app.py

import streamlit as st
import json
from datasets import load_dataset

st.set_page_config(page_title="Kaggle Notebooks inspection", layout="wide")

st.markdown("<h1 style='text-align: center; color: #00BFFF;'>Kaggle Notebooks inspection 🔍</h1>", unsafe_allow_html=True)

st.markdown("Here you can inspect Kaggle notebooks that were converted to python scripts and deduplicated.")
@st.cache()
def load_data():
 ds = load_dataset("loubnabnl/subset_kaggle_scripts", split="train")
 return ds

def show_extra_info(e):
 kv = json.loads(e["kversion"])[0]
 try:
 data_v = json.loads(e["dataset_versions"])[0]
 except:
 data_v = ""
 if data_v:
 data_title = data_v["Title"]
 import numpy as np
 decription = data_v["Description"] if str(data_v["Description"]) != 'nan' else "<empty_description>"
 data_text = f"<br><h4 style=color: #00BFFF;'📚 Dataset description:</h4><br>Title: **{data_title}**, described as: {decription}."
 else:
 data_text = ""
 
 text = f"The title of the notebook is: **{kv['Title']}** and it has **{kv['TotalVotes']} ⬆️ upvotes**.{data_text}"
 return text

samples = load_data()
index_example = st.number_input(f"Chose a sample from the existing {len(samples)} notebooks:", min_value=0, max_value=len(samples)-1, value=0, step=1)
st.markdown(show_extra_info(samples[index_example]), unsafe_allow_html=True)
st.markdown("<h4 style=color: #00BFFF;'💻 Kaggle Notebook</h4>", unsafe_allow_html=True)
st.code(samples[index_example]["script"])