//
//  ContentExtractor.swift
//  Zyquo Atlas
//
//  Author: Simon-Pierre Boucher
//  Mail: contact@spboucher.ai
//
//  Injects the extraction scripts (Mozilla Readability + AtlasExtractor driver)
//  into an isolated WKContentWorld and pulls a normalized PageContext out of a
//  live tab via callAsyncJavaScript. Isolation keeps the extractor from
//  colliding with or being observed by page JS (docs/AI-BROWSER-RESEARCH.md §2).
//  All page content is treated as untrusted downstream (Phase 3.C prompt
//  hygiene).
//

import Foundation
import WebKit

enum ContentExtractionError: LocalizedError {
    case scriptsUnavailable
    case scriptFailure(String)
    case decodeFailure(String)

    var errorDescription: String? {
        switch self {
        case .scriptsUnavailable:
            return "Content extraction scripts are missing from the app bundle."
        case .scriptFailure(let detail):
            return "Couldn't read this page's content (\(detail))."
        case .decodeFailure(let detail):
            return "Couldn't parse the extracted page content (\(detail))."
        }
    }
}

@MainActor
final class ContentExtractor {
    /// Dedicated world so extraction never touches the page's JS globals.
    static let world = WKContentWorld.world(name: "ZyquoAtlasContent")

    /// Installs the extraction user scripts into a configuration. Call once per
    /// WKWebViewConfiguration (ProfileStore) before any tab loads.
    static func install(into config: WKWebViewConfiguration) {
        let ucc = config.userContentController
        for name in ["Readability", "Readability-readerable", "AtlasExtractor"] {
            guard let source = script(name) else { continue }
            ucc.addUserScript(WKUserScript(
                source: source,
                injectionTime: .atDocumentEnd,
                forMainFrameOnly: true,
                in: world
            ))
        }
    }

    /// Extracts a PageContext from a tab's web view. `mode` "auto" tries
    /// Readability first; "rawText" forces the visible-text fallback.
    static func extract(from webView: WKWebView, mode: String = "auto") async throws -> PageContext {
        let result: Any?
        do {
            result = try await webView.callAsyncJavaScript(
                "return window.__zyquoAtlas ? window.__zyquoAtlas.extract(mode, maxBytes) : { ok:false, error:'driver-missing' };",
                arguments: ["mode": mode, "maxBytes": 600_000],
                in: nil,
                contentWorld: world
            )
        } catch {
            throw ContentExtractionError.scriptFailure(error.localizedDescription)
        }

        guard let dict = result as? [String: Any] else {
            throw ContentExtractionError.scriptFailure("no result")
        }
        if dict["ok"] as? Bool != true {
            throw ContentExtractionError.scriptFailure((dict["error"] as? String) ?? "unknown")
        }
        guard let context = dict["context"] else {
            throw ContentExtractionError.scriptFailure("empty context")
        }
        do {
            let data = try JSONSerialization.data(withJSONObject: context)
            return try JSONDecoder().decode(PageContext.self, from: data)
        } catch {
            throw ContentExtractionError.decodeFailure(error.localizedDescription)
        }
    }

    // MARK: - Script loading

    private static func script(_ name: String) -> String? {
        guard let url = Bundle.module.url(forResource: name, withExtension: "js"),
              let source = try? String(contentsOf: url, encoding: .utf8) else {
            return nil
        }
        return source
    }
}