后端完成json翻译工作流(待检查)

This commit is contained in:
xunbu
2025-08-03 18:08:06 +08:00
parent f054a1e0a9
commit b41daddd4f
8 changed files with 189 additions and 34 deletions

View File

@@ -47,3 +47,11 @@ class TXTExportable(Protocol[T_ExporterConfig]):
def save_as_txt(self, name: str, output_dir: Path | str, config: T_ExporterConfig | None = None) -> Self:
...
@runtime_checkable
class JsonExportable(Protocol[T_ExporterConfig]):
def export_to_json(self, config: T_ExporterConfig | None = None) -> str:
...
def save_as_json(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.js.json2html_exporter import Json2HTMLExporterConfig, Json2HTMLExporter
from docutranslate.exporter.js.json2json_exporter import Json2JsonExporter
from docutranslate.ir.document import Document
from docutranslate.translator.ai_translator.json_translator import JsonTranslatorConfig, JsonTranslator
from docutranslate.workflow.base import Workflow, WorkflowConfig
from docutranslate.workflow.interfaces import HTMLExportable, JsonExportable
@dataclass(kw_only=True)
class JsonWorkflowConfig(WorkflowConfig):
translator_config: JsonTranslatorConfig
html_exporter_config: Json2HTMLExporterConfig
class JsonWorkflow(Workflow[JsonWorkflowConfig, Document, Document], HTMLExportable[Json2HTMLExporterConfig],
JsonExportable[ExporterConfig]):
def __init__(self, config: JsonWorkflowConfig):
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 = JsonTranslator(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: Json2HTMLExporterConfig = None) -> str:
config = config or self.config.html_exporter_config
docu = self._export(Json2HTMLExporter(config))
return docu.content.decode()
def export_to_json(self, _: ExporterConfig | None = None) -> str:
docu = self._export(Json2JsonExporter())
return docu.content.decode()
def save_as_html(self, name: str = None, output_dir: Path | str = "./output",
config: Json2HTMLExporter | None = None) -> Self:
config = config or self.config.html_exporter_config
self._save(exporter=Json2HTMLExporter(config), name=name, output_dir=output_dir)
return self
def save_as_json(self, name: str = None, output_dir: Path | str = "./output",
_: ExporterConfig | None = None) -> Self:
self._save(exporter=Json2JsonExporter(), name=name, output_dir=output_dir)
return self