195 lines
7.6 KiB
Python
195 lines
7.6 KiB
Python
"""
|
|
翻译处理模块：自动识别外文文章，通过AI并发生成中文翻译
|
|
"""
|
|
import re
|
|
import json
|
|
from concurrent.futures import ThreadPoolExecutor, as_completed
|
|
from .utils import is_chinese, call_ai
|
|
from .config import AI_CONCURRENCY
|
|
from .logger import get_logger
|
|
|
|
|
|
def translate_title(article, api_key, base_url, model, index=0, total=0):
|
|
"""仅翻译外文文章的标题，返回更新后的文章"""
|
|
logger = get_logger()
|
|
try:
|
|
title = article['title']
|
|
prompt = f"""请将以下军事相关的英文标题翻译成专业中文，仅返回翻译结果，不要其他内容：
|
|
{title}"""
|
|
|
|
result = call_ai([
|
|
{'role': 'system', 'content': '你是专业的军事科技翻译专家，翻译准确、专业、简洁，只返回翻译结果。'},
|
|
{'role': 'user', 'content': prompt}
|
|
], api_key, base_url, model, temperature=0.2, max_tokens=200, purpose="标题翻译")
|
|
|
|
article['translated_title'] = result.strip()
|
|
if index > 0 and total > 0:
|
|
logger.info("标题翻译完成 %d/%d: %s", index, total, article['translated_title'][:30])
|
|
return article
|
|
|
|
except Exception as e:
|
|
article['translated_title'] = article['title']
|
|
if index > 0 and total > 0:
|
|
logger.warning("标题翻译失败 %d/%d: %s", index, total, e)
|
|
return article
|
|
|
|
|
|
def translate_content(article, api_key, base_url, model, index=0, total=0):
|
|
"""翻译外文文章的正文前500词，返回更新后的文章"""
|
|
logger = get_logger()
|
|
try:
|
|
content = article['content'][:500] if article['content'] else ''
|
|
prompt = f"""请将以下军事相关的英文正文翻译成专业中文，仅返回翻译结果：
|
|
{content}"""
|
|
|
|
result = call_ai([
|
|
{'role': 'system', 'content': '你是专业的军事科技翻译专家，翻译准确、专业、简洁，只返回翻译结果。'},
|
|
{'role': 'user', 'content': prompt}
|
|
], api_key, base_url, model, temperature=0.2, max_tokens=800, purpose="正文翻译")
|
|
|
|
article['translated_content'] = result.strip()[:300]
|
|
if index > 0 and total > 0:
|
|
logger.info("正文翻译完成 %d/%d: %s", index, total, article['translated_title'][:30])
|
|
return article
|
|
|
|
except Exception as e:
|
|
article['translated_content'] = article['content'][:300] if article['content'] else ''
|
|
if index > 0 and total > 0:
|
|
logger.warning("正文翻译失败 %d/%d: %s", index, total, e)
|
|
return article
|
|
|
|
|
|
def _translate_single_article(article, api_key, base_url, model, index=0, total=0):
|
|
"""翻译单篇外文文章的标题和正文前500词，返回处理后的文章（兼容旧接口）"""
|
|
article = translate_title(article, api_key, base_url, model, index, total)
|
|
article = translate_content(article, api_key, base_url, model, index, total)
|
|
return article
|
|
|
|
|
|
def batch_translate_titles(articles, api_key, base_url, model, max_concurrent=None):
|
|
"""批量翻译所有外文文章的标题，中文文章直接填充translated_title
|
|
所有文章处理后都包含translated_title字段，无需后续判断
|
|
"""
|
|
logger = get_logger()
|
|
processed = []
|
|
chinese_articles = []
|
|
foreign_articles = []
|
|
|
|
for a in articles:
|
|
if is_chinese(a['title']):
|
|
a['translated_title'] = a['title']
|
|
chinese_articles.append(a)
|
|
else:
|
|
foreign_articles.append(a)
|
|
|
|
processed.extend(chinese_articles)
|
|
logger.info("中文文章：%d 篇，待翻译标题的外文：%d 篇", len(chinese_articles), len(foreign_articles))
|
|
|
|
if not foreign_articles:
|
|
return processed
|
|
|
|
if max_concurrent is None:
|
|
max_concurrent = AI_CONCURRENCY
|
|
|
|
logger.info("开始并发翻译标题，并发数：%d", max_concurrent)
|
|
translated_results = []
|
|
|
|
with ThreadPoolExecutor(max_workers=max_concurrent) as executor:
|
|
future_to_index = {
|
|
executor.submit(translate_title, article, api_key, base_url, model, i+1, len(foreign_articles)): i
|
|
for i, article in enumerate(foreign_articles)
|
|
}
|
|
|
|
for future in as_completed(future_to_index):
|
|
translated_article = future.result()
|
|
translated_results.append(translated_article)
|
|
|
|
processed.extend(translated_results)
|
|
logger.info("标题翻译完成，共处理 %d 篇文章", len(processed))
|
|
return processed
|
|
|
|
|
|
def batch_translate_contents(articles, api_key, base_url, model, max_concurrent=None):
|
|
"""批量翻译指定文章的正文，仅翻译需要的文章
|
|
调用前需要确保文章已经翻译了标题，并且经过筛选确实需要翻译正文
|
|
"""
|
|
logger = get_logger()
|
|
if not articles:
|
|
return []
|
|
|
|
# 分离不需要翻译正文的中文和需要翻译的外文
|
|
chinese_articles = []
|
|
foreign_articles = []
|
|
for a in articles:
|
|
if is_chinese(a['title']):
|
|
a['translated_content'] = a['content'][:300] if a['content'] else ''
|
|
chinese_articles.append(a)
|
|
else:
|
|
foreign_articles.append(a)
|
|
|
|
logger.info("无需翻译正文的中文文章：%d 篇，待翻译正文的外文：%d 篇", len(chinese_articles), len(foreign_articles))
|
|
|
|
if not foreign_articles:
|
|
return chinese_articles
|
|
|
|
if max_concurrent is None:
|
|
max_concurrent = AI_CONCURRENCY
|
|
|
|
logger.info("开始并发翻译正文，并发数：%d", max_concurrent)
|
|
translated_results = []
|
|
|
|
with ThreadPoolExecutor(max_workers=max_concurrent) as executor:
|
|
future_to_index = {
|
|
executor.submit(translate_content, article, api_key, base_url, model, i+1, len(foreign_articles)): i
|
|
for i, article in enumerate(foreign_articles)
|
|
}
|
|
|
|
for future in as_completed(future_to_index):
|
|
translated_article = future.result()
|
|
translated_results.append(translated_article)
|
|
|
|
all_articles = chinese_articles + translated_results
|
|
logger.info("正文翻译完成，共处理 %d 篇文章", len(all_articles))
|
|
return all_articles
|
|
|
|
|
|
def batch_translate_articles(articles, api_key, base_url, model, max_concurrent=None):
|
|
"""批量翻译外文文章：中英分离，外文并发翻译（兼容旧接口）
|
|
默认并发数使用配置中的AI_CONCURRENCY
|
|
"""
|
|
logger = get_logger()
|
|
chinese_articles = []
|
|
foreign_articles = []
|
|
|
|
for a in articles:
|
|
if is_chinese(a['title']) and is_chinese(a['content'][:100]):
|
|
a['translated_title'] = a['title']
|
|
a['translated_content'] = a['content'][:300] if a['content'] else ''
|
|
chinese_articles.append(a)
|
|
else:
|
|
foreign_articles.append(a)
|
|
|
|
logger.info("中文文章：%d 篇，待翻译外文：%d 篇", len(chinese_articles), len(foreign_articles))
|
|
|
|
if not foreign_articles:
|
|
return articles
|
|
|
|
if max_concurrent is None:
|
|
max_concurrent = AI_CONCURRENCY
|
|
|
|
logger.info("开始并发翻译，并发数：%d", max_concurrent)
|
|
translated_results = []
|
|
|
|
with ThreadPoolExecutor(max_workers=max_concurrent) as executor:
|
|
future_to_index = {
|
|
executor.submit(_translate_single_article, article, api_key, base_url, model, i+1, len(foreign_articles)): i
|
|
for i, article in enumerate(foreign_articles)
|
|
}
|
|
|
|
for future in as_completed(future_to_index):
|
|
translated_article = future.result()
|
|
translated_results.append(translated_article)
|
|
|
|
logger.info("所有翻译任务完成，共翻译 %d 篇外文", len(translated_results))
|
|
return chinese_articles + translated_results
|