//
//  ContentBlock.swift
//  Prisme
//
//  Author: Simon-Pierre Boucher <contact@spboucher.ai>
//

import Foundation

/// A typed block produced by the deterministic DOM extraction (Distiller
/// step 2). `domPath` is the child-index path from `document.body` to the
/// source element — the anchor that lets every generated summary point back
/// to the real page (CLAUDE.md §4, the easiest constraint to forget).
struct ContentBlock: Codable, Hashable, Sendable {
    enum Kind: String, Codable, Sendable {
        case heading
        case paragraph
        case listItem
        case quote
        case code
        case table
        case caption
    }

    let kind: Kind
    /// Heading depth (1–6) for headings, 0 otherwise.
    let level: Int
    let text: String
    let domPath: String
}

/// Everything the extractor returns for one page.
struct ExtractedContent: Codable, Sendable {
    let title: String
    let lang: String?
    let blocks: [ContentBlock]

    /// First sentence of the first paragraph — the deterministic essence
    /// used wherever the model has nothing better to offer.
    var leadSentence: String? {
        guard let paragraph = blocks.first(where: { $0.kind == .paragraph })?.text else {
            return nil
        }
        if let end = paragraph.firstIndex(where: { ".!?".contains($0) }) {
            return String(paragraph[...end])
        }
        return String(paragraph.prefix(200))
    }
}

/// A text selection captured in the page, with its anchor and context.
struct SelectionExcerpt: Codable, Sendable {
    let text: String
    let domPath: String?
    let section: String?
}