import { exec } from 'child_process'
import { promisify } from 'util'
import fs from 'fs'

const execAsync = promisify(exec)

const PYTHON_BRIDGE = `
import scrapling
from scrapling.fetchers import StealthyFetcher
from bs4 import BeautifulSoup
import json
import re

def clean_code_block(text):
    raw_lines = text.split('\\n')
    cleaned_lines = []
    current_line = []

    for l in raw_lines:
        s = l.strip()
        if not s:
            if current_line:
                cleaned_lines.append(" ".join(current_line))
                current_line = []
            continue

        if s.startswith('#') or s.startswith('$') or s.startswith('>>>'):
            if current_line:
                cleaned_lines.append(" ".join(current_line))
                current_line = []
            cleaned_lines.append(s)
        else:
            current_line.append(s)

    if current_line:
        cleaned_lines.append(" ".join(current_line))

    res = "\\n".join(cleaned_lines)
    res = re.sub(r'\\s+([,.:;\\]\\)\\}])', r'\\1', res)
    res = re.sub(r'([\\[\\(\\{])\\s+', r'\\1', res)
    res = re.sub(r'\\s*=\\s*', ' = ', res)
    res = re.sub(r'\\s*\\+\\=\\s*', ' += ', res)
    res = re.sub(r'\\s*\\-\\=\\s*', ' -= ', res)
    return res.strip()

def fetch_cheatsheet():
    try:
        fetcher = StealthyFetcher()
        page = fetcher.fetch('https://realpython.com/cheatsheets/python/', network_idle=True)
        if page.status != 200:
            print(json.dumps({"status": False, "error": f"HTTP {page.status}"}))
            return

        soup = BeautifulSoup(page.html_content, 'html.parser')
        title = "Python 3 Cheat Sheet"

        sections = []

        # Blacklist of promotional keywords
        PROMO_WORDS = [
            'real python', 'show/hide', 'continue your learning', 'bonus',
            'download this', 'newsletter', 'quizzes', 'podcast', 'get a python cheat',
            'learning paths', 'video courses', 'written tutorials', 'interactive quizzes',
            'do you want to go deeper'
        ]

        for sec_tag in soup.find_all('section', class_='section2'):
            h2 = sec_tag.find('h2')
            if not h2:
                continue
            sec_title = h2.get_text().strip()

            # Skip promotional sections
            if any(pw in sec_title.lower() for pw in ['related tutorial', 'bonus', 'resources', 'curriculum']):
                continue

            items = []
            for child in sec_tag.find_all(['pre', 'p']):
                if child.name == 'pre':
                    code_clean = clean_code_block(child.get_text())
                    if code_clean:
                        items.append({'type': 'code', 'language': 'python', 'content': code_clean})
                elif child.name == 'p':
                    t = child.get_text().strip()
                    if t and len(t) > 3 and not any(pw in t.lower() for pw in PROMO_WORDS):
                        items.append({'type': 'text', 'content': t})

            if any(it['type'] == 'code' for it in items):
                sections.append({
                    'title': sec_title,
                    'items': items
                })

        print(json.dumps({
            "status": True,
            "title": title,
            "total_sections": len(sections),
            "sections": sections
        }))
    except Exception as e:
        print(json.dumps({"status": False, "error": str(e)}))

if __name__ == "__main__":
    fetch_cheatsheet()
`

let cacheCheatsheet = null
let lastFetch = 0

export async function getPythonCheatSheet(forceRefresh = false) {
    const now = Date.now()
    if (!forceRefresh && cacheCheatsheet && (now - lastFetch < 3600000)) {
        return cacheCheatsheet
    }

    try {
        const tmpScript = '/tmp/rp_cheatsheet_clean.py'
        await fs.promises.writeFile(tmpScript, PYTHON_BRIDGE, 'utf-8')

        const { stdout } = await execAsync(`${process.env.SCRAPLING_PYTHON || "/tmp/scrapping-venv/bin/python"} ${tmpScript}`, { maxBuffer: 15 * 1024 * 1024 })
        const json = JSON.parse(stdout.trim().split('\n').pop())

        if (!json.status) throw new Error(json.error || 'Gagal mengekstrak cheatsheet.')

        cacheCheatsheet = json
        lastFetch = now
        return json
    } catch (e) {
        console.error('[RealPython CheatSheet Error]:', e.message)
        throw e
    }
}

export default {
    getPythonCheatSheet
}