Update app.py

app.py

@@ -39,7 +39,7 @@ from rule_detector import DEFAULT_RULEBASE_ALLOWED_ENTITY_TYPES, detect_by_rules
 
 DEFAULT_MODEL_ID = os.getenv("MODEL_ID", "AITeamUIT/gliner2-multi-full-21-6-base")
 DEFAULT_THRESHOLD = float(os.getenv("THRESHOLD", "0.5"))
-DEFAULT_CHUNK_CHARS = int(os.getenv("CHUNK_CHARS", "2000"))
+DEFAULT_CHUNK_CHARS = int(os.getenv("CHUNK_CHARS", "1000"))
 MAX_WIDTH = int(os.getenv("MAX_WIDTH", "30"))
 
 PII_LABEL_DESCRIPTIONS: Dict[str, str] = {
@@ -126,22 +126,21 @@ NUMERIC_MASK_LABELS = {
 try:
     from gliner2.processor import WhitespaceTokenSplitter
 
-    
     IMPROVED_PATTERN = re.compile(
         r"""
         # Scheme / www URL
         (?:(?:https?|ftp)://|www\.)[^\s<>"'`{}|\\^]*[^\s<>"'`{}|\\^.,;:!?()\[\]]
-
+    
         # Email address
         |[a-z0-9.!#$%&'*+/=?^_`{|}~-]+@(?:[a-z0-9](?:[a-z0-9-]{0,61}[a-z0-9])?\.)+[a-z]{2,63}
-
+    
         # Social/user handle
         |@[a-z0-9_](?:[a-z0-9_.-]{0,28}[a-z0-9_])?
-
+    
         # IPv4 address. Allows sentence punctuation after the IP, but avoids
         # matching inside a longer dotted numeric sequence.
         |(?:25[0-5]|2[0-4]\d|1\d\d|[1-9]?\d)(?:\.(?:25[0-5]|2[0-4]\d|1\d\d|[1-9]?\d)){3}(?=$|[^\w])(?!\.\d)
-
+    
         # IPv6 address, including compressed forms and optional zone IDs.
         |(?:
             (?:[0-9a-f]{1,4}:){7}[0-9a-f]{1,4}
@@ -154,13 +153,13 @@ try:
             |:(?:(?::[0-9a-f]{1,4}){1,7}|:)
             |(?:[0-9a-f]{1,4}:){1,7}:
         )(?:%[0-9a-z_.-]+)?(?![0-9a-f:])
-
+    
         # Strict JWT-like token
         |[a-z0-9_-]{10,}\.[a-z0-9_-]{10,}\.[a-z0-9_-]{10,}
-
+    
         # Shorter dot-separated token-like string, e.g. sk-abc.def.ghi
         |(?=[a-z0-9_.-]*[-_\d])[a-z0-9_-]{2,}(?:\.[a-z0-9_-]{2,}){2,}
-
+    
         # Bare domain / URL-like candidate. Validation belongs downstream; this
         # tokenizer only keeps plausible URL spans from being split into pieces.
         |(?<![\w@])
@@ -169,18 +168,18 @@ try:
             (?=$|/|:|[^\w-])
             (?::\d{2,5})?
             (?:/(?:[^\s<>"'`{}|\\^]*[^\s<>"'`{}|\\^.,;:!?()\[\]])?)?
-
+    
         # Letter-containing word token: must have at least one letter or
         # underscore (the [^\W\d] anchor). Freely joins via hyphen/underscore,
         # so Jean-Pierre, B-52, 30A-12345, well-known, and 3D-model all stay
         # atomic. The leading \d* allows digit-run prefixes (30A, 13th, 3D).
         |(?:\d*[^\W\d]\w*)(?:[-_]\w+)*
-
+    
         # Pure digit sequence: hyphens/underscores only join other digit groups.
         # "555-1234" stays atomic; "13-tuổi" → "13" / "-" / "tuổi" so the
         # numeric entity (AGE, DATE, …) is not fused with the surrounding word.
         |\d+(?:[-_]\d+)*
-
+    
         # Fallback: any single non-space character.
         |\S
         """,