尝试增加对docx的支持

docutranslate/exporter/docx/base.py

@@ -0,0 +1,8 @@
+from docutranslate.exporter.base import Exporter
+from docutranslate.ir.document import Document
+
+#TODO:看情况是否需要为json单独写一个document类型
+class DocxExporter(Exporter[Document]):
+
+    def export(self,document:Document)->Document:
+        ...

docutranslate/exporter/docx/docx2docx_exporter.py

@@ -0,0 +1,7 @@
+from docutranslate.exporter.docx.base import DocxExporter
+from docutranslate.ir.document import Document
+
+
+class Docx2DocxExporter(DocxExporter):
+    def export(self, document: Document) -> Document:
+        return document.copy()

docutranslate/exporter/docx/docx2html_exporter.py

@@ -0,0 +1,25 @@
+from dataclasses import dataclass
+from io import BytesIO
+
+import mammoth
+
+from docutranslate.exporter.base import ExporterConfig
+from docutranslate.exporter.xlsx.base import XlsxExporter
+from docutranslate.ir.document import Document
+
+
+@dataclass
+class Docx2HTMLExporterConfig(ExporterConfig):
+    cdn: bool = True
+
+
+class Docx2HTMLExporter(XlsxExporter):
+    def __init__(self, config: Docx2HTMLExporterConfig = None):
+        config = config or Docx2HTMLExporterConfig()
+        super().__init__(config=config)
+        self.cdn = config.cdn
+
+    def export(self, document: Document) -> Document:
+        html_content = mammoth.convert_to_html(BytesIO(document.content))
+
+        return Document.from_bytes(content=html_content.encode("utf-8"), suffix=".html", stem=document.stem)

docutranslate/exporter/xlsx/xlsx2html_exporter.py

@@ -1,14 +1,11 @@
 from dataclasses import dataclass
 from io import BytesIO
 
-import jinja2
-import openpyxl
 from xlsx2html import xlsx2html
 
 from docutranslate.exporter.base import ExporterConfig
 from docutranslate.exporter.xlsx.base import XlsxExporter
 from docutranslate.ir.document import Document
-from docutranslate.utils.resource_utils import resource_path
 
 
 @dataclass

docutranslate/translator/ai_translator/docx_translator.py

@@ -0,0 +1,157 @@
+import asyncio
+from dataclasses import dataclass
+from io import BytesIO
+from typing import Self, Literal, List, Dict, Any, Tuple
+
+import docx
+from docx.document import Document as DocumentObject
+from docx.table import _Cell
+
+from docutranslate.agents.segments_agent import SegmentsTranslateAgentConfig, SegmentsTranslateAgent
+from docutranslate.ir.document import Document
+from docutranslate.translator.ai_translator.base import AiTranslatorConfig
+from docutranslate.translator.base import Translator
+
+
+@dataclass
+class DocxTranslatorConfig(AiTranslatorConfig):
+    """
+    DocxTranslator 的配置类。
+    """
+    insert_mode: Literal["replace", "append", "prepend"] = "replace"
+    separator: str = "\n"
+
+
+class DocxTranslator(Translator):
+    """
+    用于翻译 .docx 文件的翻译器。
+    """
+
+    def __init__(self, config: DocxTranslatorConfig):
+        super().__init__(config=config)
+        self.chunk_size = config.chunk_size
+        agent_config = SegmentsTranslateAgentConfig(
+            custom_prompt=config.custom_prompt,
+            to_lang=config.to_lang,
+            baseurl=config.base_url,
+            key=config.api_key,
+            model_id=config.model_id,
+            system_prompt=None,
+            temperature=config.temperature,
+            thinking=config.thinking,
+            max_concurrent=config.concurrent,
+            timeout=config.timeout,
+            logger=self.logger
+        )
+        self.translate_agent = SegmentsTranslateAgent(agent_config)
+        self.insert_mode = config.insert_mode
+        self.separator = config.separator
+
+    def _pre_translate(self, document: Document) -> Tuple[DocumentObject, List[Dict[str, Any]], List[str]]:
+        """
+        预处理 .docx 文件，提取所有需要翻译的文本。
+
+        :param document: 包含 .docx 文件内容的 Document 对象。
+        :return: 一个元组，包含：
+                 - docx.Document 对象
+                 - 一个包含文本元素信息的列表 (e.g., paragraph, cell)
+                 - 一个包含所有待翻译原文的列表
+        """
+        doc = docx.Document(BytesIO(document.content))
+        elements_to_translate = []
+        original_texts = []
+
+        # 遍历所有段落
+        for para in doc.paragraphs:
+            if para.text.strip():  # 确保段落有实际内容
+                elements_to_translate.append({"type": "paragraph", "element": para})
+                original_texts.append(para.text)
+
+        # 遍历所有表格
+        for table in doc.tables:
+            for row in table.rows:
+                for cell in row.cells:
+                    if cell.text.strip():  # 确保单元格有实际内容
+                        elements_to_translate.append({"type": "cell", "element": cell})
+                        original_texts.append(cell.text)
+
+        return doc, elements_to_translate, original_texts
+
+    def _after_translate(self, doc: DocumentObject, elements_to_translate: List[Dict[str, Any]],
+                         translated_texts: List[str], original_texts: List[str]) -> bytes:
+        """
+        将翻译后的文本写回到 .docx 对象中。
+
+        :param doc: docx.Document 对象。
+        :param elements_to_translate: 包含文本元素信息的列表。
+        :param translated_texts: 翻译后的文本列表。
+        :param original_texts: 原始文本列表。
+        :return: 更新后的 .docx 文件内容的字节流。
+        """
+        for i, element_info in enumerate(elements_to_translate):
+            element = element_info["element"]
+            original_text = original_texts[i]
+            translated_text = translated_texts[i]
+
+            # 清空原有内容并写入新内容
+            if isinstance(element, docx.text.paragraph.Paragraph):
+                # 清空段落内容
+                element.clear()
+                # 根据插入模式添加文本
+                if self.insert_mode == "replace":
+                    element.add_run(translated_text)
+                elif self.insert_mode == "append":
+                    element.add_run(original_text + self.separator + translated_text)
+                elif self.insert_mode == "prepend":
+                    element.add_run(translated_text + self.separator + original_text)
+                else:
+                    self.logger.error("不正确的DocxTranslatorConfig参数")
+
+            elif isinstance(element, _Cell):
+                # 根据插入模式设置单元格文本
+                if self.insert_mode == "replace":
+                    element.text = translated_text
+                elif self.insert_mode == "append":
+                    element.text = original_text + self.separator + translated_text
+                elif self.insert_mode == "prepend":
+                    element.text = translated_text + self.separator + original_text
+                else:
+                    self.logger.error("不正确的DocxTranslatorConfig参数")
+
+        # 将修改后的文档保存到 BytesIO 流
+        doc_output_stream = BytesIO()
+        doc.save(doc_output_stream)
+        return doc_output_stream.getvalue()
+
+    def translate(self, document: Document) -> Self:
+        """
+        同步翻译 .docx 文件。
+        """
+        doc, elements_to_translate, original_texts = self._pre_translate(document)
+        if not elements_to_translate:
+            print("\n文件中没有找到需要翻译的文本内容。")
+            return self
+
+        # 调用翻译 agent
+        translated_texts = self.translate_agent.send_segments(original_texts, self.chunk_size)
+
+        # 将翻译结果写回文档
+        document.content = self._after_translate(doc, elements_to_translate, translated_texts, original_texts)
+        return self
+
+    async def translate_async(self, document: Document) -> Self:
+        """
+        异步翻译 .docx 文件。
+        """
+        doc, elements_to_translate, original_texts = await asyncio.to_thread(self._pre_translate, document)
+        if not elements_to_translate:
+            print("\n文件中没有找到需要翻译的文本内容。")
+            return self
+
+        # 异步调用翻译 agent
+        translated_texts = await self.translate_agent.send_segments_async(original_texts, self.chunk_size)
+
+        # 将翻译结果写回文档
+        document.content = await asyncio.to_thread(self._after_translate, doc, elements_to_translate, translated_texts,
+                                                   original_texts)
+        return self

docutranslate/workflow/docx_workflow.py

@@ -0,0 +1,66 @@
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Self
+
+from docutranslate.exporter.base import ExporterConfig
+from docutranslate.exporter.docx.docx2docx_exporter import Docx2DocxExporter
+from docutranslate.exporter.docx.docx2html_exporter import Docx2HTMLExporterConfig, Docx2HTMLExporter
+
+from docutranslate.ir.document import Document
+from docutranslate.translator.ai_translator.docx_translator import DocxTranslatorConfig, DocxTranslator
+from docutranslate.workflow.base import Workflow, WorkflowConfig
+from docutranslate.workflow.interfaces import HTMLExportable, DocxExportable
+
+
+@dataclass(kw_only=True)
+class DocxWorkflowConfig(WorkflowConfig):
+    translator_config: DocxTranslatorConfig
+    html_exporter_config: Docx2HTMLExporterConfig
+
+
+class DocxWorkflow(Workflow[DocxWorkflowConfig, Document, Document], HTMLExportable[Docx2HTMLExporterConfig],
+                   DocxExportable[ExporterConfig]):
+    def __init__(self, config: DocxWorkflowConfig):
+        super().__init__(config=config)
+        if config.logger:
+            for sub_config in [self.config.translator_config]:
+                if sub_config:
+                    sub_config.logger = config.logger
+
+    def _pre_translate(self, document_original: Document):
+        document = document_original.copy()
+        translate_config = self.config.translator_config
+        translator = DocxTranslator(translate_config)
+        return document, translator
+
+    def translate(self) -> Self:
+        document, translator = self._pre_translate(self.document_original)
+        translator.translate(document)
+        self.document_translated = document
+        return self
+
+    async def translate_async(self) -> Self:
+        document, translator = self._pre_translate(self.document_original)
+        await translator.translate_async(document)
+        self.document_translated = document
+        return self
+
+    def export_to_html(self, config: Docx2HTMLExporterConfig = None) -> str:
+        config = config or self.config.html_exporter_config
+        docu = self._export(Docx2HTMLExporter(config))
+        return docu.content.decode()
+
+    def export_to_xlsx(self, _: ExporterConfig | None = None) -> bytes:
+        docu = self._export(Docx2DocxExporter())
+        return docu.content
+
+    def save_as_html(self, name: str = None, output_dir: Path | str = "./output",
+                     config: Docx2HTMLExporter | None = None) -> Self:
+        config = config or self.config.html_exporter_config
+        self._save(exporter=Docx2HTMLExporter(config), name=name, output_dir=output_dir)
+        return self
+
+    def save_as_xlsx(self, name: str = None, output_dir: Path | str = "./output",
+                     _: ExporterConfig | None = None) -> Self:
+        self._save(exporter=Docx2DocxExporter(), name=name, output_dir=output_dir)
+        return self

docutranslate/workflow/interfaces.py

@@ -63,3 +63,11 @@ class XlsxExportable(Protocol[T_ExporterConfig]):
 
     def save_as_xlsx(self, name: str, output_dir: Path | str, config: T_ExporterConfig | None = None) -> Self:
         ...
+
+@runtime_checkable
+class DocxExportable(Protocol[T_ExporterConfig]):
+    def export_to_docx(self, config: T_ExporterConfig | None = None) -> bytes:
+        ...
+
+    def save_as_docx(self, name: str, output_dir: Path | str, config: T_ExporterConfig | None = None) -> Self:
+        ...

pyproject.toml

@@ -11,6 +11,7 @@ dependencies = [
     "openpyxl>=3.1.5",
     "xlsx2html>=0.6.2",
     "json-repair>=0.48.0",
+    "mammoth>=1.10.0",
 ]
 dynamic = ["version"]
 

uv.lock

@@ -142,6 +142,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/7e/d4/7ebdbd03970677812aac39c869717059dbb71a4cfc033ca6e5221787892c/click-8.1.8-py3-none-any.whl", hash = "sha256:63c132bbbed01578a06712a2d1f497bb62d9c1c0d329b7903a866228027263b2", size = 98188 },
 ]
 
+[[package]]
+name = "cobble"
+version = "0.1.4"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/54/7a/a507c709be2c96e1bb6102eb7b7f4026c5e5e223ef7d745a17d239e9d844/cobble-0.1.4.tar.gz", hash = "sha256:de38be1539992c8a06e569630717c485a5f91be2192c461ea2b220607dfa78aa", size = 3805 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/d5/e1/3714a2f371985215c219c2a70953d38e3eed81ef165aed061d21de0e998b/cobble-0.1.4-py3-none-any.whl", hash = "sha256:36c91b1655e599fd428e2b95fdd5f0da1ca2e9f1abb0bc871dec21a0e78a2b44", size = 3984 },
+]
+
 [[package]]
 name = "colorama"
 version = "0.4.6"
@@ -294,6 +303,7 @@ dependencies = [
     { name = "httpx" },
     { name = "json-repair" },
     { name = "jsonpath-ng" },
+    { name = "mammoth" },
     { name = "markdown2" },
     { name = "openpyxl" },
     { name = "xlsx2html" },
@@ -319,6 +329,7 @@ requires-dist = [
     { name = "httpx", specifier = "==0.27.2" },
     { name = "json-repair", specifier = ">=0.48.0" },
     { name = "jsonpath-ng", specifier = ">=1.7.0" },
+    { name = "mammoth", specifier = ">=1.10.0" },
     { name = "markdown2", specifier = ">=2.5.3" },
     { name = "opencv-python", marker = "extra == 'docling'", specifier = ">=4.11.0.86" },
     { name = "openpyxl", specifier = ">=3.1.5" },
@@ -727,6 +738,18 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/d1/5d/c059c180c84f7962db0aeae7c3b9303ed1d73d76f2bfbc32bc231c8be314/macholib-1.16.3-py2.py3-none-any.whl", hash = "sha256:0e315d7583d38b8c77e815b1ecbdbf504a8258d8b3e17b61165c6feb60d18f2c", size = 38094 },
 ]
 
+[[package]]
+name = "mammoth"
+version = "1.10.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "cobble" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/89/0d/2ab86f37021b4c50fe72354acd226b1e31a10497e51f6cbd7e3d1eca1181/mammoth-1.10.0.tar.gz", hash = "sha256:cb6fbba41ccf8b5502859c457177d87a833fef0e0b1d4e6fd23ec372fe892c30", size = 52285 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/a6/67/36eeb3a8726df3b282ba99ec126323871cffdbcf3b7a1db64ca9bbe4abc1/mammoth-1.10.0-py2.py3-none-any.whl", hash = "sha256:a1c87d5b98ca30230394267f98614b58b14b50f8031dc33ac9a535c6ab04eb99", size = 53823 },
+]
+
 [[package]]
 name = "markdown-it-py"
 version = "3.0.0"