Spaces:
Sleeping
Sleeping
| import streamlit as st | |
| from dotenv import load_dotenv | |
| # from langchain.text_splitter import CharacterTextSplitter, RecursiveCharacterTextSplitter | |
| # from langchain.vectorstores import FAISS | |
| # from langchain.embeddings import HuggingFaceEmbeddings # General embeddings from HuggingFace models. | |
| from langchain.memory import ConversationBufferMemory | |
| from langchain.chains import ConversationalRetrievalChain | |
| from htmlTemplates import css, bot_template, user_template | |
| # from langchain.llms import LlamaCpp # For loading transformer models. | |
| # from langchain.document_loaders import PyPDFLoader, TextLoader, JSONLoader, CSVLoader | |
| # ํ ์คํธ ์คํ๋ฆฌํฐ | |
| from langchain_text_splitters import CharacterTextSplitter, RecursiveCharacterTextSplitter | |
| # ๋ฒกํฐ์คํ ์ด/์๋ฒ ๋ฉ/LLM | |
| from langchain_community.vectorstores import FAISS | |
| from langchain_community.embeddings import HuggingFaceEmbeddings | |
| # ๋ก๋๋ค (pebblo/pwd ๋๋ ค์ค์ง ์๊ฒ ์๋ธ๋ชจ๋๋ก) | |
| from langchain_community.document_loaders.pdf import PyPDFLoader | |
| from langchain_community.document_loaders.text import TextLoader | |
| from langchain_community.document_loaders.csv_loader import CSVLoader | |
| from langchain_community.document_loaders.json_loader import JSONLoader | |
| import tempfile # ์์ ํ์ผ์ ์์ฑํ๊ธฐ ์ํ ๋ผ์ด๋ธ๋ฌ๋ฆฌ์ ๋๋ค. | |
| import os | |
| import json | |
| from langchain.docstore.document import Document | |
| from langchain_groq import ChatGroq | |
| # PDF ๋ฌธ์๋ก๋ถํฐ ํ ์คํธ๋ฅผ ์ถ์ถํ๋ ํจ์ | |
| def get_pdf_text(pdf_docs): | |
| temp_dir = tempfile.TemporaryDirectory() | |
| temp_filepath = os.path.join(temp_dir.name, pdf_docs.name) | |
| with open(temp_filepath, "wb") as f: | |
| f.write(pdf_docs.getvalue()) | |
| pdf_loader = PyPDFLoader(temp_filepath) | |
| pdf_doc = pdf_loader.load() | |
| return pdf_doc | |
| # txt ํ์ผ๋ก๋ถํฐ text ์ถ์ถ | |
| def get_text_file(txt_docs): | |
| temp_dir = tempfile.TemporaryDirectory() | |
| temp_filepath = os.path.join(temp_dir.name, txt_docs.name) | |
| with open(temp_filepath, "wb") as f: | |
| f.write(txt_docs.getvalue()) | |
| text_loader = TextLoader(temp_filepath) | |
| text_doc = text_loader.load() | |
| return text_doc | |
| # csv ํ์ผ๋ก๋ถํฐ text ์ถ์ถ | |
| def get_csv_file(csv_docs): | |
| temp_dir = tempfile.TemporaryDirectory() | |
| temp_filepath = os.path.join(temp_dir.name, csv_docs.name) | |
| with open(temp_filepath,"wb") as f: | |
| f.write(csv_docs.getvalue()) | |
| csv_loader = CSVLoader(temp_filepath) | |
| csv_doc = csv_loader.load() | |
| return csv_doc | |
| # def get_json_file(docs): | |
| # temp_dir = tempfile.TemporaryDirectory() | |
| # temp_filepath = os.path.join(temp_dir.name, docs.name) | |
| # with open(temp_filepath, "wb") as f: | |
| # f.write(docs.getvalue()) | |
| # json_loader = JSONLoader(temp_filepath, | |
| # jq_schema='.scans[].relationships', | |
| # text_content=False) | |
| # | |
| # json_doc = json_loader.load() | |
| # # print('json_doc = ',json_doc) | |
| # return json_doc | |
| def get_json_file(file) -> list[Document]: | |
| # Streamlit UploadedFile -> str | |
| raw = file.getvalue().decode("utf-8", errors="ignore") | |
| data = json.loads(raw) | |
| docs = [] | |
| # ์์ jq ๊ฒฝ๋ก๊ฐ '.scans[].relationships'์๋ค๋ฉด, ๋์ผํ ์๋ฏธ๋ก ํ์ฑ: | |
| # ์กด์ฌํ๋ฉด ๊ทธ๊ฒ๋ง ๋ฝ๊ณ , ์์ผ๋ฉด ํต์ผ๋ก ๋ฌธ์ํ | |
| def add_doc(x): | |
| docs.append(Document(page_content=json.dumps(x, ensure_ascii=False))) | |
| if isinstance(data, dict) and "scans" in data and isinstance(data["scans"], list): | |
| for s in data["scans"]: | |
| rels = s.get("relationships", []) | |
| if isinstance(rels, list) and rels: | |
| for r in rels: | |
| add_doc(r) | |
| if not docs: # ๊ทธ๋๋ ๋ชป ๋ฝ์์ผ๋ฉด ์ ์ฒด๋ฅผ ํ๋๋ก | |
| add_doc(data) | |
| elif isinstance(data, list): | |
| for item in data: | |
| add_doc(item) | |
| else: | |
| add_doc(data) | |
| return docs | |
| # ๋ฌธ์๋ค์ ์ฒ๋ฆฌํ์ฌ ํ ์คํธ ์ฒญํฌ๋ก ๋๋๋ ํจ์์ ๋๋ค. | |
| def get_text_chunks(documents): | |
| text_splitter = RecursiveCharacterTextSplitter( | |
| chunk_size=1000, # ์ฒญํฌ์ ํฌ๊ธฐ๋ฅผ ์ง์ ํฉ๋๋ค. | |
| chunk_overlap=200, # ์ฒญํฌ ์ฌ์ด์ ์ค๋ณต์ ์ง์ ํฉ๋๋ค. | |
| length_function=len # ํ ์คํธ์ ๊ธธ์ด๋ฅผ ์ธก์ ํ๋ ํจ์๋ฅผ ์ง์ ํฉ๋๋ค. | |
| ) | |
| documents = text_splitter.split_documents(documents) # ๋ฌธ์๋ค์ ์ฒญํฌ๋ก ๋๋๋๋ค. | |
| return documents # ๋๋ ์ฒญํฌ๋ฅผ ๋ฐํํฉ๋๋ค. | |
| # ํ ์คํธ ์ฒญํฌ๋ค๋ก๋ถํฐ ๋ฒกํฐ ์คํ ์ด๋ฅผ ์์ฑํ๋ ํจ์์ ๋๋ค. | |
| def get_vectorstore(text_chunks): | |
| # ์ํ๋ ์๋ฒ ๋ฉ ๋ชจ๋ธ์ ๋ก๋ํฉ๋๋ค. | |
| embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/clip-ViT-B-32-multilingual-v1', | |
| model_kwargs={'device': 'cpu'}) # ์๋ฒ ๋ฉ ๋ชจ๋ธ์ ์ค์ ํฉ๋๋ค. | |
| vectorstore = FAISS.from_documents(text_chunks, embeddings) # FAISS ๋ฒกํฐ ์คํ ์ด๋ฅผ ์์ฑํฉ๋๋ค. | |
| return vectorstore # ์์ฑ๋ ๋ฒกํฐ ์คํ ์ด๋ฅผ ๋ฐํํฉ๋๋ค. | |
| def get_conversation_chain(vectorstore): | |
| # Groq LLM | |
| llm = ChatGroq( | |
| groq_api_key=os.environ.get("GROQ_API_KEY"), | |
| model_name="llama-3.1-8b-instant", | |
| temperature=0.75, # ํ์์ ๋ง๊ฒ ํ๋ | |
| max_tokens=512 # ์ปจํ ์คํธ ์ด๊ณผ ๋ฐฉ์ง์ฉ (ํ์์ ์กฐ์ ) | |
| ) | |
| memory = ConversationBufferMemory( | |
| memory_key="chat_history", | |
| return_messages=True | |
| ) | |
| retriever = vectorstore.as_retriever(search_kwargs={"k": 3}) | |
| conversation_chain = ConversationalRetrievalChain.from_llm( | |
| llm=llm, | |
| retriever=retriever, | |
| memory=memory, | |
| ) | |
| return conversation_chain | |
| # ์ฌ์ฉ์ ์ ๋ ฅ์ ์ฒ๋ฆฌํ๋ ํจ์์ ๋๋ค. | |
| def handle_userinput(user_question): | |
| print('user_question => ', user_question) | |
| # ๋ํ ์ฒด์ธ์ ์ฌ์ฉํ์ฌ ์ฌ์ฉ์ ์ง๋ฌธ์ ๋ํ ์๋ต์ ์์ฑํฉ๋๋ค. | |
| response = st.session_state.conversation({'question': user_question}) | |
| # ๋ํ ๊ธฐ๋ก์ ์ ์ฅํฉ๋๋ค. | |
| st.session_state.chat_history = response['chat_history'] | |
| for i, message in enumerate(st.session_state.chat_history): | |
| if i % 2 == 0: | |
| st.write(user_template.replace( | |
| "{{MSG}}", message.content), unsafe_allow_html=True) | |
| else: | |
| st.write(bot_template.replace( | |
| "{{MSG}}", message.content), unsafe_allow_html=True) | |
| def main(): | |
| load_dotenv() | |
| st.set_page_config(page_title="Basic_RAG_AI_Chatbot_with_Llama", | |
| page_icon=":books:") | |
| st.write(css, unsafe_allow_html=True) | |
| if "conversation" not in st.session_state: | |
| st.session_state.conversation = None | |
| if "chat_history" not in st.session_state: | |
| st.session_state.chat_history = None | |
| st.header("Basic_RAG_AI_Chatbot_with_Llama3 :books:") | |
| user_question = st.text_input("Ask a question about your documents:") | |
| if user_question: | |
| handle_userinput(user_question) | |
| with st.sidebar: | |
| st.subheader("Your documents") | |
| docs = st.file_uploader( | |
| "Upload your Files here and click on 'Process'", accept_multiple_files=True) | |
| if st.button("Process[PDF]"): | |
| with st.spinner("Processing"): | |
| # get pdf text | |
| doc_list = [] | |
| for file in docs: | |
| print('file - type : ', file.type) | |
| if file.type in ['application/octet-stream', 'application/pdf']: | |
| # file is .pdf | |
| doc_list.extend(get_pdf_text(file)) | |
| else: | |
| st.error("PDF ํ์ผ์ด ์๋๋๋ค.") | |
| if not doc_list: | |
| st.error("์ฒ๋ฆฌ ๊ฐ๋ฅํ ๋ฌธ์๋ฅผ ์ฐพ์ง ๋ชปํ์ต๋๋ค.") | |
| st.stop() | |
| text_chunks = get_text_chunks(doc_list) | |
| vectorstore = get_vectorstore(text_chunks) | |
| st.session_state.conversation = get_conversation_chain(vectorstore) | |
| ################## TXT, CSV ๋ฒํผ ๊ตฌํ | |
| # TXT ๋ฒํผ ๊ตฌํ ์ฐธ๊ณ : if file.type == 'text/plain': | |
| # CSV ๋ฒํผ ๊ตฌํ ์ฐธ๊ณ : if file.type == 'text/csv': | |
| if st.button("Process[JSON]"): | |
| with st.spinner("Processing"): | |
| doc_list = [] | |
| for file in docs: | |
| print('file - type : ', file.type) | |
| if file.type == 'application/json': | |
| # file is .json | |
| doc_list.extend(get_json_file(file)) | |
| else: | |
| st.error("JSON ํ์ผ์ด ์๋๋๋ค.") | |
| if not doc_list: | |
| st.error("์ฒ๋ฆฌ ๊ฐ๋ฅํ ๋ฌธ์๋ฅผ ์ฐพ์ง ๋ชปํ์ต๋๋ค.") | |
| st.stop() | |
| text_chunks = get_text_chunks(doc_list) | |
| vectorstore = get_vectorstore(text_chunks) | |
| st.session_state.conversation = get_conversation_chain(vectorstore) | |
| if st.button("Process[TXT]"): | |
| with st.spinner("Processing"): | |
| # get txt text | |
| doc_list = [] | |
| for file in docs: | |
| print('file - type : ', file.type) | |
| if file.type == 'text/plain': | |
| doc_list.extend(get_text_file(file)) | |
| else: | |
| st.error("TXT ํ์ผ์ด ์๋๋๋ค.") | |
| if not doc_list: | |
| st.error("์ฒ๋ฆฌ ๊ฐ๋ฅํ ๋ฌธ์๋ฅผ ์ฐพ์ง ๋ชปํ์ต๋๋ค.") | |
| st.stop() | |
| text_chunks = get_text_chunks(doc_list) | |
| vectorstore = get_vectorstore(text_chunks) | |
| st.session_state.conversation = get_conversation_chain(vectorstore) | |
| if st.button("Process[CSV]"): | |
| with st.spinner("Processing"): | |
| # get csv text | |
| doc_list = [] | |
| for file in docs: | |
| print('file - type : ', file.type) | |
| if file.type == 'text/csv': | |
| doc_list.extend(get_csv_file(file)) | |
| else: | |
| st.error("csv ํ์ผ์ด ์๋๋๋ค.") | |
| if not doc_list: | |
| st.error("์ฒ๋ฆฌ ๊ฐ๋ฅํ ๋ฌธ์๋ฅผ ์ฐพ์ง ๋ชปํ์ต๋๋ค.") | |
| st.stop() | |
| text_chunks = get_text_chunks(doc_list) | |
| vectorstore = get_vectorstore(text_chunks) | |
| st.session_state.conversation = get_conversation_chain(vectorstore) | |
| if __name__ == '__main__': | |
| main() |