added chunking

app.py

@@ -1,9 +1,12 @@
 from fastapi import FastAPI
 from pydantic import BaseModel
-from transformers import pipeline
+from transformers import pipeline, AutoTokenizer
 
 app = FastAPI()
-summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+
+model_name = "facebook/bart-large-cnn"
+summarizer = pipeline("summarization", model=model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
 
 class SummarizationRequest(BaseModel):
     inputs: str
@@ -11,15 +14,36 @@ class SummarizationRequest(BaseModel):
 class SummarizationResponse(BaseModel):
     summary: str
 
+
+def chunk_text(text, max_tokens=900):
+    tokens = tokenizer.encode(text, truncation=False)
+    chunks = []
+
+    for i in range(0, len(tokens), max_tokens):
+        chunk = tokens[i:i + max_tokens]
+        chunks.append(tokenizer.decode(chunk, skip_special_tokens=True))
+
+    return chunks
+
+
 @app.post("/summarize", response_model=SummarizationResponse)
 async def summarize_text(request: SummarizationRequest):
-    input_length = len(request.inputs.split())
+    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)
+        summaries.append(summary[0]["summary_text"])
+
+    final_summary = " ".join(summaries)
 
-    max_length = min(250, max(100, int(input_length * 0.4)))
-    min_length = min(100, max(50, int(input_length * 0.2)))
+    return {"summary": final_summary}
 
-    summary = summarizer(request.inputs, max_length=max_length, min_length=min_length, do_sample=False)
-    return {"summary": summary[0]["summary_text"]}
 
 @app.get("/")
 def greet_json():