feat(epub): align wxread css and overlay pagination behavior

This commit is contained in:
shen
2026-05-25 22:15:24 +08:00
parent c0aac56083
commit 23182e8b5a
31 changed files with 2585 additions and 1691 deletions
@@ -193,7 +193,8 @@ enum RDEPUBTextRendererSupport {
rawHTML: String,
baseURL: URL?,
style: RDEPUBTextRenderStyle,
resourceResolver: RDEPUBResourceResolver?
resourceResolver: RDEPUBResourceResolver?,
contentLanguageCode: String? = nil
) -> RDEPUBTextChapterRenderRequest {
let normalizedHTML = injectPaginationSemanticMarkers(into: normalizeHTML(rawHTML))
let stylesheetHrefReplacements = inlineLinkedStyleSheets(
@@ -204,7 +205,9 @@ enum RDEPUBTextRendererSupport {
)
let layers = makeStyleSheetLayers(
style: style,
epubCSS: stylesheetHrefReplacements.inlinedCSS
epubCSS: stylesheetHrefReplacements.inlinedCSS,
contentLanguageCode: contentLanguageCode,
sourceHTML: rawHTML
)
let htmlWithBase = injectBaseHref(into: stylesheetHrefReplacements.html, baseURL: baseURL)
let htmlWithDefaultLayers = injectStyleTag(
@@ -293,11 +296,17 @@ enum RDEPUBTextRendererSupport {
private static func makeStyleSheetLayers(
style: RDEPUBTextRenderStyle,
epubCSS: String
epubCSS: String,
contentLanguageCode: String?,
sourceHTML: String
) -> [RDEPUBTextStyleSheetLayer] {
let useLatinReplace = prefersLatinLanguageCSS(
languageCode: contentLanguageCode,
sourceHTML: sourceHTML
)
var layers: [RDEPUBTextStyleSheetLayer] = [
.init(kind: .default, css: defaultCSS()),
.init(kind: .replace, css: replaceCSS())
.init(kind: .replace, css: replaceCSS(useLatinVariant: useLatinReplace))
]
if isDarkTheme(style: style) {
layers.append(.init(kind: .dark, css: darkCSS(style: style)))
@@ -310,65 +319,12 @@ enum RDEPUBTextRendererSupport {
}
private static func defaultCSS() -> String {
"""
html, body {
margin: 0;
padding: 0;
background: transparent;
}
body {
word-break: break-word;
overflow-wrap: break-word;
-webkit-text-size-adjust: 100%;
}
p, div, li, blockquote {
margin-top: 0;
margin-bottom: 0.75em;
}
"""
RDEPUBAssetRepository.string(for: .wxReadDefaultCSS)
}
private static func replaceCSS() -> String {
"""
img, svg, video, canvas {
max-width: 100%;
height: auto;
}
.frontCover,
.bodyPic,
.qrbodyPic,
figure {
text-align: center;
text-indent: 0;
}
.frontCover img,
.rd-front-cover-image,
.bodyPic img,
.qrbodyPic img,
figure img {
display: block;
margin-left: auto;
margin-right: auto;
}
.qqreader-footnote,
.s-pic,
.h-pic,
.g-pic {
display: inline;
vertical-align: middle;
}
h1, h2, h3, h4, h5, h6 {
page-break-after: avoid;
break-after: avoid;
}
pre, code {
white-space: pre-wrap;
word-break: break-word;
}
table {
max-width: 100%;
}
"""
private static func replaceCSS(useLatinVariant: Bool) -> String {
let asset: RDEPUBAsset = useLatinVariant ? .wxReadLatinReplaceCSS : .wxReadReplaceCSS
return RDEPUBAssetRepository.string(for: asset)
}
private static func normalizeAttachmentHTMLMarkers(in html: String) -> String {
@@ -517,7 +473,7 @@ enum RDEPUBTextRendererSupport {
private static func darkCSS(style: RDEPUBTextRenderStyle) -> String {
let background = style.backgroundColor?.ss_cssString ?? "rgba(0, 0, 0, 1.000)"
let text = style.textColor?.ss_cssString ?? "rgba(255, 255, 255, 1.000)"
return """
return RDEPUBAssetRepository.string(for: .wxReadDarkCSS) + "\n\n" + """
html, body {
background: \(background) !important;
color: \(text) !important;
@@ -556,6 +512,118 @@ enum RDEPUBTextRendererSupport {
return luminance < 0.5
}
private static func prefersLatinLanguageCSS(
languageCode: String?,
sourceHTML: String
) -> Bool {
let candidateCodes = inferredLanguageCodes(
explicitLanguageCode: languageCode,
sourceHTML: sourceHTML
)
if candidateCodes.contains(where: isExplicitLatinLanguageCode) {
return true
}
if candidateCodes.contains(where: isExplicitCJKLanguageCode) {
return false
}
let textSample = plainTextSample(from: sourceHTML)
guard !textSample.isEmpty else { return false }
var alphabeticCount = 0
var latinCount = 0
for scalar in textSample.unicodeScalars {
guard CharacterSet.letters.contains(scalar) else { continue }
alphabeticCount += 1
if isLatinScalar(scalar) {
latinCount += 1
}
}
guard alphabeticCount >= 80 else { return false }
return (Double(latinCount) / Double(alphabeticCount)) >= 0.6
}
private static func inferredLanguageCodes(
explicitLanguageCode: String?,
sourceHTML: String
) -> [String] {
var codes: [String] = []
if let explicitLanguageCode {
let normalized = explicitLanguageCode.trimmingCharacters(in: .whitespacesAndNewlines).lowercased()
if !normalized.isEmpty {
codes.append(normalized)
}
}
if let regex = try? NSRegularExpression(
pattern: #"\b(?:xml:lang|lang)\s*=\s*["']([^"']+)["']"#,
options: [.caseInsensitive]
) {
let nsHTML = sourceHTML as NSString
let range = NSRange(location: 0, length: min(nsHTML.length, 8_000))
for match in regex.matches(in: sourceHTML, options: [], range: range) {
guard match.numberOfRanges > 1 else { continue }
let code = nsHTML.substring(with: match.range(at: 1))
.trimmingCharacters(in: .whitespacesAndNewlines)
.lowercased()
if !code.isEmpty {
codes.append(code)
}
}
}
return Array(NSOrderedSet(array: codes)) as? [String] ?? codes
}
private static func isExplicitLatinLanguageCode(_ code: String) -> Bool {
let normalized = code.lowercased()
if normalized.contains("latn") {
return true
}
let prefix = normalized.split(separator: "-").first.map(String.init) ?? normalized
let latinPrefixes: Set<String> = [
"en", "fr", "de", "es", "it", "pt", "nl", "sv", "da", "no", "fi",
"is", "ga", "cy", "pl", "cs", "sk", "sl", "hr", "hu", "ro", "tr",
"vi", "id", "ms", "tl", "sw", "af", "sq", "et", "lv", "lt"
]
return latinPrefixes.contains(prefix)
}
private static func isExplicitCJKLanguageCode(_ code: String) -> Bool {
let prefix = code.lowercased().split(separator: "-").first.map(String.init) ?? code.lowercased()
return ["zh", "ja", "ko"].contains(prefix)
}
private static func plainTextSample(from html: String) -> String {
let maxLength = min(html.count, 20_000)
let sample = String(html.prefix(maxLength))
let withoutTags = sample.replacingOccurrences(
of: #"<[^>]+>"#,
with: " ",
options: .regularExpression
)
return withoutTags.replacingOccurrences(
of: #"&[A-Za-z0-9#]+;"#,
with: " ",
options: .regularExpression
)
}
private static func isLatinScalar(_ scalar: UnicodeScalar) -> Bool {
switch scalar.value {
case 0x0041...0x007A,
0x00C0...0x00FF,
0x0100...0x024F,
0x1E00...0x1EFF:
return true
default:
return false
}
}
private static func injectBaseHref(into html: String, baseURL: URL?) -> String {
guard let baseURL else {
return html