const express = require('express');
const axios = require('axios');
const xml2js = require('xml2js');
const cheerio = require('cheerio');

const app = express();
const port = process.env.PORT || 8080;

// Middleware to parse JSON requests
app.use(express.json());

function textSplitter(text, chunkSize, overlap) {
  if (text.length <= chunkSize) {
    return [text];
  }

  let end = chunkSize;
  
  // Find the position to split the text (avoiding word boundary)
  while (end > 0 && text.charAt(end) !== ' ') {
    end--;
  }

  // Find the position for overlap to split the text (avoiding word boundary)
  let from = end - overlap;
  while (from > 0 && text.charAt(from) !== ' ') {
    from--;
  }
  
  const chunk = text.substring(0, end).trim();
  const remainingText = text.substring(from).trim(); // Adjust for overlap

  return [chunk, ...textSplitter(remainingText, chunkSize, overlap)];
}

// REST API endpoint
app.post('/extract', async (req, res) => {
  const { source, selector, chunkSize, chunkOverlap } = req.body;
  let selectedContent = '';
  let chunks = [];
  try {
    const response = await axios.get(source);
    const htmlContent = response.data;
    try {
      const $ = cheerio.load(htmlContent);

      selectedContent = $(selector).text().trim().replace(/<(?:"[^"]*"['"]*|'[^']*'['"]*|[^'">])+>/g, '');
      chunks = textSplitter(selectedContent, chunkSize, chunkOverlap)
    } catch(err) {
      console.error(err);
    }
    res.json({content: selectedContent, chunks})
  } catch (error) {
    console.error(error);
    res.status(500).json({ error: `Error fetching content from ${source}: ${error.message}` });
  }
});

// REST API endpoint
app.post('/split', async (req, res) => {
  const { content, chunkSize, chunkOverlap } = req.body;
  try {
    const selectedContent = content.trim().replace(/<(?:"[^"]*"['"]*|'[^']*'['"]*|[^'">])+>/g, '');
    const chunks = textSplitter(selectedContent, chunkSize, chunkOverlap)
    res.json({content: selectedContent, chunks})
  } catch (error) {
    console.error(error);
    res.status(500).json({ error: `Error splitting content: ${error.message}` });
  }
});

// Start the server
app.listen(port, async () => {
  console.log(`Server is running on port ${port}`);
});