use reqwest::blocking::Client;
use reqwest::header::{HeaderMap, HeaderValue, USER_AGENT};
use std::collections::HashMap;
use std::env;
use std::fs;
use std::io::{self, prelude::*, Error, Write};
use std::path::Path;
use std::process;
use std::time::Duration;

use monolith::html::{
    add_favicon, create_metadata_tag, get_base_url, has_favicon, html_to_dom, set_base_url,
    stringify_document, walk_and_embed_assets,
};
use monolith::opts::Options;
use monolith::url::{
    data_to_data_url, is_data_url, is_file_url, is_http_url, parse_data_url, resolve_url,
};
use monolith::utils::retrieve_asset;

mod macros;

enum Output {
    Stdout(io::Stdout),
    File(fs::File),
}

impl Output {
    fn new(file_path: &str) -> Result<Output, Error> {
        if file_path.is_empty() || file_path.eq("-") {
            Ok(Output::Stdout(io::stdout()))
        } else {
            Ok(Output::File(fs::File::create(file_path)?))
        }
    }

    fn writeln_str(&mut self, s: &str) -> Result<(), Error> {
        match self {
            Output::Stdout(stdout) => {
                writeln!(stdout, "{}", s)?;
                stdout.flush()
            }
            Output::File(f) => {
                writeln!(f, "{}", s)?;
                f.flush()
            }
        }
    }
}

pub fn read_stdin() -> String {
    let mut buffer = String::new();
    for line in io::stdin().lock().lines() {
        buffer += line.unwrap_or_default().as_str();
        buffer += "\n";
    }
    buffer
}

fn main() {
    let options = Options::from_args();
    let original_target: &str = &options.target;
    let target_url: &str;
    let mut base_url: String;
    let mut dom;
    let mut use_stdin: bool = false;

    // Pre-process the input
    let cwd_normalized: String =
        str!(env::current_dir().unwrap().to_str().unwrap()).replace("\\", "/");
    let path = Path::new(original_target);
    let mut target: String = str!(original_target.clone()).replace("\\", "/");
    let path_is_relative: bool = path.is_relative();

    // Determine exact target URL
    if target.clone().len() == 0 {
        if !options.silent {
            eprintln!("No target specified");
        }
        process::exit(1);
    } else if target.clone() == "-" {
        // Read from pipe (stdin)
        use_stdin = true;
        // Default target URL to empty data URL; the user can control it via --base-url
        target_url = "data:text/html,"
    } else if is_http_url(target.clone()) || is_data_url(target.clone()) {
        target_url = target.as_str();
    } else if is_file_url(target.clone()) {
        target_url = target.as_str();
    } else if path.exists() {
        if !path.is_file() {
            if !options.silent {
                eprintln!("Local target is not a file: {}", original_target);
            }
            process::exit(1);
        }
        target.insert_str(0, if cfg!(windows) { "file:///" } else { "file://" });
        if path_is_relative {
            target.insert_str(if cfg!(windows) { 8 } else { 7 }, &cwd_normalized);
            target.insert_str(
                if cfg!(windows) { 8 } else { 7 } + &cwd_normalized.len(),
                "/",
            );
        }
        target_url = target.as_str();
    } else {
        target.insert_str(0, "http://");
        target_url = target.as_str();
    }

    // Define output
    let mut output = Output::new(&options.output).expect("Could not prepare output");

    // Initialize client
    let mut cache = HashMap::new();
    let mut header_map = HeaderMap::new();
    if let Some(user_agent) = &options.user_agent {
        header_map.insert(
            USER_AGENT,
            HeaderValue::from_str(&user_agent).expect("Invalid User-Agent header specified"),
        );
    }
    let timeout: u64 = if options.timeout > 0 {
        options.timeout
    } else {
        std::u64::MAX / 4
    };
    let client = Client::builder()
        .timeout(Duration::from_secs(timeout))
        .danger_accept_invalid_certs(options.insecure)
        .default_headers(header_map)
        .build()
        .expect("Failed to initialize HTTP client");

    // At this stage we assume that the base URL is the same as the target URL
    base_url = str!(target_url);

    // Retrieve target document
    if use_stdin {
        dom = html_to_dom(&read_stdin());
    } else if is_file_url(target_url) || is_http_url(target_url) {
        match retrieve_asset(&mut cache, &client, target_url, target_url, &options, 0) {
            Ok((data, final_url, _media_type)) => {
                if options.base_url.clone().unwrap_or(str!()).is_empty() {
                    base_url = final_url
                }
                dom = html_to_dom(&String::from_utf8_lossy(&data));
            }
            Err(_) => {
                if !options.silent {
                    eprintln!("Could not retrieve target document");
                }
                process::exit(1);
            }
        }
    } else if is_data_url(target_url) {
        let (media_type, data): (String, Vec<u8>) = parse_data_url(target_url);
        if !media_type.eq_ignore_ascii_case("text/html") {
            if !options.silent {
                eprintln!("Unsupported data URL media type");
            }
            process::exit(1);
        }
        dom = html_to_dom(&String::from_utf8_lossy(&data));
    } else {
        process::exit(1);
    }

    // Use custom base URL if specified, read and use what's in the DOM otherwise
    if !options.base_url.clone().unwrap_or(str!()).is_empty() {
        if is_data_url(options.base_url.clone().unwrap()) {
            if !options.silent {
                eprintln!("Data URLs cannot be used as base URL");
            }
            process::exit(1);
        } else {
            base_url = options.base_url.clone().unwrap();
        }
    } else {
        if let Some(existing_base_url) = get_base_url(&dom.document) {
            base_url = resolve_url(target_url, existing_base_url).unwrap();
        }
    }

    // Embed remote assets
    walk_and_embed_assets(&mut cache, &client, &base_url, &dom.document, &options, 0);

    // Update or add new BASE tag to reroute network requests and hash-links in the final document
    if let Some(new_base_url) = options.base_url.clone() {
        dom = set_base_url(&dom.document, new_base_url);
    }

    // Request and embed /favicon.ico (unless it's already linked in the document)
    if !options.no_images && is_http_url(target_url) && !has_favicon(&dom.document) {
        let favicon_ico_url: String = resolve_url(&base_url, "/favicon.ico").unwrap();

        match retrieve_asset(
            &mut cache,
            &client,
            &base_url,
            &favicon_ico_url,
            &options,
            0,
        ) {
            Ok((data, final_url, media_type)) => {
                let favicon_data_url: String = data_to_data_url(&media_type, &data, &final_url);
                dom = add_favicon(&dom.document, favicon_data_url);
            }
            Err(_) => {
                // Failed to retrieve favicon.ico
            }
        }
    }

    // Remove charset meta-tag
    // set_charset_meta_to_utf8(&dom.document);

    // Serialize DOM tree
    let mut result: String = stringify_document(&dom.document, &options);

    // Add metadata tag
    if !options.no_metadata {
        let metadata_comment: String = create_metadata_tag(&target_url);
        result.insert_str(0, &metadata_comment);
        if metadata_comment.len() > 0 {
            result.insert_str(metadata_comment.len(), "\n");
        }
    }

    // Write result into stdout or file
    output
        .writeln_str(&result)
        .expect("Could not write HTML output");
}