Update utils/translator.py

utils/translator.py

@@ -17,7 +17,7 @@ except OSError:
     spacy.cli.download("en_core_web_sm")
     nlp = spacy.load("en_core_web_sm")
 
-def split_into_chunks(text, max_chunk_len=500):
+def split_into_chunks(text, max_chunk_len=300):
     """
     Split input into sentence-aware chunks using spaCy.
     """
@@ -39,14 +39,16 @@ def split_into_chunks(text, max_chunk_len=500):
 
 def translate_text(text):
     """
-    Translate contract text from English to Brazilian Portuguese using Hugging Face model.
-    Sentence-level chunking and timing included.
+    Translate contract from English to Brazilian Portuguese using Hugging Face model.
+    Includes chunking, cleanup, and progress timing.
     """
     import time
 
     if not text.strip():
         return "No input to translate."
 
+    # Clean input
+    text = text.replace("\n", " ").replace("  ", " ").strip()
     chunks = split_into_chunks(text)
     translated_chunks = []