app.py

import json
import logging
from logging import Logger
from logging.handlers import SysLogHandler

import streamlit as st
import tokenizers
import torch
from transformers import Pipeline, pipeline

from utils import get_answer, get_context


@st.cache(
    hash_funcs={
        torch.nn.parameter.Parameter: lambda _: None,
        tokenizers.Tokenizer: lambda _: None,
        tokenizers.AddedToken: lambda _: None,
    },
    allow_output_mutation=True,
    show_spinner=False,
)
def load_engine() -> Pipeline:

    nlp_qa = pipeline(
        "question-answering",
        model="mrm8488/bert-italian-finedtuned-squadv1-it-alfa",
        tokenizer="mrm8488/bert-italian-finedtuned-squadv1-it-alfa",
    )

    return nlp_qa


if ("syslog" not in st.session_state) and ("logger" not in st.session_state):
    syslog = SysLogHandler(
        address=(st.secrets["logging_address"], int(st.secrets["logging_port"]))
    )
    logger = logging.getLogger()
    logger.setLevel(logging.INFO)
    logger.addHandler(syslog)
    st.session_state["syslog"] = syslog
    st.session_state["logger"] = logger

with st.spinner(
    text="Sto preparando il necessario per rispondere alle tue domande personali..."
):

    engine = load_engine()

st.title("Le risposte alle tue domande personali")

input = st.text_input("Scrivi una domanda in italiano e comparirà la risposta!")

if input:
    try:
        context = get_context()
        st.session_state["logger"].info(input)
        answer = get_answer(input, context, engine)
        st.subheader(answer)
    except:

        st.error(
            "Qualcosa é andato storto. Prova di nuovo con un'altra domanda magari!"
        )