// src/scraper/webdriver.rs

use anyhow::{anyhow, Context, Result};
use fantoccini::{Client, ClientBuilder};
use serde_json::{Map, Value};
use std::pin::Pin;
use std::process::Stdio;
use std::sync::Arc;
use tokio::io::{AsyncBufReadExt, BufReader};
use tokio::process::{Child, Command};
use tokio::sync::{Mutex, Semaphore};
use tokio::time::{sleep, timeout, Duration};

/// Manages a pool of ChromeDriver instances for parallel scraping.
///
/// This struct maintains multiple ChromeDriver processes and allows controlled
/// concurrent access via a semaphore. Instances are reused across tasks to avoid
/// the overhead of spawning new processes.
pub struct ChromeDriverPool {
    instances: Vec<Arc<Mutex<ChromeInstance>>>,
    semaphore: Arc<Semaphore>,
    tasks_per_instance: usize,
}

impl ChromeDriverPool {
    /// Creates a new pool with the specified number of ChromeDriver instances.
    ///
    /// # Arguments
    /// * `pool_size` - Number of concurrent ChromeDriver instances to maintain
    pub async fn new(pool_size: usize) -> Result<Self> {
        let mut instances = Vec::with_capacity(pool_size);

        println!(
            "Initializing ChromeDriver pool with {} instances...",
            pool_size
        );

        for i in 0..pool_size {
            match ChromeInstance::new().await {
                Ok(instance) => {
                    println!("  ✓ Instance {} ready", i + 1);
                    instances.push(Arc::new(Mutex::new(instance)));
                }
                Err(e) => {
                    eprintln!("  ✗ Failed to create instance {}: {}", i + 1, e);
                    // Clean up already created instances
                    drop(instances);
                    return Err(e);
                }
            }
        }

        Ok(Self {
            instances,
            semaphore: Arc::new(Semaphore::new(pool_size)),
            tasks_per_instance: 0,
        })
    }

    /// Executes a scrape task using an available instance from the pool.
    pub async fn execute<T, F, Fut>(&self, url: String, parse: F) -> Result<T>
    where
        T: Send + 'static,
        F: FnOnce(Client) -> Fut + Send + 'static,
        Fut: std::future::Future<Output = Result<T>> + Send + 'static,
    {
        // Acquire semaphore permit
        let _permit = self
            .semaphore
            .acquire()
            .await
            .map_err(|_| anyhow!("Semaphore closed"))?;

        // Find an available instance (round-robin or first available)
        let instance = self.instances[0].clone(); // Simple: use first, could be round-robin
        let mut guard = instance.lock().await;

        // Create a new session for this task
        let client = guard.new_session().await?;

        // Release lock while we do the actual scraping
        drop(guard);

        // Navigate and parse
        client.goto(&url).await.context("Failed to navigate")?;
        let result = timeout(Duration::from_secs(60), parse(client))
            .await
            .context("Parse function timed out after 60s")??;

        Ok(result)
    }

    pub fn get_number_of_instances(&self) -> usize {
        self.instances.len()
    }
}

/// Represents a single instance of chromedriver process.
pub struct ChromeInstance {
    process: Child,
    base_url: String,
}

impl ChromeInstance {
    /// Creates a new ChromeInstance by spawning chromedriver with random port.
    ///
    /// This spawns `chromedriver --port=0` to avoid port conflicts, reads stdout to extract
    /// the listening address, and waits for the success message. If timeout occurs or
    /// spawning fails, returns an error with context.
    ///
    /// # Errors
    ///
    /// Returns an error if chromedriver fails to spawn (e.g., not in PATH, version mismatch),
    /// if the process exits early, or if the address/success message isn't found within 30s.
    pub async fn new() -> Result<Self> {
        let mut command = Command::new("chromedriver-win64/chromedriver.exe");
        command
            .arg("--port=0") // Use random available port to support pooling
            .stdout(Stdio::piped())
            .stderr(Stdio::piped());

        let mut process = command
            .spawn()
            .context("Failed to spawn chromedriver. Ensure it's installed and in PATH.")?;

        let mut stdout =
            BufReader::new(process.stdout.take().context("Failed to capture stdout")?).lines();

        let mut stderr =
            BufReader::new(process.stderr.take().context("Failed to capture stderr")?).lines();

        let start_time = std::time::Instant::now();
        let mut address: Option<String> = None;
        let mut success = false;

        // Log stderr in background for debugging
        tokio::spawn(async move {
            while let Ok(Some(line)) = stderr.next_line().await {
                eprintln!("ChromeDriver stderr: {}", line);
            }
        });

        // Wait for address and success (up to 30s)
        while start_time.elapsed() < Duration::from_secs(30) {
            if let Ok(Ok(Some(line))) = timeout(Duration::from_secs(1), stdout.next_line()).await {
                if let Some(addr) = parse_chromedriver_address(&line) {
                    address = Some(addr.to_string());
                }

                if line.contains("ChromeDriver was started successfully") {
                    success = true;
                }

                if let (Some(addr), true) = (&address, success) {
                    return Ok(Self {
                        process,
                        base_url: addr.clone(),
                    });
                }
            }

            sleep(Duration::from_millis(100)).await;
        }

        // Cleanup on failure
        let _ = process.kill().await;
        Err(anyhow!("Timeout: ChromeDriver did not start within 30 seconds. Check version match with Chrome browser and system resources."))
    }

    /// Creates a new browser session (client) from this ChromeDriver instance.
    /// Each session is independent and can be closed without affecting the driver.
    pub async fn new_session(&self) -> Result<Client> {
        ClientBuilder::native()
            .capabilities(Self::chrome_args())
            .connect(&self.base_url)
            .await
            .context("Failed to create new session")
    }

    fn chrome_args() -> Map<String, Value> {
        let args = serde_json::json!({
            "goog:chromeOptions": {
                "args": [
                    "--headless=new",
                    "--disable-gpu",
                    "--no-sandbox",
                    "--disable-dev-shm-usage",
                    "--disable-infobars",
                    "--disable-extensions",
                    "--disable-popup-blocking",
                    "--disable-notifications",
                    "--disable-logging",
                    "--disable-autofill",
                    "--disable-features=TranslateUI,OptimizationGuideModelDownloading",
                    "--window-size=1920,1080",
                    "--disable-blink-features=AutomationControlled",
                    "--user-agent=Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
                ],
                "excludeSwitches": ["enable-logging", "enable-automation"],
                "useAutomationExtension": false,
                "prefs": {
                    "profile.default_content_setting_values.notifications": 2
                }
            }
        });
        args.as_object()
            .expect("Capabilities should be a JSON object")
            .clone()
    }
}

/// Parses the ChromeDriver address from a log line.
///
/// Looks for the "Starting ChromeDriver ... on port XXXX" line and extracts the port.
/// Returns `Some("http://localhost:XXXX")` if found, else `None`.
fn parse_chromedriver_address(line: &str) -> Option<String> {
    if line.contains("Starting ChromeDriver") {
        if let Some(port_str) = line.split("on port ").nth(1) {
            if let Some(port) = port_str.split_whitespace().next() {
                if port.parse::<u16>().is_ok() {
                    return Some(format!("http://localhost:{}", port));
                }
            }
        }
    }
    // Fallback for other formats (e.g., explicit port mentions)
    for word in line.split_whitespace() {
        if let Ok(port) = word.trim_matches(|c: char| !c.is_numeric()).parse::<u16>() {
            if port > 1024 && port < 65535 && line.to_lowercase().contains("port") {
                return Some(format!("http://localhost:{}", port));
            }
        }
    }
    None
}

impl Drop for ChromeInstance {
    fn drop(&mut self) {
        let _ = self.process.start_kill();
        std::thread::sleep(std::time::Duration::from_millis(100));
    }
}

/// Simplified task execution - now uses the pool pattern.
///
/// For backwards compatibility with existing code.
pub struct ScrapeTask<T> {
    url: String,
    parse: Box<
        dyn FnOnce(Client) -> Pin<Box<dyn std::future::Future<Output = Result<T>> + Send>> + Send,
    >,
}

impl<T: Send + 'static> ScrapeTask<T> {
    pub fn new<F, Fut>(url: String, parse: F) -> Self
    where
        F: FnOnce(Client) -> Fut + Send + 'static,
        Fut: std::future::Future<Output = Result<T>> + Send + 'static,
    {
        Self {
            url,
            parse: Box::new(move |client| Box::pin(parse(client))),
        }
    }

    /// Executes using a provided pool (more efficient for multiple tasks).
    pub async fn execute_with_pool(self, pool: &ChromeDriverPool) -> Result<T> {
        let url = self.url;
        let parse = self.parse;

        pool.execute(url, move |client| async move { (parse)(client).await })
            .await
    }
}