Spaces:
Running
Running
| # Import streamlit, langchanin, PyMuPDFLoader, dan file loadllm | |
| # PyMuPDFLoader adalah library untuk mengekstraksi, menganalisa, dan mengkonversi data dari dokumen PDF | |
| import streamlit as st | |
| from langchain.document_loaders import PyMuPDFLoader | |
| from loadllm import Loadllm | |
| from streamlit_chat import message | |
| import tempfile | |
| from langchain.embeddings import HuggingFaceEmbeddings | |
| from langchain.vectorstores import FAISS | |
| from langchain.chains import ConversationalRetrievalChain | |
| # Load model directly | |
| #from transformers import AutoModel | |
| # Path dimana hasil vectore score dari FAISS akan disimpan | |
| # FAISS (Facebook AI Similarity Search) adalah sebuah library untuk mencari embedding dalam dokumen yang serupa satu dengan yang lainnya | |
| # FAISS mempunyai algoritma yang mencari kesamaan di set vector dengan ukuran apapun | |
| # FAISS bisa mencari melalui banyak informasi dengan cepat dan memilih mereka yang penting | |
| DB_FAISS_PATH = 'vectorstore/db_faiss' | |
| class FileIngestor: | |
| def __init__(self, uploaded_file): | |
| self.uploaded_file = uploaded_file | |
| def handlefileandingest(self): | |
| with tempfile.NamedTemporaryFile(delete=False) as tmp_file: | |
| tmp_file.write(self.uploaded_file.getvalue()) | |
| tmp_file_path = tmp_file.name | |
| loader = PyMuPDFLoader(file_path=tmp_file_path) | |
| data = loader.load() | |
| # Create embeddings using Sentence Transformers | |
| # Word embedding dari dokumen akan dibuat menggunakan sentence-transformers yang disediakan HuggingFace | |
| # Transformer ini berbasis BERT dan bisa memetakan kalimat dan paragraf menjadi vector space dengan | |
| # densitas 384 dimensi | |
| embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2') | |
| # Create a FAISS vector store and save embeddings | |
| db = FAISS.from_documents(data, embeddings) | |
| db.save_local(DB_FAISS_PATH) | |
| # Load the language model | |
| # Load model Llama 2 yang telah disiapkan di file loadllm.py | |
| llm = Loadllm.load_llm() | |
| #llm = AutoModel.from_pretrained("TheBloke/Llama-2-7B-Chat-GGUF") | |
| # Create a conversational chain | |
| # Membuat chain conversation dari Llama 2 | |
| chain = ConversationalRetrievalChain.from_llm(llm=llm, retriever=db.as_retriever()) | |
| # Function for conversational chat | |
| # Memasukkan chat baru bagi Streamlit | |
| # Query adalah pertanyaan yang kita berikan, answer jawaban, dan history agar Llama mengetahui | |
| # konteks untuk percakapan kita dengan dia | |
| def conversational_chat(query): | |
| result = chain({"question": query, "chat_history": st.session_state['history']}) | |
| st.session_state['history'].append((query, result["answer"])) | |
| return result["answer"] | |
| # Initialize chat history | |
| if 'history' not in st.session_state: | |
| st.session_state['history'] = [] | |
| # Initialize messages | |
| if 'generated' not in st.session_state: | |
| st.session_state['generated'] = ["Hello ! Ask me(LLAMA2) about " + self.uploaded_file.name + " 🤗"] | |
| if 'past' not in st.session_state: | |
| st.session_state['past'] = ["Hey ! 👋"] | |
| # Create containers for chat history and user input | |
| # Buat container untuk display UI | |
| response_container = st.container() | |
| container = st.container() | |
| # User input form | |
| with container: | |
| with st.form(key='my_form', clear_on_submit=True): | |
| user_input = st.text_input("Query:", placeholder="Talk to PDF data 🧮", key='input') | |
| submit_button = st.form_submit_button(label='Send') | |
| # Jika kita mengklik tombol submit/enter dan user input telah diisi, maka conversation akan kita mulai | |
| if submit_button and user_input: | |
| output = conversational_chat(user_input) | |
| st.session_state['past'].append(user_input) | |
| st.session_state['generated'].append(output) | |
| # Display chat history | |
| if st.session_state['generated']: | |
| with response_container: | |
| for i in range(len(st.session_state['generated'])): | |
| message(st.session_state["past"][i], is_user=True, key=str(i) + '_user', avatar_style="big-smile") | |
| message(st.session_state["generated"][i], key=str(i), avatar_style="thumbs") |