From c9ffa8f19590c299d443cb571f22424398aee3f0 Mon Sep 17 00:00:00 2001 From: xunbu Date: Wed, 17 Dec 2025 21:00:36 +0800 Subject: [PATCH] =?UTF-8?q?=E4=BF=AE=E5=A4=8Depub=E6=8E=92=E7=89=88?= =?UTF-8?q?=E9=97=AE=E9=A2=98?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docutranslate/__init__.py | 2 +- .../ai_translator/epub_translator.py | 90 ++++++++++++++----- 2 files changed, 68 insertions(+), 24 deletions(-) diff --git a/docutranslate/__init__.py b/docutranslate/__init__.py index d255e6a..dbaaa05 100644 --- a/docutranslate/__init__.py +++ b/docutranslate/__init__.py @@ -1,3 +1,3 @@ # SPDX-FileCopyrightText: 2025 QinHan # SPDX-License-Identifier: MPL-2.0 -__version__="1.5.6a1" \ No newline at end of file +__version__="1.5.6a2" \ No newline at end of file diff --git a/docutranslate/translator/ai_translator/epub_translator.py b/docutranslate/translator/ai_translator/epub_translator.py index 61c1712..ae3a073 100644 --- a/docutranslate/translator/ai_translator/epub_translator.py +++ b/docutranslate/translator/ai_translator/epub_translator.py @@ -18,7 +18,7 @@ from docutranslate.translator.ai_translator.base import AiTranslatorConfig, AiTr @dataclass class EpubTranslatorConfig(AiTranslatorConfig): insert_mode: Literal["replace", "append", "prepend"] = "replace" - # 建议使用 \n，代码会将其转换为
，更灵活 + # 建议使用 \n，代码会将其转换为
或 换行，更灵活 separator: str = "\n" @@ -26,8 +26,10 @@ class EpubTranslator(AiTranslator): """ 一个用于翻译 EPUB 文件中内容的翻译器。 【高级版】此版本直接翻译HTML内容，以保留内联格式，并支持表格翻译。 - 【结构化修改版 v2】借鉴 DocxTranslator 的实现，在 append/prepend 模式下， - 对常规块级元素创建新标签存放译文，对表格单元格则在内部追加内容，以保证文档结构的正确性。 + 【结构化修改版 v3】 + 1. 修复了 BeautifulSoup 自动添加 导致嵌套错误的问题。 + 2. 对 li, div, td, th 采用内部追加模式，保护文档结构。 + 3. 复制标签时自动移除 ID 属性，防止锚点冲突。 """ def __init__(self, config: EpubTranslatorConfig): @@ -66,10 +68,12 @@ class EpubTranslator(AiTranslator): items_to_translate = [] original_texts = [] + # 读取 Zip 内容 with zipfile.ZipFile(BytesIO(document.content), 'r') as zf: for filename in zf.namelist(): all_files[filename] = zf.read(filename) + # 解析 container.xml 寻找 OPF container_xml = all_files.get('META-INF/container.xml') if not container_xml: raise ValueError("无效的 EPUB：找不到 META-INF/container.xml") @@ -78,6 +82,7 @@ class EpubTranslator(AiTranslator): opf_path = root.find('cn:rootfiles/cn:rootfile', ns).get('full-path') opf_dir = os.path.dirname(opf_path) + # 解析 OPF 获取文件清单 opf_xml = all_files.get(opf_path) if not opf_xml: raise ValueError(f"无效的 EPUB：找不到 {opf_path}") @@ -91,6 +96,7 @@ class EpubTranslator(AiTranslator): full_href = os.path.join(opf_dir, href).replace('\\', '/') manifest_items[item_id] = {'href': full_href, 'media_type': item.get('media-type')} + # 定义需要翻译的标签 TAGS_TO_TRANSLATE = ['p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'li', 'div', 'td', 'th'] for item_id, item_data in manifest_items.items(): @@ -112,6 +118,7 @@ class EpubTranslator(AiTranslator): tags_to_process = [] for tag in all_potential_tags: + # 仅处理叶子节点块（不包含其他待翻译块的块） contains_other_block = tag.find( lambda child_tag: child_tag in all_potential_tags_set and child_tag is not tag ) @@ -147,34 +154,54 @@ class EpubTranslator(AiTranslator): original_html = original_texts[i] translated_html = translated_texts[i] - # --- 关键逻辑：根据标签类型选择不同的处理策略 --- - is_table_cell = original_tag.name in ['td', 'th'] + # --- 核心修复：解析并剥离 HTML/BODY 外壳 --- + # BeautifulSoup(html, 'lxml') 会自动补全 ，必须剥离 + new_content_soup = BeautifulSoup(translated_html, 'lxml') + content_nodes = [] + if new_content_soup.body: + content_nodes = list(new_content_soup.body.contents) + elif new_content_soup.html: + content_nodes = list(new_content_soup.html.contents) + else: + content_nodes = list(new_content_soup.contents) + + # 策略 A: 替换模式（Replace） if self.insert_mode == "replace": original_tag.clear() - new_content_soup = BeautifulSoup(translated_html, 'lxml') - for node in list(new_content_soup.children): + for node in content_nodes: original_tag.append(node.extract()) + continue - elif is_table_cell: - # --- 表格单元格处理：在标签内部组合内容 --- + # 策略 B: 容器型元素 (td, th, li, div) -> 在内部追加 + # 这样可以保护 ul/ol 结构不被打断，div 布局不被破坏 + is_container_node = original_tag.name in ['td', 'th', 'li', 'div'] + + if is_container_node: original_tag.clear() - # 解析HTML片段 - original_nodes = BeautifulSoup(original_html, 'lxml').contents - translated_nodes = BeautifulSoup(translated_html, 'lxml').contents + # 解析原文 (注意防范原文被解析出多余标签) + orig_soup = BeautifulSoup(original_html, 'lxml') + original_nodes = list(orig_soup.body.contents) if orig_soup.body else list(orig_soup.contents) - # 创建分隔符节点 + # 译文节点使用上面剥离好的 content_nodes + translated_nodes = content_nodes + + # 构建分隔符 (容器内部使用 span/br，不使用 p/div 以防破坏行内流) separator_nodes = [] if self.separator: lines = self.separator.split('\n') for j, line in enumerate(lines): if line: - separator_nodes.append(NavigableString(line)) + sep_span = soup.new_tag('span', attrs={'class': 'translate-separator'}) + sep_span.string = line + separator_nodes.append(sep_span) if j < len(lines) - 1: separator_nodes.append(soup.new_tag('br')) + # 额外加一个换行区分原文和译文 + separator_nodes.append(soup.new_tag('br')) - # 根据模式按顺序重新填充 + # 组装内容 if self.insert_mode == "append": nodes_order = [original_nodes, separator_nodes, translated_nodes] else: # prepend @@ -182,18 +209,32 @@ class EpubTranslator(AiTranslator): for node_list in nodes_order: for node in node_list: - original_tag.append(node.extract() if isinstance(node, Tag) else node) + if node: + original_tag.append(node.extract() if isinstance(node, Tag) else node) + # 策略 C: 独立文本块 (p, h1-h6) -> 创建兄弟标签 else: - # --- 常规块级元素处理：创建新标签 --- - translated_tag = soup.new_tag(original_tag.name, attrs=original_tag.attrs) - new_content_soup = BeautifulSoup(translated_html, 'lxml') - for node in list(new_content_soup.children): - translated_tag.append(node.extract()) + # 复制属性，但必须删除 ID 以防冲突 + new_attrs = dict(original_tag.attrs) + if 'id' in new_attrs: + del new_attrs['id'] + translated_tag = soup.new_tag(original_tag.name, attrs=new_attrs) + + # 填充译文 + for node in content_nodes: + # 额外检查：防止

里面套

+ # 如果译文被识别为

翻译

，而外层容器也是

，则只取内部文本 + if isinstance(node, Tag) and node.name == original_tag.name and node.name == 'p': + for inner_child in list(node.contents): + translated_tag.append(inner_child.extract()) + else: + translated_tag.append(node.extract()) + + # 创建块级分隔符 separator_tag = None if self.separator: - separator_tag = soup.new_tag('p') + separator_tag = soup.new_tag('div', attrs={'class': 'translate-separator'}) lines = self.separator.split('\n') for j, line in enumerate(lines): if line: @@ -202,16 +243,19 @@ class EpubTranslator(AiTranslator): separator_tag.append(soup.new_tag('br')) if self.insert_mode == "append": + # 插入顺序：原文 -> 分隔符 -> 译文 current_node = original_tag if separator_tag: current_node.insert_after(separator_tag) current_node = separator_tag current_node.insert_after(translated_tag) elif self.insert_mode == "prepend": + # 插入顺序：译文 -> 分隔符 -> 原文 original_tag.insert_before(translated_tag) if separator_tag: translated_tag.insert_after(separator_tag) + # 重新打包 EPUB for file_path, soup in soups.items(): all_files[file_path] = str(soup).encode('utf-8') @@ -266,4 +310,4 @@ class EpubTranslator(AiTranslator): document.content = await asyncio.to_thread( self._after_translate, all_files, soups, items_to_translate, translated_texts, original_texts ) - return self + return self \ No newline at end of file