import Foundation

/// A named "row + fields" extraction recipe, auto-applied when a page's URL
/// matches. The Discogs profile mirrors the selectors in the scraperrr notes.
struct ScrapeProfile {
    let name: String
    let urlPattern: String
    let rowSelector: String
    let fields: [(name: String, selector: String)]

    func matchesURL(_ url: String) -> Bool {
        url.range(of: urlPattern, options: [.regularExpression, .caseInsensitive]) != nil
    }

    static let discogsMarketplace = ScrapeProfile(
        name: "discogs_marketplace",
        urlPattern: #"discogs\.com/(sell/list|sell/mywants|seller/[^/]+/profile)"#,
        rowSelector: "tr.shortcut_navigable",
        fields: [
            ("item_id", "a.cart-button@data-item-id"),
            ("release_id", "tr@data-release-id"),
            ("title", "a.item_description_title"),
            ("listing_url", "a.item_description_title@href"),
            ("media_condition", ".item_condition span[!class]"),
            ("sleeve_condition", "span.item_sleeve_condition"),
            ("seller_username", ".seller_block a"),
            ("seller_id", "li.seller_mywants@data-seller-id"),
            // Co-located with seller_id on the same element — a reliable
            // username fallback so a seller is never referenced unmapped.
            ("seller_username_alt", "li.seller_mywants@data-username"),
            ("price_value", "span.price@data-pricevalue"),
            ("price_currency", "span.price@data-currency"),
            ("shipping", "span.item_shipping"),
        ]
    )

    static let builtIn: [ScrapeProfile] = [discogsMarketplace]

    static func matching(url: String?) -> ScrapeProfile? {
        guard let url else { return nil }
        return builtIn.first { $0.matchesURL(url) }
    }
}

struct DetectedPage {
    let url: String?
    let html: String
}

enum PageDetector {
    /// Recognizes the "URL on the first line, then pasted HTML" convention
    /// (URL optional). Returns nil when the document doesn't look like HTML.
    static func detect(in text: String) -> DetectedPage? {
        var url: String?
        var html = String(text.drop(while: { $0.isWhitespace }))
        if html.hasPrefix("http://") || html.hasPrefix("https://") {
            if let newline = html.firstIndex(of: "\n") {
                let candidate = String(html[..<newline]).trimmingCharacters(in: .whitespaces)
                if !candidate.contains(" "), !candidate.contains("<") {
                    url = candidate
                    html = String(html[newline...])
                }
            }
        }
        let htmlish = html.contains("<body") || html.contains("<html") || html.contains("<div")
            || html.contains("<table") || html.drop(while: { $0.isWhitespace }).hasPrefix("<")
        guard htmlish else { return nil }
        return DetectedPage(url: url, html: html)
    }
}