Update main.py

main.py

@@ -2,13 +2,11 @@ from fastapi import FastAPI, File, UploadFile, Form
 import os
 import shutil
 import tempfile
-from langchain_community.document_loaders import PyMuPDFLoader
-from LoadLLM import Loadllm
-from langchain_community.embeddings import HuggingFaceEmbeddings
-from langchain_community.vectorstores import FAISS
-from langchain.chains import ConversationalRetrievalChain
+from langchain.document_loaders import PyPDFLoader
+from langchain_community.llms import CTransformers
+from langchain.chains import LLMChain
+from langchain.prompts import PromptTemplate
 
-DB_FAISS_PATH = 'vectorstore/db_faiss'
 
 app = FastAPI()
 
@@ -25,23 +23,19 @@ async def PromptLLM(file: UploadFile = File(...), query: str = Form(...)):
         shutil.copyfileobj(file.file, f)
     
 
-    loader = PyMuPDFLoader(file_path= temp_file_path)
+    loader = PyPDFLoader(temp_file_path)
     data = loader.load()
 
-    # Create embeddings using Sentence Transformers
-    embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2')
+    llm = CTransformers(model="llama-2-7b-chat.ggmlv3.q4_1.bin", model_type="llama",
+                            config={'max_new_tokens': 1024,  'context_length': 2048, 'temperature': 0.01})
 
-    # Create a FAISS vector store and save embeddings
-    db = FAISS.from_documents(data, embeddings)
-    db.save_local(DB_FAISS_PATH)
+        template = """Summarise the report {pages} 
+        """
+        prompt_template = PromptTemplate(input_variables=["pages"], template=template)
+        chain = LLMChain(llm=llm, prompt=prompt_template)
 
-    # Load the language model
-    llm = Loadllm.load_llm()
 
-    # Create a conversational chain
-    chain = ConversationalRetrievalChain.from_llm(llm=llm, retriever=db.as_retriever())
-
-    result = chain({"question": query, "chat_history": ''})
-    return result['answer']
+        result = chain.run(pages=data[0].page_content)
+    return result