bex/grammar.py

@@ -170,7 +170,7 @@ def _match_rep(child, seq, pos, min_rep):
 # Counting (for MDL scoring)
 # ---------------------------------------------------------------------------
 
-_COUNT_CAP = 10**12
+_COUNT_CAP = 10 ** 30
 
 @lru_cache(maxsize=None)
 def count_words(node, length):

bex/tag_preprocessor/analyze.py

@@ -757,8 +757,8 @@ def analyze_directory(
     crx_method='standard',
     min_structure=0.0,
     context_strategy="dir",
-    decompose=False,
-    max_seq_length=5,
+    decompose=True,
+    max_seq_length=4,
     reduce_threshold=0.15,
     split_mixed=False,
     idregex_refine=False,
@@ -1040,12 +1040,16 @@ def _parse_args(argv=None):
         help="Method to split mixed groups: first-symbol (fast, crude) or distributional (slower, smarter clustering)",
     )
     parser.add_argument(
-        "--decompose", action="store_true",
-        help="Decompose long sequences into shorter fragments before inference",
+        "--decompose", action="store_true", default=True,
+        help="Decompose long sequences into shorter fragments before inference (default: on)",
     )
     parser.add_argument(
-        "--max-seq-length", type=int, default=5,
-        help="Maximum sequence length after decomposition (default: 5)",
+        "--no-decompose", dest="decompose", action="store_false",
+        help="Disable sequence decomposition",
+    )
+    parser.add_argument(
+        "--max-seq-length", type=int, default=4,
+        help="Maximum sequence length after decomposition (default: 4)",
     )
     parser.add_argument(
         "--idregex-refine", action="store_true",