFFICULTY)
+    return clampi(current_difficulty, MIN_DIFFICULTY, MAX_DIFFICULTY)