|
import streamlit as st |
|
from langchain.llms import HuggingFaceHub |
|
import os |
|
|
|
from PyPDF2 import PdfReader |
|
from langchain.text_splitter import RecursiveCharacterTextSplitter |
|
from langchain.embeddings import HuggingFaceEmbeddings |
|
from langchain.vectorstores import FAISS |
|
from langchain.chat_models import ChatOpenAI |
|
from langchain.chains.question_answering import load_qa_chain |
|
|
|
st.set_page_config('preguntaDOC') |
|
st.header("Pregunta a tu PDF") |
|
OPENAI_API_KEY = st.text_input('OpenAI API Key', type='password') |
|
pdf_obj = st.file_uploader("Carga tu documento", type="pdf", on_change=st.cache_resource.clear) |
|
|
|
@st.cache_resource |
|
def create_embeddings(pdf): |
|
pdf_reader = PdfReader(pdf) |
|
text = "" |
|
for page in pdf_reader.pages: |
|
text += page.extract_text() |
|
|
|
text_splitter = RecursiveCharacterTextSplitter( |
|
chunk_size=800, |
|
chunk_overlap=100, |
|
length_function=len |
|
) |
|
chunks = text_splitter.split_text(text) |
|
|
|
|
|
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/paraphrase-multilingual-mpnet-base-v2") |
|
knowledge_base = FAISS.from_texts(chunks, embeddings) |
|
|
|
return knowledge_base |
|
|
|
if pdf_obj: |
|
knowledge_base = create_embeddings(pdf_obj) |
|
user_question = st.text_input("Haz una pregunta sobre tu PDF:") |
|
|
|
if user_question: |
|
os.environ["HUGGINGFACEHUB_API_TOKEN"] = "" |
|
docs = knowledge_base.similarity_search(user_question, 3) |
|
|
|
|
|
llm = HuggingFaceHub(repo_id="google/flan-t5-xxl", model_kwargs={"temperature":0.5, "max_length":512}) |
|
chain = load_qa_chain(llm, chain_type="stuff") |
|
respuesta = chain.run(input_documents=docs, question=user_question) |
|
|
|
st.write(respuesta) |
