使用asyncio来构建异步翻译任务提高翻译效率

目前存在退出的时候报错async runtime error还未解决
自动检测并安装openpyxl和openai库
This commit is contained in:
Ekko.bao 2024-11-04 08:42:00 +08:00
parent 1f2de5d8ac
commit d3ab696d10
2 changed files with 145 additions and 90 deletions

View File

@ -1,14 +1,25 @@
from dataclasses import dataclass
import shutil
import openpyxl
from typing import List, Tuple
import os
import sys
from openai import OpenAI
import json
try:
import openpyxl
except ImportError:
print("openpyxl is not installed, installing...")
os.system("pip install openpyxl")
import openpyxl
try:
from openai import AsyncOpenAI
except ImportError:
print("openai is not installed, installing...")
os.system("pip install openai")
from openai import AsyncOpenAI
import asyncio
# Set up OpenAI API key
API_KEY = "sk-ckFgxmnjJAJoVfcVF918CbFbEc5a459eA72cA51e4dB24dAf"
API_KEY = "sk-ckFgxmnjJAJoVfcVF918CbFbEc5a459eA72cA51e4dB24dAf" #来自V3API
API_URL = "https://api.gpt.ge/v1"
client = OpenAI(api_key=API_KEY, base_url=API_URL)
# completion = client.chat.completions.create(
# model="gpt-4o-mini",
# messages=[
@ -37,15 +48,13 @@ class Model:
# 翻译时请结合所有内容整体去进行理解含义而不仅仅是单个单元格的内容。
# 注意仅输出翻译后的内容即可,不要保留原文任何内容!!!
# 请翻译:\n"""
PROMT = """
# 将我所提供的如下内容翻译为英文,
# 注意仅输出翻译后的内容即可,不要保留原文任何内容!!!
# 请翻译:\n"""
def chinese2english(text, model=Model.gpt_4o_mini):
PROMT = "将如下内容翻译为英文,仅输出翻译后的内容,不输出任何原文:"
async def chinese2english(text, model=Model.gpt_4o_mini):
# print("start translate")
#Translate the text using OpenAI
response = client.chat.completions.create(
model = Model.gpt_4o_mini,
client = AsyncOpenAI(api_key=API_KEY, base_url=API_URL)
response = await client.chat.completions.create(
model = model,
messages = [
{
"role": "user",
@ -57,49 +66,94 @@ def chinese2english(text, model=Model.gpt_4o_mini):
)
# print("translate done")
translated_text = response.choices[0].message.content
client.close()
return translated_text
def chinese2english_stream(text, model=Model.gpt_4o_mini):
with client.chat.completions.with_streaming_response.create(
model = Model.gpt_4o_mini,
messages = [
{
"role": "user",
"content": [
{"type": "text", "text": PROMT + text},
],
}
]
) as response:
result = response.json()
print(result)
# def chinese2english_stream(text, model=Model.gpt_4o_mini):
# client = OpenAI(api_key=API_KEY, base_url=API_URL)
# with client.chat.completions.with_streaming_response.create(
# model = Model.gpt_4o_mini,
# messages = [
# {
# "role": "user",
# "content": [
# {"type": "text", "text": PROMT + text},
# ],
# }
# ]
# ) as response:
# result = response.json()
# print(result)
# return ''
def idx2excel_pos(row_idx, col_idx):
if col_idx < 1:
return ''
ret = ''
while col_idx:
ret += chr(ord('A') + (col_idx - 1) % 26)
col_idx = (col_idx - 1) // 26
return f"{ret[::-1]}{row_idx}"
def update_translate(output_sheet, map_list, translated_text):
# json_obj = json.loads(translated_text)
# print(json_obj)
json_obj = [translated_text]
for idx, (row_idx, col_idx) in enumerate(map_list):
output_sheet.cell(row=row_idx, column=col_idx).value = json_obj[idx]
pos = [f"{idx2excel_pos(row_idx, col_idx)}" for row_idx, col_idx in map_list]
pos = ','.join(pos)
print(f"========================= [{pos}] Update Translated Text :\n{translated_text}")
for row_idx, col_idx in map_list:
output_sheet.cell(row=row_idx, column=col_idx).value = translated_text
def get_json_content(content):
content = content.strip().split('\n')
#去除掉api返回的code代码提示信息
if content[0].strip().startswith('```') and content[-1].strip().endswith('```'):
content = content[1:-1]
return '\n'.join(content)
# def get_json_content(content):
# content = content.strip().split('\n')
# #去除掉api返回的code代码提示信息
# if content[0].strip().startswith('```') and content[-1].strip().endswith('```'):
# content = content[1:-1]
# return '\n'.join(content)
def is_all_ascii(s):
return all(ord(char) < 128 for char in s)
def main():
if len(sys.argv) >= 3:
print("Usage: python excel_translate.py input_file_path output_file_path")
input_file_path = sys.argv[1]
output_file_path = sys.argv[2]
translate_excel_process()
def translate_excel_process(input_file_path="test.xlsx", output_file_path="output.xlsx"):
@dataclass
class TaskInfo:
original_text: str
translated_text: str
pos:List[Tuple[int, int]]
class TaskManager:
def __init__(self, output_sheet):
self.task_list = []
self.output_sheet = output_sheet
async def add_task(self, pos_list, original_text):
task = TaskInfo(original_text, '', pos_list)
self.task_list.append(task)
if len(self.task_list) >= 20:
await self.do_task()
async def do_task(self):
tasks = []
for task in self.task_list:
tasks.append(asyncio.create_task(chinese2english(task.original_text)))
results = await asyncio.gather(*tasks)
for idx, task in enumerate(self.task_list):
task.translated_text = results[idx]
update_translate(self.output_sheet, task.pos, task.translated_text)
self.task_list = []
async def finish(self):
if len(self.task_list) > 0:
await self.do_task()
async def main() -> None:
if len(sys.argv) < 2:
print("Usage: python excel_translate.py input_file_path")
sys.exit(0)
input_file = sys.argv[1]
output = os.path.splitext(input_file)[0] + '_translated' + os.path.splitext(input_file)[1]
await translate_excel_process(input_file, output)
print(f"{input_file} Translation complete => {output}.")
async def translate_excel_process(input_file_path="input.xlsx", output_file_path="output.xlsx"):
# Check if the input file exists
if not os.path.exists(input_file_path):
print("Input file not found.")
@ -108,54 +162,55 @@ def translate_excel_process(input_file_path="test.xlsx", output_file_path="outpu
# Open the input file
input_workbook = openpyxl.load_workbook(input_file_path)
input_sheet = input_workbook.active
# Create a new output workbook
output_workbook = openpyxl.load_workbook(output_file_path)
output_sheet = output_workbook.active
# Loop through each row in the input sheet
sheets = input_workbook.sheetnames
map_list = [] #合并的单元格的翻译位置记录
original_text = ""
col_idx= 1
map_list = []
last_cell_value = ""
last_cell_value_trans = ""
# debug_test_count = 0
try:
for sheet_name in sheets:
print(f"Processing sheet: {sheet_name}")
input_sheet = input_workbook[sheet_name]
# Create a new output sheet
output_sheet = output_workbook[sheet_name]
task_manager = TaskManager(output_sheet)
# Loop through each row in the input sheet
for col_idx, col in enumerate(input_sheet.iter_cols(min_col=1, values_only=True), start=1):
# Get the original text and language code
for row_idx, content in enumerate(col, start=1):
if not content:
if not content or str(content).strip() == '': #没内容不用翻译
# print(f"Skip empty cell: {idx2excel_pos(row_idx, col_idx)}")
continue
elif isinstance(content, int):#如果是纯数字也不用翻译
continue
#如果全是英文,则不翻译
elif is_all_ascii(content):
continue
try:
int(content)
except ValueError:
pass
else:
continue
#和最近一次的cell内容相同则不翻译直接使用结果即可
if content != last_cell_value:
original_text += f'{content}\n'
if last_cell_value == '': #第一个单元格先记录一下,等待后面有不一样的才翻译
last_cell_value = content
original_text = content
map_list.append((row_idx, col_idx))
#if len(original_text) > 100:
continue
last_cell_value = content
# print(f"Original text: {original_text}")
translated_text = chinese2english(original_text)
# translated_text = get_json_content(translated_text)
# print(f"Translated text: {translated_text}")
update_translate(output_sheet, map_list, translated_text)
await task_manager.add_task(map_list, original_text)
original_text = content
map_list = []
original_text = ""
# Add the original text and translated text to the output sheet
# output_sheet.append([original_text, translated_text])
map_list.append((row_idx, col_idx))
await task_manager.finish()
except Exception as e:
print(f"Error: {e}")
pass
except KeyboardInterrupt:
print("KeyboardInterrupt")
output_workbook.save(output_file_path)
sys.exit()
# Save the output workbook
output_workbook.save(output_file_path)
print("Translation complete.")
if __name__ == '__main__':
main()
pass
if __name__ == "__main__":
asyncio.run(main())
exit(0)

Binary file not shown.