BabyHemingway D2+D3: entities, base-rate gender resolver, rename preset, leak gate passes

2026-09-16 07:55:34 -07:00
parent 9598d0b4a7
commit 03b4a3f62c
5 changed files with 329 additions and 1 deletions
@@ -47,6 +47,16 @@ PRESETS = {
    "yarros": {"a": ("us", ["en_US", "en_CA"]),
               "b": ("eu", ["es_ES", "es_MX", "it_IT", "de_DE", "fr_FR"]),
               "share": {}, "default_share": 0.62},
+    # Hemingway's cast is genuinely multilingual -- Spanish, Italian, Cuban and American
+    # characters in the same books -- and the corpus alphabet agrees: 1,496 non-ASCII letters
+    # across 23 forms, against Yarros's 2. The F02 rule (a pool's character inventory must be
+    # a SUBSET of the corpus's) therefore permits accents here where it forbade them for
+    # Yarros, and forbidding them would strand every Spanish and Italian name in the cast.
+    # Share favours the romance-language pool because the two largest works by far -- For Whom
+    # the Bell Tolls and Islands in the Stream -- are set in Spain and the Caribbean.
+    "hemingway": {"a": ("us", ["en_US", "en_GB"]),
+                  "b": ("rom", ["es_ES", "es_MX", "it_IT", "fr_FR"]),
+                  "share": {}, "default_share": 0.45},
 }