Adds tests for 'próf.' and 'bandar.'

mideind · Feb 1, 2024 · 6bcf195 · 6bcf195
1 parent 051041f
commit 6bcf195
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/src/icespeak/transcribe/__init__.py b/src/icespeak/transcribe/__init__.py
@@ -533,7 +533,7 @@ def _split_substring_types(t: str) -> Iterable[str]:
     (TOK.WORD, TOK.PERSON, TOK.ENTITY, TOK.TIMESTAMP, TOK.UNKNOWN)
 )
 # These should not be interpreted as abbreviations
-# if they aren't followed by a period
+# unless they include a period
 _IGNORED_ABBREVS = frozenset(("mið", "fim", "bandar", "mao", "próf", "tom", "mar"))
 _HYPHEN_SYMBOLS = frozenset(HYPHENS)
 

diff --git a/tests/test_transcribe.py b/tests/test_transcribe.py
@@ -473,6 +473,12 @@ def test_dt_token_transcribe_basic() -> None:
     t = "maðurinn tom fékk mar eftir strembið próf í síðustu viku"
     n = DT.token_transcribe(t)
     assert n == t
+    t = "Undirritað, próf. Jónína"
+    n = DT.token_transcribe(t)
+    assert "prófessor" in n
+    t = "Hann er bandar. ríkisborgari"
+    n = DT.token_transcribe(t)
+    assert "bandarískur" in n
 
 
 def test_dt_token_transcribe_experimental():