添加xlsxworkflow

This commit is contained in:
xunbu
2025-08-04 18:35:26 +08:00
parent 141de321db
commit c364752b2d
4 changed files with 79 additions and 7 deletions

View File

@@ -1,6 +1,5 @@
from abc import abstractmethod
from dataclasses import dataclass
from logging import Logger
from typing import TypeVar
from docutranslate.agents.agent import ThinkingMode

View File

@@ -12,7 +12,7 @@ from docutranslate.translator.base import Translator
@dataclass
class XlsxTranslatorConfig(AiTranslatorConfig):
position: Literal["replace", "append", "prepend"] = "replace"
insert_mode: Literal["replace", "append", "prepend"] = "replace"
separator: str = "\n"
@@ -32,7 +32,7 @@ class XlsxTranslator(Translator):
timeout=config.timeout,
logger=self.logger)
self.translate_agent = SegmentsTranslateAgent(agent_config)
self.position = config.position
self.insert_mode = config.insert_mode
self.separator = config.separator
def _pre_translate(self, document: Document):
@@ -66,11 +66,11 @@ class XlsxTranslator(Translator):
# 定位到工作表和单元格
sheet = workbook[sheet_name]
if self.position == "replace":
if self.insert_mode == "replace":
sheet[coordinate] = translated_text
elif self.position == "append":
elif self.insert_mode == "append":
sheet[coordinate] = original_text + self.separator + translated_text
elif self.position == "prepend":
elif self.insert_mode == "prepend":
sheet[coordinate] = translated_text + self.separator + original_text
else:
self.logger.error("不正确的XlsxTranslatorConfig参数")
@@ -120,7 +120,7 @@ if __name__ == '__main__':
api_key=r"969ba51b61914cc2b710d1393dca1a3c.hSuATex5IoNVZNGu",
model_id=r"glm-4-flash",
to_lang="英文",
position="append"
insert_mode="append"
)
translator = XlsxTranslator(config)
document = Document.from_path(r"C:\Users\jxgm\Desktop\translate\docutranslate\tests\files\工业互联分组表.xlsx")

View File

@@ -55,3 +55,11 @@ class JsonExportable(Protocol[T_ExporterConfig]):
def save_as_json(self, name: str, output_dir: Path | str, config: T_ExporterConfig | None = None) -> Self:
...
@runtime_checkable
class XlsxExportable(Protocol[T_ExporterConfig]):
def export_to_xlsx(self, config: T_ExporterConfig | None = None) -> bytes:
...
def save_as_xlsx(self, name: str, output_dir: Path | str, config: T_ExporterConfig | None = None) -> Self:
...

View File

@@ -0,0 +1,65 @@
from dataclasses import dataclass
from pathlib import Path
from typing import Self
from docutranslate.exporter.base import ExporterConfig
from docutranslate.exporter.xlsx.xlsx2html_exporter import Xlsx2HTMLExporterConfig, Xlsx2HTMLExporter
from docutranslate.exporter.xlsx.xlsx2xlsx_exporter import Xlsx2XlsxExporter
from docutranslate.ir.document import Document
from docutranslate.translator.ai_translator.xlsx_translator import XlsxTranslatorConfig, XlsxTranslator
from docutranslate.workflow.base import Workflow, WorkflowConfig
from docutranslate.workflow.interfaces import HTMLExportable, XlsxExportable
@dataclass(kw_only=True)
class XlsxWorkflowConfig(WorkflowConfig):
translator_config: XlsxTranslatorConfig
html_exporter_config: Xlsx2HTMLExporterConfig
class XlsxWorkflow(Workflow[XlsxWorkflowConfig, Document, Document], HTMLExportable[Xlsx2HTMLExporterConfig],
XlsxExportable[ExporterConfig]):
def __init__(self, config: XlsxWorkflowConfig):
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 = XlsxTranslator(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: Xlsx2HTMLExporterConfig = None) -> str:
config = config or self.config.html_exporter_config
docu = self._export(Xlsx2HTMLExporter(config))
return docu.content.decode()
def export_to_xlsx(self, _: ExporterConfig | None = None) -> bytes:
docu = self._export(Xlsx2XlsxExporter())
return docu.content
def save_as_html(self, name: str = None, output_dir: Path | str = "./output",
config: Xlsx2HTMLExporter | None = None) -> Self:
config = config or self.config.html_exporter_config
self._save(exporter=Xlsx2HTMLExporter(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=Xlsx2XlsxExporter(), name=name, output_dir=output_dir)
return self