feat(wxread): align pagination, rendering, and docs

This commit is contained in:
shen
2026-05-24 15:36:01 +08:00
parent a318c0e3d0
commit 2a94921e88
103 changed files with 6124 additions and 4623 deletions
@@ -40,13 +40,29 @@ public struct RDEPUBTextChapter: Equatable {
public var pages: [RDEPUBTextPage]
}
public struct RDEPUBTextBook: Equatable {
public struct RDEPUBTextBook {
public var chapters: [RDEPUBTextChapter]
public var pages: [RDEPUBTextPage]
public let indexTable: RDEPUBTextIndexTable
public init(chapters: [RDEPUBTextChapter], pages: [RDEPUBTextPage]) {
self.chapters = chapters
self.pages = pages
self.indexTable = RDEPUBTextIndexTable(chapters: chapters)
}
public static func == (lhs: RDEPUBTextBook, rhs: RDEPUBTextBook) -> Bool {
lhs.chapters == rhs.chapters && lhs.pages == rhs.pages
}
public func chapterData(for href: String) -> RDEPUBChapterData? {
guard let chapter = chapters.first(where: { $0.href == href }) else { return nil }
return RDEPUBChapterData(chapter: chapter, indexTable: indexTable)
}
public func chapterData(atChapterIndex index: Int) -> RDEPUBChapterData? {
guard chapters.indices.contains(index) else { return nil }
return RDEPUBChapterData(chapter: chapters[index], indexTable: indexTable)
}
public func page(at pageNumber: Int) -> RDEPUBTextPage? {
@@ -62,6 +78,11 @@ public struct RDEPUBTextBook: Equatable {
return nil
}
if let anchor = normalizedLocation.rangeAnchor?.start,
let page = indexTable.pageNumber(for: anchor, in: self) {
return page + 1
}
let targetOffset: Int
if let fragment = normalizedLocation.fragment, let fragmentOffset = chapter.fragmentOffsets[fragment] {
targetOffset = fragmentOffset
@@ -84,12 +105,18 @@ public struct RDEPUBTextBook: Equatable {
}
let totalLength = max(chapter.attributedContent.length - 1, 1)
let startAnchor = indexTable.anchor(forAbsoluteIndex: page.pageStartOffset, in: chapter)
let endAnchor = indexTable.anchor(forAbsoluteIndex: page.pageEndOffset, in: chapter)
let rangeAnchor = RDEPUBTextRangeAnchor(start: startAnchor, end: endAnchor)
return RDEPUBLocation(
bookIdentifier: bookIdentifier,
href: page.href,
progression: Double(page.pageStartOffset) / Double(totalLength),
lastProgression: Double(page.pageEndOffset) / Double(totalLength),
fragment: nil
fragment: nil,
rangeAnchor: rangeAnchor
)
}
}
@@ -243,7 +270,12 @@ public final class RDEPUBTextBookBuilder {
}
let paginateDuration = CFAbsoluteTimeGetCurrent() - paginateStart
let effectiveFrames = layoutFrames.isEmpty && content.length > 0
let normalizedFrames = normalizeTrailingFrames(
layoutFrames,
content: content,
href: item.href
)
let effectiveFrames = normalizedFrames.isEmpty && content.length > 0
? [
RDEPUBTextLayoutFrame(
contentRange: NSRange(location: 0, length: content.length),
@@ -261,7 +293,7 @@ public final class RDEPUBTextBookBuilder {
]
)
]
: layoutFrames
: normalizedFrames
if item.href.lowercased().contains("cover") {
print("[EPUB][Cover] paginated href=\(item.href) pages=\(effectiveFrames.count) firstRange=\(effectiveFrames.first.map { NSStringFromRange($0.contentRange) } ?? "none")")
}
@@ -423,6 +455,111 @@ public final class RDEPUBTextBookBuilder {
return attachmentCount(in: content) > 0 && trimmed.count <= 1
}
private func normalizeTrailingFrames(
_ frames: [RDEPUBTextLayoutFrame],
content: NSAttributedString,
href: String
) -> [RDEPUBTextLayoutFrame] {
guard frames.count > 1 else { return frames }
var normalized = frames
while let lastFrame = normalized.last,
shouldDropWhitespaceOnlyTrailingFrame(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 {
print("[EPUB][Pagination] href=\(href) dropped trailing frame \(NSStringFromRange(lastFrame.contentRange))")
}
}
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 shouldDropWhitespaceOnlyTrailingFrame(
_ 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
}
let previousVisibleCount = visibleCharacterCount(in: content, range: previousFrame.contentRange)
return previousVisibleCount >= max(visibleCount * 8, 12)
}
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 uniqueValues<T: Equatable>(from values: [T]) -> [T] {
values.reduce(into: [T]()) { result, value in
if !result.contains(value) {
@@ -431,6 +568,25 @@ public final class RDEPUBTextBookBuilder {
}
}
private func uniqueRanges(from ranges: [NSRange]) -> [NSRange] {
ranges.reduce(into: [NSRange]()) { result, value in
if !result.contains(value) {
result.append(value)
}
}
}
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 makeCacheKey(
bookID: String,
pageSize: CGSize,