ReadViewSDK/Sources/RDReaderView/EPUBCore/Notes/RDEPUBNoteResolver.swift

import Foundation

public final class RDEPUBNoteResolver {

    private let resourceResolver: RDEPUBResourceResolver

    public init(resourceResolver: RDEPUBResourceResolver) {
        self.resourceResolver = resourceResolver
    }

    public func resolveInternalLink(
        sourceHref: String,
        sourceCFI: String?,
        targetLocation: RDEPUBLocation,
        sourceLocation: RDEPUBLocation? = nil,
        relativeToSpineIndex spineIndex: Int?
    ) -> RDEPUBResolvedNote? {
        guard let normalizedTarget = resourceResolver.normalizedLocation(
            targetLocation,
            relativeToSpineIndex: spineIndex,
            bookIdentifier: targetLocation.bookIdentifier
        ) else {
            return nil
        }

        let fragment = normalizedTarget.fragment
        guard let elementHTML = extractElementHTML(href: normalizedTarget.href, fragment: fragment) else {
            return nil
        }

        guard let reference = RDEPUBNoteDetector.makeReferenceIfLikelyNote(
            sourceHref: sourceHref,
            sourceCFI: sourceCFI,
            targetHref: normalizedTarget.href,
            targetFragment: fragment,
            targetCFI: normalizedTarget.cfi,
            targetElementHTML: elementHTML
        ) else {
            return nil
        }

        let sanitizedHTML = stripBacklinks(from: elementHTML)
        let plainText = plainText(from: sanitizedHTML)
        guard !plainText.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty else {
            return nil
        }

        return RDEPUBResolvedNote(
            reference: reference,
            sourceLocation: sourceLocation,
            targetLocation: normalizedTarget,
            title: title(for: reference),
            html: sanitizedHTML,
            plainText: plainText
        )
    }

    public func extractElementHTML(href: String, fragment: String?) -> String? {
        guard let fragment,
              !fragment.isEmpty,
              let fileURL = resourceResolver.fileURL(forRelativePath: href),
              let html = chapterHTML(at: fileURL) else {
            return nil
        }

        return elementHTML(withID: fragment, in: html)
    }

    /// 读取章节 HTML：优先经加密资源 provider 解密，明文文件保持原直读路径
    private func chapterHTML(at fileURL: URL) -> String? {
        if let provided = resourceResolver.providedResourceData(at: fileURL) {
            return String(data: provided, encoding: .utf8)
                ?? String(data: provided, encoding: .utf16)
        }
        return try? String(contentsOf: fileURL, encoding: .utf8)
    }

    private func elementHTML(withID id: String, in html: String) -> String? {
        let escapedID = NSRegularExpression.escapedPattern(for: id)
        let pattern = #"<([A-Za-z][A-Za-z0-9:_-]*)(?=[^>]*(?:id|xml:id)\s*=\s*(['"])"# + escapedID + #"\2)[^>]*>"#
        guard let regex = try? NSRegularExpression(pattern: pattern, options: [.caseInsensitive]) else {
            return nil
        }
        let nsRange = NSRange(html.startIndex..<html.endIndex, in: html)
        guard let match = regex.firstMatch(in: html, options: [], range: nsRange),
              let startRange = Range(match.range, in: html),
              let tagRange = Range(match.range(at: 1), in: html) else {
            return nil
        }
        let tagName = String(html[tagRange])
        guard !String(html[startRange]).hasSuffix("/>") else {
            return String(html[startRange])
        }
        guard let endRange = matchingEndTagRange(
            tagName: tagName,
            in: html,
            after: startRange.upperBound
        ) else {
            return String(html[startRange])
        }
        return String(html[startRange.lowerBound..<endRange.upperBound])
    }

    private func stripBacklinks(from html: String) -> String {
        let backlinkPattern = #"<a\b[^>]*(?:rev\s*=\s*['"]footnote['"]|epub:type\s*=\s*['"][^'"]*(?:backlink|return)[^'"]*['"]|class\s*=\s*['"][^'"]*(?:backlink|return)[^'"]*['"])[^>]*>[\s\S]*?</a>"#
        guard let regex = try? NSRegularExpression(pattern: backlinkPattern, options: [.caseInsensitive]) else {
            return html
        }
        let nsRange = NSRange(html.startIndex..<html.endIndex, in: html)
        return regex.stringByReplacingMatches(in: html, options: [], range: nsRange, withTemplate: "")
    }

    private func plainText(from html: String) -> String {
        let withoutScripts = html
            .replacingOccurrences(of: #"<script[\s\S]*?</script>"#, with: "", options: .regularExpression)
            .replacingOccurrences(of: #"<style[\s\S]*?</style>"#, with: "", options: .regularExpression)
        let withoutTags = withoutScripts.replacingOccurrences(of: #"<[^>]+>"#, with: " ", options: .regularExpression)
        return withoutTags
            .replacingOccurrences(of: "&nbsp;", with: " ")
            .replacingOccurrences(of: "&lt;", with: "<")
            .replacingOccurrences(of: "&gt;", with: ">")
            .replacingOccurrences(of: "&amp;", with: "&")
            .components(separatedBy: .whitespacesAndNewlines)
            .filter { !$0.isEmpty }
            .joined(separator: " ")
    }

    private func title(for reference: RDEPUBNoteReference) -> String {
        switch reference.kind {
        case .footnote:
            return "脚注"
        case .endnote:
            return "尾注"
        case .rearnote:
            return "后注"
        case .unknown:
            return "注释"
        }
    }

    private func matchingEndTagRange(
        tagName: String,
        in html: String,
        after startIndex: String.Index
    ) -> Range<String.Index>? {
        let escapedTagName = NSRegularExpression.escapedPattern(for: tagName)
        let pattern = #"</?\s*"# + escapedTagName + #"\b[^>]*>"#
        guard let regex = try? NSRegularExpression(pattern: pattern, options: [.caseInsensitive]) else {
            return nil
        }

        let searchRange = NSRange(startIndex..<html.endIndex, in: html)
        var depth = 1
        for match in regex.matches(in: html, options: [], range: searchRange) {
            guard let range = Range(match.range, in: html) else { continue }
            let tag = String(html[range])
            if tag.hasPrefix("</") {
                depth -= 1
            } else if !tag.hasSuffix("/>") {
                depth += 1
            }
            if depth == 0 {
                return range
            }
        }
        return nil
    }
}