Leon/docutranslate
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

AiTranslatorConfig添加skip_translate参数

Browse Source
This commit is contained in:
xunbu
2025-08-28 20:35:36 +08:00
parent 7e1fb345c6
commit 7d71e6382c
11 changed files with 237 additions and 150 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
docutranslate/translator/ai_translator/md_translator.py
View File

@@ -19,18 +19,20 @@ class MDTranslator(AiTranslator):
def __init__(self, config: MDTranslatorConfig):
super().__init__(config=config)
self.chunk_size = config.chunk_size
agent_config = MDTranslateAgentConfig(custom_prompt=config.custom_prompt,
to_lang=config.to_lang,
baseurl=config.base_url,
key=config.api_key,
model_id=config.model_id,
temperature=config.temperature,
thinking=config.thinking,
max_concurrent=config.concurrent,
timeout=config.timeout,
logger=self.logger,
glossary_dict=config.glossary_dict)
self.translate_agent = MDTranslateAgent(agent_config)
self.translate_agent = None
if not self.skip_translate:
agent_config = MDTranslateAgentConfig(custom_prompt=config.custom_prompt,
to_lang=config.to_lang,
baseurl=config.base_url,
key=config.api_key,
model_id=config.model_id,
temperature=config.temperature,
thinking=config.thinking,
max_concurrent=config.concurrent,
timeout=config.timeout,
logger=self.logger,
glossary_dict=config.glossary_dict)
self.translate_agent = MDTranslateAgent(agent_config)
def translate(self, document: MarkdownDocument) -> Self:
self.logger.info("正在翻译markdown")
@@ -38,9 +40,13 @@ class MDTranslator(AiTranslator):
chunks: list[str] = split_markdown_text(document.content.decode(), self.chunk_size)
if self.glossary_agent:
self.glossary_dict_gen = self.glossary_agent.send_segments(chunks, self.chunk_size)
self.translate_agent.update_glossary_dict(self.glossary_dict_gen)
if self.translate_agent:
self.translate_agent.update_glossary_dict(self.glossary_dict_gen)
self.logger.info(f"markdown分为{len(chunks)}块")
result: list[str] = self.translate_agent.send_chunks(chunks)
if self.translate_agent:
result: list[str] = self.translate_agent.send_chunks(chunks)
else:
result = chunks
content = join_markdown_texts(result)
# 做一些加强鲁棒性的操作
content = content.replace(r'\（', r'\(')
@@ -57,10 +63,14 @@ class MDTranslator(AiTranslator):
if self.glossary_agent:
self.glossary_dict_gen = await self.glossary_agent.send_segments_async(chunks, self.chunk_size)
self.translate_agent.update_glossary_dict(self.glossary_dict_gen)
if self.translate_agent:
self.translate_agent.update_glossary_dict(self.glossary_dict_gen)
self.logger.info(f"markdown分为{len(chunks)}块")
result: list[str] = await self.translate_agent.send_chunks_async(chunks)
if self.translate_agent:
result: list[str] = await self.translate_agent.send_chunks_async(chunks)
else:
result = chunks
def run():
content = join_markdown_texts(result)