Files
ReadViewSDK/Sources/RDReaderView/EPUBTextRendering/RDPlainTextBookBuilder.swift
T
2026-05-21 19:40:51 +08:00

130 lines
5.1 KiB
Swift

import UIKit
/// 将纯文本文件构建为 RDEPUBTextBook,复用 EPUB 文本渲染和分页管线
public final class RDPlainTextBookBuilder {
private let renderer: RDEPUBTextRenderer
public init(renderer: RDEPUBTextRenderer = RDEPUBDTCoreTextRenderer()) {
self.renderer = renderer
}
public func build(
textFileURL: URL,
pageSize: CGSize,
style: RDEPUBTextRenderStyle
) throws -> RDEPUBTextBook {
let rawText = rd_decodeTextFile(url: textFileURL)
let chapterSpecs = splitChapters(from: rawText)
var chapters: [RDEPUBTextChapter] = []
var flatPages: [RDEPUBTextPage] = []
for (index, spec) in chapterSpecs.enumerated() {
let html = wrapTextAsHTML(spec.content)
let rendered = try renderer.renderChapter(html: html, baseURL: nil, style: style)
let content = NSMutableAttributedString(attributedString: rendered.attributedString)
let pageRanges = content.length > 0 ? content.ss_pageRanges(size: pageSize) : []
let effectivePageRanges = pageRanges.isEmpty && content.length > 0
? [NSRange(location: 0, length: content.length)]
: pageRanges
let href = "chapter_\(index).xhtml"
let pages = effectivePageRanges.enumerated().map { localPageIndex, range in
RDEPUBTextPage(
absolutePageIndex: flatPages.count + localPageIndex,
chapterIndex: index,
spineIndex: index,
href: href,
chapterTitle: spec.title ?? "第 \(index + 1) 章",
pageIndexInChapter: localPageIndex,
totalPagesInChapter: effectivePageRanges.count,
content: content.attributedSubstring(from: range),
contentRange: range,
pageStartOffset: range.location,
pageEndOffset: range.location + max(range.length - 1, 0)
)
}
chapters.append(
RDEPUBTextChapter(
chapterIndex: index,
spineIndex: index,
href: href,
title: spec.title ?? "第 \(index + 1) 章",
attributedContent: content.copy() as! NSAttributedString,
fragmentOffsets: [:],
pages: pages
)
)
flatPages.append(contentsOf: pages)
}
return RDEPUBTextBook(chapters: chapters, pages: flatPages)
}
// MARK: - 章节拆分
private struct ChapterSpec {
let title: String?
let content: String
}
/// 按中文章节正则拆分文本。无匹配时整体作为一章。
private func splitChapters(from text: String) -> [ChapterSpec] {
let pattern = #"^(第[零一二三四五六七八九十百千万\d]+[章节回卷].*)$"#
guard let regex = try? NSRegularExpression(pattern: pattern, options: [.anchorsMatchLines]) else {
return [ChapterSpec(title: nil, content: text)]
}
let nsText = text as NSString
let matches = regex.matches(in: text, range: NSRange(location: 0, length: nsText.length))
guard !matches.isEmpty else {
return [ChapterSpec(title: nil, content: text)]
}
var specs: [ChapterSpec] = []
for (i, match) in matches.enumerated() {
let titleRange = match.range(at: 1)
let title = nsText.substring(with: titleRange).trimmingCharacters(in: .whitespacesAndNewlines)
let contentStart = titleRange.location + titleRange.length
let contentEnd = (i + 1 < matches.count) ? matches[i + 1].range.location : nsText.length
let contentRange = NSRange(location: contentStart, length: contentEnd - contentStart)
let content = nsText.substring(with: contentRange).trimmingCharacters(in: .whitespacesAndNewlines)
if !content.isEmpty {
specs.append(ChapterSpec(title: title, content: content))
}
}
// 拆分结果为空(所有章节内容为空),整体作为一章
if specs.isEmpty {
return [ChapterSpec(title: nil, content: text)]
}
return specs
}
// MARK: - HTML 包装
/// 将纯文本包装为 HTML 段落
private func wrapTextAsHTML(_ text: String) -> String {
let paragraphs = text.components(separatedBy: "\n").filter { !$0.isEmpty }
let body = paragraphs.map { "<p>\($0)</p>" }.joined(separator: "\n")
return "<html><body>\(body)</body></html>"
}
// MARK: - 文本解码
private func rd_decodeTextFile(url: URL) -> String {
if let content = try? NSString(contentsOf: url, encoding: String.Encoding.utf8.rawValue) as String {
return content
}
if let content = try? NSString(contentsOf: url, encoding: 0x80000632) as String {
return content
}
if let content = try? NSString(contentsOf: url, encoding: 0x80000631) as String {
return content
}
return ""
}
}