Update app.py

app.py

@@ -10,22 +10,10 @@ app = FastAPI()
 
 # Load models
 model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
-question_model = "deepset/tinyroberta-squad2"
-nlp = pipeline('question-answering', model=question_model, tokenizer=question_model)
 
 summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
 
 
-class ModifyQueryRequest_v3(BaseModel):
-    query_string_list: List[str]
-
-
-class T5QuestionRequest(BaseModel):
-    context: str
-
-class T5Response(BaseModel):
-    answer: str
-
 # API endpoints
 @app.post("/modify_query")
 async def modify_query(request: Request):
@@ -46,44 +34,13 @@ async def modify_query_v3(request: Request):
     except Exception as e:
         raise HTTPException(status_code=500, detail=f"Error in modifying query v3: {str(e)}")
 
-@app.post("/answer_question")
-async def answer_question(request: Request):
-    try:
-        raw_data = await request.json()
-        res_locs = []
-        context_string = ''
-        corpus_embeddings = model.encode(raw_data['context'], convert_to_tensor=True)
-        query_embeddings = model.encode(raw_data['question'], convert_to_tensor=True)
-        hits = util.semantic_search(query_embeddings, corpus_embeddings)
-        
-        # Collect relevant contexts
-        for hit in hits[0]:
-            if hit['score'] > 0.4:
-                loc = hit['corpus_id']
-                res_locs.append(raw_data['locations'][loc])
-                context_string += raw_data['context'][loc] + ' '
-        
-        # If no relevant contexts are found
-        if not res_locs:
-            answer = "Sorry, I couldn't find any results for your query. Please try again!"
-        else:
-            # Use the question-answering pipeline
-            QA_input = {
-                'question': raw_data['question'],
-                'context': context_string.replace('\n', ' ')
-            }
-            result = nlp(QA_input)
-            answer = result['answer']
-        
-        return JSONResponse(content={'answer':answer, "location":res_locs})
-    except Exception as e:
-        raise HTTPException(status_code=500, detail=f"Error in answering question: {str(e)}")
 
-@app.post("/t5answer", response_model=T5Response)
-async def t5answer(request: T5QuestionRequest):
+@app.post("/makeanswer")
+async def makeAnswer(request: Request):
     try:
         # Summarize the context
-        response = summarizer(request.context, max_length=130, min_length=30, do_sample=False)
+        raw_data = await request.json()
+        response = summarizer(raw_data['context'], max_length=130, min_length=30, do_sample=False)
         return T5Response(answer=response[0]["summary_text"])
     except Exception as e:
         raise HTTPException(status_code=500, detail=f"Error in T5 summarization: {str(e)}")