Add text chunking and Persian language helpers

utils/persian.py

@@ -0,0 +1,45 @@
+import re
+import unicodedata
+
+# Normalize Arabic lookalikes to Persian
+_ARABIC_TO_PERSIAN = str.maketrans({
+    'ي': 'ی',  # Arabic ya  → Persian ya
+    'ك': 'ک',  # Arabic kaf → Persian kaf
+    'ى': 'ی',  # Alef maqsura → Persian ya
+})
+
+_PUPPY_RE = re.compile(r'puppy|پاپی|پاپي', re.IGNORECASE | re.UNICODE)
+
+# Rough threshold: if >15% of chars are Arabic/Persian script → Persian
+_PERSIAN_BLOCK = re.compile(r'[؀-ۿ]')
+
+
+def normalize(text: str) -> str:
+    text = unicodedata.normalize('NFC', text)
+    text = text.translate(_ARABIC_TO_PERSIAN)
+    text = text.replace('‌', '').replace('‍', '')  # ZWNJ / ZWJ
+    return text
+
+
+def is_puppy_mentioned(text: str) -> bool:
+    if not text:
+        return False
+    return bool(_PUPPY_RE.search(normalize(text)))
+
+
+def detect_language(text: str) -> str:
+    if not text:
+        return 'en'
+    persian_count = len(_PERSIAN_BLOCK.findall(text))
+    return 'fa' if persian_count > len(text) * 0.15 else 'en'
+
+
+def strip_puppy_prefix(text: str) -> str:
+    """Remove 'Puppy/پاپی' name prefix from a command-style message."""
+    cleaned = re.sub(
+        r'^[\s]*(puppy|پاپی|پاپي)[،,:\s]+',
+        '',
+        normalize(text),
+        flags=re.IGNORECASE | re.UNICODE,
+    )
+    return cleaned.strip() or text.strip()