app.py

import streamlit as st
import datasets
from datasets import load_dataset
import pandas as pd
from streamlit.components.v1 import html
from streamlit import markdown
import re
import os
import time

import json

st.title('StackOverflow Question Demo')

library = st.radio('Select a library', ('numpy', 'tensorflow', 'pytorch', 'scipy', 'scikit-learn', 'pandas'))

question_path = './{}.txt'.format(library)
# question_path = 'numpy.txt'
# loading stackoverflow questions.
# using huggingface load_dataset function.
# not done yet
# @st.cache
# def load_data(path):
#    return load_dataset('text', data_files = path, cache_dir = './data')


dataset = []
# dataset = load_data(question_path)
with open(question_path) as f:
    lines = f.readlines()
    question = ''
    temp = {}
    tag = ''
    for line in lines:
        if line == 'Origin:\n' or line == 'Function:\n' or re.match(r'A\d:\n', line):
            if not tag:
                tag = line[:-2]
            else:
                temp[tag] = question
                question = ''
                tag = line[:-2]
        elif re.match(r'\d*\.\n', line):
            if tag:
                temp[tag] = question
                dataset.append(temp)
                question = ''
                tag = ''
                temp = {}
        else:
            if tag:
                question += line + '\n'

temp[tag] = question
dataset.append(temp)

# Select index
number = st.number_input("Insert a index: range from",
                         min_value=0, max_value=len(dataset) - 1)
st.write('The current index is ', number)
data_index = int(number)

# Selece modification
options = tuple(dataset[data_index].keys())
modification = st.radio('Modification:',
                        options=options
                        )

st.write(dataset[data_index][modification])