Generate provisional dictation audio so dictée missions can be played

Manual testing confirmed the "La clairière de l'écho" dictée mission
was unplayable: no audio file existed yet (decisions.md point 8), so
there was nothing to listen to before transcribing the sentence.

Adds tools/generate_dictation_audio.py (offline eSpeak NG synthesis +
ffmpeg encoding to Ogg Vorbis, exposed as `make
generate-dictation-audio`) and uses it to generate
ce1_accord_adjectif_dictee_01.ogg. A new headless test
(tests/data/test_dictation_audio.gd) guards against future dictées
being added without regenerating their audio.

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

tests/data/test_dictation_audio.gd

@@ -0,0 +1,26 @@
+extends GutTest
+## L'absence de fichier audio pour une dictee bloque tout test manuel de ce
+## type d'exercice (voir docs/specs/decisions.md, point 19) : ce test garantit
+## que chaque dictee marquee `audio_provisional: true` a reellement un fichier
+## `.ogg` chargeable, pas seulement une entree de donnees qui le promet.
+
+
+func test_every_provisional_dictation_has_a_loadable_audio_file() -> void:
+    var dictation_paths := DataLoader.find_json_files("res://data/dictations")
+    assert_false(dictation_paths.is_empty(), "il doit y avoir au moins une dictee a verifier")
+
+    for path: String in dictation_paths:
+        var dictation: Dictionary = DataLoader.read_json(path)
+        if not dictation.get("audio_provisional", false):
+            continue
+
+        var audio_id: String = dictation["audio_id"]
+        var audio_path := "res://assets/audio/dictations/%s.ogg" % audio_id
+        assert_true(
+            ResourceLoader.exists(audio_path),
+            "audio manquant pour %s : %s" % [audio_id, audio_path],
+        )
+        var stream: Variant = load(audio_path)
+        assert_true(
+            stream is AudioStream, "le fichier charge pour %s doit etre un AudioStream" % audio_id
+        )

tests/data/test_dictation_audio.gd.uid

@@ -0,0 +1 @@
+uid://cjjn1kkkc0is6