ReadViewSDK/Sources/RDReaderView/EPUBTextRendering/BuildPipeline/RDEPUBChapterTailNormalizer.swift

import Foundation

struct RDEPUBChapterTailNormalizer {

    func normalize(
        _ frames: [RDEPUBTextLayoutFrame],
        content: NSAttributedString,
        href: String
    ) -> [RDEPUBTextLayoutFrame] {

        guard frames.count > 1 else { return frames }

        var normalized = frames

        var compacted: [RDEPUBTextLayoutFrame] = []
        compacted.reserveCapacity(normalized.count)
        for frame in normalized {
            if shouldDropWhitespaceOnlyFrame(frame, in: content) {
                let note = "normalized: dropped whitespace-only intermediate page \(NSStringFromRange(frame.contentRange))"

                if var previous = compacted.popLast() {
                    previous.diagnostics.append(note)
                    compacted.append(previous)
                } else {

                    #if DEBUG
                    print("[EPUB][Pagination] href=\(href) dropped leading/intermediate whitespace frame \(NSStringFromRange(frame.contentRange))")
                    #endif
                }
                continue
            }
            compacted.append(frame)
        }
        normalized = compacted

        while let lastFrame = normalized.last,
              shouldDropWhitespaceOnlyFrame(lastFrame, in: content) {
            normalized.removeLast()
            let note = "normalized: dropped whitespace-only trailing page \(NSStringFromRange(lastFrame.contentRange))"
            if var previousFrame = normalized.popLast() {
                previousFrame.diagnostics.append(note)
                normalized.append(previousFrame)
            } else {
                #if DEBUG
                print("[EPUB][Pagination] href=\(href) dropped trailing frame \(NSStringFromRange(lastFrame.contentRange))")
                #endif
            }
        }

        guard normalized.count > 1,
              let lastFrame = normalized.last,
              let previousFrame = normalized.dropLast().last,
              shouldMergeShortTrailingFrame(lastFrame, previousFrame: previousFrame, in: content) else {
            return normalized
        }

        let mergedFrame = mergeTrailingFrame(previousFrame, with: lastFrame)
        normalized.removeLast(2)
        normalized.append(mergedFrame)
        return normalized
    }

    private func shouldDropWhitespaceOnlyFrame(
        _ frame: RDEPUBTextLayoutFrame,
        in content: NSAttributedString
    ) -> Bool {
        guard frame.contentRange.length > 0,
              attachmentCount(in: content, range: frame.contentRange) == 0 else {
            return false
        }
        return visibleCharacterCount(in: content, range: frame.contentRange) == 0
    }

    private func shouldMergeShortTrailingFrame(
        _ trailingFrame: RDEPUBTextLayoutFrame,
        previousFrame: RDEPUBTextLayoutFrame,
        in content: NSAttributedString
    ) -> Bool {

        guard trailingFrame.contentRange.length > 0,
              NSMaxRange(previousFrame.contentRange) == trailingFrame.contentRange.location else {
            return false
        }

        let visibleCount = visibleCharacterCount(in: content, range: trailingFrame.contentRange)
        let trailingAttachmentCount = attachmentCount(in: content, range: trailingFrame.contentRange)

        guard visibleCount <= 2,
              trailingFrame.contentRange.length <= 2,
              visibleCount > 0 || trailingAttachmentCount > 0 else {
            return false
        }

        guard !frameHasAvoidPageBreakInside(trailingFrame, in: content) else {
            return false
        }

        let previousVisibleCount = visibleCharacterCount(in: content, range: previousFrame.contentRange)
        return previousVisibleCount >= max(visibleCount * 8, 12)
    }

    private func frameHasAvoidPageBreakInside(
        _ frame: RDEPUBTextLayoutFrame,
        in content: NSAttributedString
    ) -> Bool {
        if frame.semanticHints.contains(.avoidPageBreakInside) {
            return true
        }

        guard content.length > 0,
              let safeRange = clampedRange(frame.contentRange, in: content),
              safeRange.length > 0 else {
            return false
        }

        var found = false
        content.enumerateAttribute(.rdPageSemanticHints, in: safeRange) { value, _, stop in
            guard let rawValue = value as? String else { return }
            let hints = rawValue
                .split(separator: ",")
                .compactMap { RDEPUBTextSemanticHint(rawValue: String($0)) }
            if hints.contains(.avoidPageBreakInside) {
                found = true
                stop.pointee = true
            }
        }
        return found
    }

    private func mergeTrailingFrame(
        _ previousFrame: RDEPUBTextLayoutFrame,
        with trailingFrame: RDEPUBTextLayoutFrame
    ) -> RDEPUBTextLayoutFrame {

        let mergedRange = NSRange(
            location: previousFrame.contentRange.location,
            length: NSMaxRange(trailingFrame.contentRange) - previousFrame.contentRange.location
        )

        return RDEPUBTextLayoutFrame(
            contentRange: mergedRange,
            breakReason: trailingFrame.breakReason,
            blockRange: trailingFrame.blockRange ?? previousFrame.blockRange,
            attachmentRanges: uniqueRanges(from: previousFrame.attachmentRanges + trailingFrame.attachmentRanges),
            attachmentKinds: uniqueValues(from: previousFrame.attachmentKinds + trailingFrame.attachmentKinds),
            blockKinds: uniqueValues(from: previousFrame.blockKinds + trailingFrame.blockKinds),
            semanticHints: uniqueValues(from: previousFrame.semanticHints + trailingFrame.semanticHints),
            attachmentPlacements: uniqueValues(from: previousFrame.attachmentPlacements + trailingFrame.attachmentPlacements),
            trailingFragmentID: trailingFrame.trailingFragmentID ?? previousFrame.trailingFragmentID,

            diagnostics: previousFrame.diagnostics
                + trailingFrame.diagnostics
                + ["normalized: merged short trailing page \(NSStringFromRange(trailingFrame.contentRange)) into previous page"]
        )
    }

    private func visibleCharacterCount(
        in content: NSAttributedString,
        range: NSRange
    ) -> Int {
        guard range.length > 0 else { return 0 }
        let string = content.attributedSubstring(from: range).string

        let filteredScalars = string.unicodeScalars.filter { scalar in
            !CharacterSet.whitespacesAndNewlines.contains(scalar)
                && !CharacterSet.controlCharacters.contains(scalar)
        }
        return filteredScalars.count
    }

    private func attachmentCount(in content: NSAttributedString, range: NSRange) -> Int {
        guard content.length > 0, range.length > 0 else { return 0 }
        var count = 0
        content.enumerateAttribute(.attachment, in: range) { value, _, _ in
            if value != nil {
                count += 1
            }
        }
        return count
    }

    private func clampedRange(_ range: NSRange, in content: NSAttributedString) -> NSRange? {
        guard range.location >= 0, range.length >= 0, range.location < content.length else {
            return nil
        }
        return NSRange(location: range.location, length: min(range.length, content.length - range.location))
    }

    private func uniqueValues<T: Equatable>(from values: [T]) -> [T] {
        values.reduce(into: [T]()) { result, value in
            if !result.contains(value) {
                result.append(value)
            }
        }
    }

    private func uniqueRanges(from ranges: [NSRange]) -> [NSRange] {
        ranges.reduce(into: [NSRange]()) { result, value in
            if !result.contains(value) {
                result.append(value)
            }
        }
    }
}