more chunkinh

app.py

@@ -14,8 +14,7 @@ class SummarizationRequest(BaseModel):
 class SummarizationResponse(BaseModel):
     summary: str
 
-
-def chunk_text(text, max_tokens=900):
+def chunk_text(text, max_tokens=800):
     tokens = tokenizer.encode(text, truncation=False)
     chunks = []
 
@@ -25,26 +24,25 @@ def chunk_text(text, max_tokens=900):
 
     return chunks
 
-
 @app.post("/summarize", response_model=SummarizationResponse)
 async def summarize_text(request: SummarizationRequest):
     chunks = chunk_text(request.inputs)
-    
     summaries = []
 
     for chunk in chunks:
-        input_length = len(chunk.split())
-        max_length = min(250, max(100, int(input_length * 0.4)))
-        min_length = min(100, max(50, int(input_length * 0.2)))
-
-        summary = summarizer(chunk, max_length=max_length, min_length=min_length, do_sample=False)
+        # Explicitly truncate inputs in pipeline
+        summary = summarizer(
+            chunk, 
+            max_length=150,          # safer summarization lengths
+            min_length=30, 
+            truncation=True,         # crucial addition!
+            do_sample=False
+        )
         summaries.append(summary[0]["summary_text"])
 
     final_summary = " ".join(summaries)
-
     return {"summary": final_summary}
 
-
 @app.get("/")
 def greet_json():
     return {"message": "BART Summarizer API is running"}