Write tests for data extraction

tests/test_extract.py

@@ -468,8 +468,8 @@ def test_extract(
 ⓢ B a r e y r e ⎵ ⎵ ⓕ J e a n ⎵ ⎵ ⓑ 2 8 . 3 . 1 1
 ⓢ R o u s s y ⎵ ⎵ ⓕ J e a n ⎵ ⎵ ⓑ 4 . 1 1 . 1 4
 ⓢ M a r i n ⎵ ⎵ ⓕ M a r c e l ⎵ ⎵ ⓑ 1 0 . 8 . 0 6
-ⓢ A m i c a l ⎵ ⎵ ⓕ E l o i ⎵ ⎵ ⓑ 1 1 . 1 0 . 0 4
-ⓢ B i r o s ⎵ ⎵ ⓕ M a e l ⎵ ⎵ ⓑ 3 0 . 1 0 . 1 0"""
+ⓢ R o q u e s ⎵ ⎵ ⓕ E l o i ⎵ ⎵ ⓑ 1 1 . 1 0 . 0 4
+ⓢ G i r o s ⎵ ⎵ ⓕ P a u l ⎵ ⎵ ⓑ 3 0 . 1 0 . 1 0"""
 
     # Transcriptions with worker version are in lowercase
     if transcription_entities_worker_version:
@@ -486,22 +486,20 @@ def test_extract(
             "⎵", expected_language_corpus
         )
 
-    assert (
-        output / "language_model" / "corpus.txt"
-    ).read_text() == expected_language_corpus
+    assert (output / "language_corpus.txt").read_text() == expected_language_corpus
 
     # Check "language_tokens.txt"
     expected_language_tokens = [
         t if t != " " else "⎵" for t in sorted(list(expected_charset))
     ]
     expected_language_tokens.append("◌")
-    assert (output / "language_model" / "tokens.txt").read_text() == "\n".join(
+    assert (output / "language_tokens.txt").read_text() == "\n".join(
         expected_language_tokens
     )
 
     # Check "language_lexicon.txt"
     expected_language_lexicon = [f"{t} {t}" for t in expected_language_tokens]
-    assert (output / "language_model" / "lexicon.txt").read_text() == "\n".join(
+    assert (output / "language_lexicon.txt").read_text() == "\n".join(
         expected_language_lexicon
     )