Update agents/language_agent.py

agents/language_agent.py

@@ -1,8 +1,17 @@
-from transformers import pipeline
+from langchain.llms import CTransformers
+from langchain.chains import RetrievalQA
+from agents.retriever_agent import create_vectorstore
 
-# Load once at module level
-summarizer = pipeline("summarization", model="google/flan-t5-small")
+def generate_brief(question):
+    # Load a local quantized model (no key)
+    llm = CTransformers(
+        model="models/mistral-7b-instruct-v0.1.Q4_K_M.gguf",  # Or use t5-small with transformers
+        model_type="mistral",
+        config={'max_new_tokens': 512, 'temperature': 0.7}
+    )
 
-def generate_brief(text):
-    response = summarizer(text, max_length=80, min_length=20, do_sample=False)
-    return response[0]['summary_text']
+    vectordb = create_vectorstore()
+    retriever = vectordb.as_retriever()
+
+    qa_chain = RetrievalQA.from_chain_type(llm=llm, retriever=retriever)
+    return qa_chain.run(question)