529 lines
17 KiB
Python
Executable File
529 lines
17 KiB
Python
Executable File
#!/usr/bin/env python3
|
||
# -*- coding: utf-8 -*-
|
||
|
||
from __future__ import annotations
|
||
|
||
import argparse
|
||
import json
|
||
import math
|
||
import re
|
||
import sys
|
||
from pathlib import Path
|
||
from typing import Any
|
||
|
||
try:
|
||
import google.generativeai as genai # type: ignore
|
||
except ImportError as exc:
|
||
raise RuntimeError(
|
||
"缺少依赖 google-generativeai，请先安装：pip install google-generativeai"
|
||
) from exc
|
||
|
||
try:
|
||
from openpyxl import load_workbook, Workbook # type: ignore
|
||
except ImportError as exc:
|
||
raise RuntimeError(
|
||
"缺少依赖 openpyxl，请先安装：pip install openpyxl"
|
||
) from exc
|
||
|
||
try:
|
||
import pandas as pd # type: ignore
|
||
except ImportError as exc:
|
||
raise RuntimeError(
|
||
"缺少依赖 pandas，请先安装：pip install pandas"
|
||
) from exc
|
||
|
||
|
||
def detect_chinese(text: str) -> bool:
|
||
"""检测文本中是否包含中文字符"""
|
||
if not text or not isinstance(text, str):
|
||
return False
|
||
return bool(re.search(r"[\u4e00-\u9fff]", text))
|
||
|
||
|
||
def format_cell_value(value: Any) -> Any:
|
||
"""格式化单元格值，保持原始类型"""
|
||
if value is None:
|
||
return None
|
||
if isinstance(value, float) and math.isnan(value):
|
||
return None
|
||
return value
|
||
|
||
|
||
def get_api_key() -> str:
|
||
"""获取 Gemini API 密钥"""
|
||
api_key = genai.configure(api_key=None)
|
||
if api_key:
|
||
return api_key
|
||
|
||
# 尝试从环境变量获取
|
||
import os
|
||
api_key = os.getenv("GEMINI_API_KEY") or os.getenv("GOOGLE_API_KEY")
|
||
if api_key:
|
||
return api_key
|
||
|
||
raise ValueError(
|
||
"未找到 Gemini API 密钥。请设置环境变量 GEMINI_API_KEY 或 GOOGLE_API_KEY，"
|
||
"或使用 --api-key 参数提供。"
|
||
)
|
||
|
||
|
||
def translate_batch(
|
||
texts: list[str],
|
||
model_name: str = "gemini-2.0-flash-lite",
|
||
api_key: str | None = None,
|
||
) -> dict[str, str]:
|
||
"""
|
||
批量翻译文本
|
||
|
||
Args:
|
||
texts: 待翻译的文本列表
|
||
model_name: 使用的模型名称
|
||
api_key: API 密钥（可选）
|
||
|
||
Returns:
|
||
原文到译文的映射字典
|
||
"""
|
||
if not texts:
|
||
return {}
|
||
|
||
# 过滤掉空文本
|
||
non_empty_texts = [(i, t) for i, t in enumerate(texts) if t and t.strip()]
|
||
if not non_empty_texts:
|
||
return {}
|
||
|
||
# 配置 API
|
||
if api_key:
|
||
genai.configure(api_key=api_key)
|
||
else:
|
||
get_api_key() # 触发自动获取
|
||
|
||
# 选择模型
|
||
try:
|
||
model = genai.GenerativeModel(model_name)
|
||
except Exception as exc:
|
||
raise RuntimeError(f"无法加载模型 {model_name}: {exc}") from exc
|
||
|
||
# 构建批量翻译提示
|
||
# 将所有待翻译文本合并为一个请求以提高效率
|
||
text_list = "\n".join(f"{i+1}. {text}" for i, (_, text) in enumerate(non_empty_texts))
|
||
|
||
prompt = f"""请将以下中文内容翻译成英文。保持专业、准确的语言风格。
|
||
注意：
|
||
1. 只翻译中文部分，保持原有的专有名词（如人名、地名、品牌名）不变
|
||
2. 保留数字、日期、时间等格式不变
|
||
3. 技术术语使用标准英文翻译
|
||
4. 每条翻译结果单独一行，格式为：序号. 译文
|
||
|
||
待翻译内容：
|
||
{text_list}
|
||
|
||
请按顺序输出翻译结果："""
|
||
|
||
try:
|
||
response = model.generate_content(prompt)
|
||
result_text = response.text
|
||
except Exception as exc:
|
||
raise RuntimeError(f"翻译请求失败: {exc}") from exc
|
||
|
||
# 解析翻译结果
|
||
result_map: dict[str, str] = {}
|
||
lines = result_text.strip().split("\n")
|
||
|
||
for line in lines:
|
||
line = line.strip()
|
||
# 匹配 "序号. 译文" 格式
|
||
match = re.match(r"^(\d+)\.\s*(.+)$", line)
|
||
if match:
|
||
index = int(match.group(1)) - 1 # 转为 0-based 索引
|
||
if index < len(non_empty_texts):
|
||
original_index, original_text = non_empty_texts[index]
|
||
result_map[original_text] = match.group(2)
|
||
else:
|
||
# 如果没有序号，尝试直接映射
|
||
if non_empty_texts:
|
||
original_index, original_text = non_empty_texts[0]
|
||
if original_text not in result_map:
|
||
result_map[original_text] = line
|
||
|
||
return result_map
|
||
|
||
|
||
def translate_excel_file(
|
||
input_path: Path,
|
||
output_path: Path,
|
||
columns: list[str] | None = None,
|
||
sheet_name: str | None = None,
|
||
model_name: str = "gemini-2.0-flash-lite",
|
||
api_key: str | None = None,
|
||
dry_run: bool = False,
|
||
) -> dict[str, Any]:
|
||
"""
|
||
翻译 Excel 文件
|
||
|
||
Args:
|
||
input_path: 输入文件路径
|
||
output_path: 输出文件路径
|
||
columns: 指定要翻译的列名列表
|
||
sheet_name: 指定工作表名称
|
||
model_name: 使用的模型名称
|
||
api_key: API 密钥
|
||
dry_run: 预览模式，不实际生成文件
|
||
|
||
Returns:
|
||
翻译结果统计信息
|
||
"""
|
||
# 加载工作簿
|
||
wb = load_workbook(input_path)
|
||
|
||
# 选择工作表
|
||
if sheet_name:
|
||
if sheet_name not in wb.sheetnames:
|
||
raise ValueError(f"工作表不存在: {sheet_name}")
|
||
sheets_to_translate = [sheet_name]
|
||
else:
|
||
sheets_to_translate = wb.sheetnames
|
||
|
||
# 统计信息
|
||
stats: dict[str, Any] = {
|
||
"sheets": {},
|
||
"total_cells": 0,
|
||
"translated_cells": 0,
|
||
"chinese_cells": 0,
|
||
"skipped_cells": 0,
|
||
}
|
||
|
||
# 处理每个工作表
|
||
for sheet_name in sheets_to_translate:
|
||
ws = wb[sheet_name]
|
||
sheet_stats = {
|
||
"total": 0,
|
||
"chinese": 0,
|
||
"translated": 0,
|
||
"skipped": 0,
|
||
"columns": [],
|
||
}
|
||
|
||
# 收集需要翻译的列
|
||
header_row = 1 # 默认第一行为表头
|
||
headers: list[str] = []
|
||
target_columns: list[int] = []
|
||
|
||
# 读取表头
|
||
for col in range(1, ws.max_column + 1):
|
||
cell_value = ws.cell(row=header_row, column=col).value
|
||
header = str(cell_value).strip() if cell_value else f"Column{col}"
|
||
headers.append(header)
|
||
|
||
# 如果指定了列名，检查是否匹配
|
||
if columns is None or header in columns:
|
||
target_columns.append(col)
|
||
sheet_stats["columns"].append(header)
|
||
|
||
if not target_columns:
|
||
stats["sheets"][sheet_name] = sheet_stats
|
||
continue
|
||
|
||
# 收集所有需要翻译的文本
|
||
texts_to_translate: list[str] = []
|
||
cell_positions: list[tuple[int, int]] = [] # (row, col)
|
||
|
||
for row in range(header_row + 1, ws.max_row + 1):
|
||
for col in target_columns:
|
||
cell = ws.cell(row=row, column=col)
|
||
value = cell.value
|
||
|
||
# 跳过空值、公式、数字
|
||
if value is None or isinstance(value, (int, float, bool)):
|
||
sheet_stats["skipped"] += 1
|
||
continue
|
||
|
||
if isinstance(value, float) and math.isnan(value):
|
||
sheet_stats["skipped"] += 1
|
||
continue
|
||
|
||
text = str(value).strip()
|
||
if not text:
|
||
sheet_stats["skipped"] += 1
|
||
continue
|
||
|
||
sheet_stats["total"] += 1
|
||
|
||
# 检测中文
|
||
if detect_chinese(text):
|
||
sheet_stats["chinese"] += 1
|
||
texts_to_translate.append(text)
|
||
cell_positions.append((row, col))
|
||
else:
|
||
sheet_stats["skipped"] += 1
|
||
|
||
# 批量翻译
|
||
if texts_to_translate:
|
||
print(f"翻译工作表 '{sheet_name}' 中的 {len(texts_to_translate)} 个单元格...")
|
||
translation_map = translate_batch(texts_to_translate, model_name, api_key)
|
||
|
||
# 应用翻译结果
|
||
for (row, col), original_text in zip(cell_positions, texts_to_translate):
|
||
translated = translation_map.get(original_text)
|
||
if translated:
|
||
ws.cell(row=row, column=col, value=translated)
|
||
sheet_stats["translated"] += 1
|
||
else:
|
||
sheet_stats["skipped"] += 1
|
||
|
||
stats["sheets"][sheet_name] = sheet_stats
|
||
stats["total_cells"] += sheet_stats["total"]
|
||
stats["chinese_cells"] += sheet_stats["chinese"]
|
||
stats["translated_cells"] += sheet_stats["translated"]
|
||
stats["skipped_cells"] += sheet_stats["skipped"]
|
||
|
||
# 保存文件
|
||
if not dry_run:
|
||
# 确保输出目录存在
|
||
output_path.parent.mkdir(parents=True, exist_ok=True)
|
||
wb.save(output_path)
|
||
print(f"已保存翻译结果到: {output_path}")
|
||
else:
|
||
print("预览模式：未生成文件")
|
||
|
||
return stats
|
||
|
||
|
||
def translate_csv_file(
|
||
input_path: Path,
|
||
output_path: Path,
|
||
columns: list[str] | None = None,
|
||
model_name: str = "gemini-2.0-flash-lite",
|
||
api_key: str | None = None,
|
||
dry_run: bool = False,
|
||
) -> dict[str, Any]:
|
||
"""
|
||
翻译 CSV 文件
|
||
|
||
Args:
|
||
input_path: 输入文件路径
|
||
output_path: 输出文件路径
|
||
columns: 指定要翻译的列名列表
|
||
model_name: 使用的模型名称
|
||
api_key: API 密钥
|
||
dry_run: 预览模式
|
||
|
||
Returns:
|
||
翻译结果统计信息
|
||
"""
|
||
# 检测编码
|
||
last_error: Exception | None = None
|
||
df = None
|
||
encoding = "utf-8-sig"
|
||
|
||
for enc in ("utf-8-sig", "utf-8", "gb18030"):
|
||
try:
|
||
df = pd.read_csv(input_path, encoding=enc)
|
||
encoding = enc
|
||
break
|
||
except UnicodeDecodeError as exc:
|
||
last_error = exc
|
||
continue
|
||
except Exception as exc:
|
||
last_error = exc
|
||
continue
|
||
|
||
if df is None:
|
||
raise ValueError(f"无法读取 CSV 文件: {last_error}")
|
||
|
||
# 确定要翻译的列
|
||
target_columns: list[str] = []
|
||
if columns:
|
||
for col in columns:
|
||
if col in df.columns:
|
||
target_columns.append(col)
|
||
else:
|
||
print(f"警告: 列 '{col}' 不存在，已跳过")
|
||
else:
|
||
target_columns = df.columns.tolist()
|
||
|
||
if not target_columns:
|
||
raise ValueError("没有可翻译的列")
|
||
|
||
# 统计信息
|
||
stats: dict[str, Any] = {
|
||
"sheets": {"main": {"total": 0, "chinese": 0, "translated": 0, "skipped": 0, "columns": target_columns}},
|
||
"total_cells": 0,
|
||
"translated_cells": 0,
|
||
"chinese_cells": 0,
|
||
"skipped_cells": 0,
|
||
}
|
||
|
||
# 收集需要翻译的文本
|
||
texts_to_translate: list[str] = []
|
||
cell_positions: list[tuple[int, str]] = [] # (row, col)
|
||
|
||
for col in target_columns:
|
||
for idx, value in enumerate(df[col], start=1):
|
||
# 跳过空值和 NaN
|
||
if pd.isna(value) or value == "":
|
||
stats["skipped_cells"] += 1
|
||
stats["sheets"]["main"]["skipped"] += 1
|
||
continue
|
||
|
||
# 跳过数字
|
||
if isinstance(value, (int, float)) and not isinstance(value, bool):
|
||
stats["skipped_cells"] += 1
|
||
stats["sheets"]["main"]["skipped"] += 1
|
||
continue
|
||
|
||
text = str(value).strip()
|
||
if not text:
|
||
stats["skipped_cells"] += 1
|
||
stats["sheets"]["main"]["skipped"] += 1
|
||
continue
|
||
|
||
stats["total_cells"] += 1
|
||
stats["sheets"]["main"]["total"] += 1
|
||
|
||
# 检测中文
|
||
if detect_chinese(text):
|
||
stats["chinese_cells"] += 1
|
||
stats["sheets"]["main"]["chinese"] += 1
|
||
texts_to_translate.append(text)
|
||
cell_positions.append((idx, col))
|
||
else:
|
||
stats["skipped_cells"] += 1
|
||
stats["sheets"]["main"]["skipped"] += 1
|
||
|
||
# 批量翻译
|
||
if texts_to_translate:
|
||
print(f"翻译 {len(texts_to_translate)} 个单元格...")
|
||
translation_map = translate_batch(texts_to_translate, model_name, api_key)
|
||
|
||
# 应用翻译结果
|
||
for (row_idx, col), original_text in zip(cell_positions, texts_to_translate):
|
||
translated = translation_map.get(original_text)
|
||
if translated:
|
||
df.at[row_idx - 1, col] = translated # pandas 使用 0-based 索引
|
||
stats["translated_cells"] += 1
|
||
stats["sheets"]["main"]["translated"] += 1
|
||
else:
|
||
stats["skipped_cells"] += 1
|
||
stats["sheets"]["main"]["skipped"] += 1
|
||
|
||
# 保存文件
|
||
if not dry_run:
|
||
output_path.parent.mkdir(parents=True, exist_ok=True)
|
||
df.to_csv(output_path, index=False, encoding="utf-8-sig")
|
||
print(f"已保存翻译结果到: {output_path}")
|
||
else:
|
||
print("预览模式：未生成文件")
|
||
|
||
return stats
|
||
|
||
|
||
def print_stats(stats: dict[str, Any], input_path: Path) -> None:
|
||
"""打印统计信息"""
|
||
print(f"\n翻译统计 - {input_path.name}")
|
||
print("=" * 60)
|
||
print(f"总单元格数: {stats['total_cells']}")
|
||
print(f"包含中文: {stats['chinese_cells']}")
|
||
print(f"已翻译: {stats['translated_cells']}")
|
||
print(f"跳过: {stats['skipped_cells']}")
|
||
|
||
for sheet_name, sheet_stats in stats["sheets"].items():
|
||
print(f"\n工作表: {sheet_name}")
|
||
print(f" 翻译列: {', '.join(sheet_stats['columns']) if sheet_stats['columns'] else '全部'}")
|
||
print(f" 总数: {sheet_stats['total']}, 中文: {sheet_stats['chinese']}, 已翻译: {sheet_stats['translated']}")
|
||
|
||
|
||
def parse_args() -> argparse.Namespace:
|
||
parser = argparse.ArgumentParser(
|
||
description="翻译 Excel (.xlsx) 或 CSV 文件中的中文内容为英文"
|
||
)
|
||
parser.add_argument("--file", required=True, help="输入文件路径")
|
||
parser.add_argument("--output", help="输出文件路径（默认：{原文件名}_en.{扩展名}）")
|
||
parser.add_argument(
|
||
"--columns",
|
||
help="指定要翻译的列名，多个列用逗号分隔，例如：'姓名,地址,备注'"
|
||
)
|
||
parser.add_argument("--sheet", help="指定工作表名称（仅 Excel 文件）")
|
||
parser.add_argument(
|
||
"--model",
|
||
default="gemini-2.0-flash-lite",
|
||
help="使用的 Gemini 模型（默认：gemini-2.0-flash-lite）"
|
||
)
|
||
parser.add_argument("--api-key", help="Gemini API 密钥（也可通过环境变量 GEMINI_API_KEY 设置）")
|
||
parser.add_argument(
|
||
"--dry-run",
|
||
action="store_true",
|
||
help="预览模式：统计需要翻译的内容但不生成文件"
|
||
)
|
||
return parser.parse_args()
|
||
|
||
|
||
def main() -> int:
|
||
args = parse_args()
|
||
input_path = Path(args.file).expanduser()
|
||
|
||
try:
|
||
# 验证输入文件
|
||
if not input_path.exists():
|
||
raise FileNotFoundError(f"文件不存在: {input_path}")
|
||
if not input_path.is_file():
|
||
raise ValueError(f"路径不是文件: {input_path}")
|
||
|
||
# 确定输出路径
|
||
if args.output:
|
||
output_path = Path(args.output).expanduser()
|
||
else:
|
||
output_path = input_path.parent / f"{input_path.stem}_en{input_path.suffix}"
|
||
|
||
# 解析列名
|
||
columns: list[str] | None = None
|
||
if args.columns:
|
||
columns = [col.strip() for col in args.columns.split(",") if col.strip()]
|
||
|
||
print(f"输入文件: {input_path}")
|
||
print(f"输出文件: {output_path}")
|
||
if columns:
|
||
print(f"翻译列: {', '.join(columns)}")
|
||
if args.sheet:
|
||
print(f"工作表: {args.sheet}")
|
||
|
||
# 根据文件类型处理
|
||
suffix = input_path.suffix.lower()
|
||
if suffix == ".xlsx":
|
||
stats = translate_excel_file(
|
||
input_path=input_path,
|
||
output_path=output_path,
|
||
columns=columns,
|
||
sheet_name=args.sheet,
|
||
model_name=args.model,
|
||
api_key=args.api_key,
|
||
dry_run=args.dry_run,
|
||
)
|
||
elif suffix == ".csv":
|
||
if args.sheet:
|
||
raise ValueError("CSV 文件不支持 --sheet 参数")
|
||
stats = translate_csv_file(
|
||
input_path=input_path,
|
||
output_path=output_path,
|
||
columns=columns,
|
||
model_name=args.model,
|
||
api_key=args.api_key,
|
||
dry_run=args.dry_run,
|
||
)
|
||
else:
|
||
raise ValueError(f"不支持的文件类型: {suffix}，仅支持 .xlsx 和 .csv")
|
||
|
||
# 打印统计信息
|
||
print_stats(stats, input_path)
|
||
|
||
return 0
|
||
except KeyboardInterrupt:
|
||
print("\n已取消。", file=sys.stderr)
|
||
return 130
|
||
except Exception as exc:
|
||
print(f"错误: {exc}", file=sys.stderr)
|
||
return 1
|
||
|
||
|
||
if __name__ == "__main__":
|
||
sys.exit(main())
|