diff --git a/flask_drop_rewrite_request.py b/flask_drop_rewrite_request.py index 5adb1de..796ce93 100644 --- a/flask_drop_rewrite_request.py +++ b/flask_drop_rewrite_request.py @@ -1,5 +1,4 @@ import os -os.environ["CUDA_VISIBLE_DEVICES"] = "2" from flask import Flask, jsonify from flask import request import requests @@ -10,7 +9,7 @@ from threading import Thread import time import re import logging -from vllm import LLM, SamplingParams +import concurrent.futures logging.basicConfig(level=logging.DEBUG, # 控制台打印的日志级别 @@ -42,7 +41,7 @@ pantten_biaoti_0 = '^[1-9一二三四五六七八九ⅠⅡⅢⅣⅤⅥⅦⅧⅨ] pantten_biaoti_1 = '^第[一二三四五六七八九]章\s{0,}?[\u4e00-\u9fa5a-zA-Z]+' pantten_biaoti_2 = '^[0-9.]+\s{0,}?[\u4e00-\u9fa5a-zA-Z]+' pantten_biaoti_3 = '^[(（][1-9一二三四五六七八九ⅠⅡⅢⅣⅤⅥⅦⅧⅨ][)_）][、.]{0,}?\s{0,}?[\u4e00-\u9fa5a-zA-Z]+' -chatgpt_url = "http://114.116.25.228:12000/predict" +chatgpt_url = "http://192.168.31.145:12001/predict" class log: @@ -141,6 +140,20 @@ def chulichangju_1(text, snetence_id, chulipangban_return_list, short_num): return chulipangban_return_list +def get_multiple_urls(urls): + with concurrent.futures.ThreadPoolExecutor() as executor: + future_to_url = {executor.submit(dialog_line_parse, url[1], url[2]): url for url in urls} + results = [] + for future in concurrent.futures.as_completed(future_to_url): + url = future_to_url[future] + try: + data = future.result() + results.append((url, data)) + except Exception as e: + results.append((url, f"Error: {str(e)}")) + return results + + def chulipangban_test_1(snetence_id, text): # 引号处理 @@ -237,6 +250,13 @@ def batch_predict(batch_data_list): return batch_data_list_new +def is_chinese(char): + if '\u4e00' <= char <= '\u9fff': + return True + else: + return False + + def predict_data_post_processing(text_list): text_list_sentence = [] # text_list_sentence.append([text_list[0][0], text_list[0][1]]) @@ -290,6 +310,8 @@ def post_sentence_ulit(sentence, text_info): sentence = sentence[sentence_lable_index + 4:] if sentence[-1] == "\n": sentence = sentence[:-1] + + else: sentence = text_info[0] return sentence @@ -304,9 +326,14 @@ def pre_sentence_ulit(sentence): sentence = str(sentence).strip() if_change = True if len(sentence) > 9: - text = "You are a helpful assistant.\n\nUser:改写下面这句话，要求意思接近但是改动幅度比较大，字数只能多不能少:\n{}\nAssistant:".format(sentence) + + if sentence[-1] != "。": + text = f"User:改写下面这段文字，要求意思接近但是改动幅度比较大，字数只能多不能少，短句前后词跟上下句衔接不能有错误，并且如果结尾有标点符号，标点不能改变，如果结尾没有标点符号，不能擅自添加标点符号:\n{sentence}\nAssistant:" + else: + text = f"User:改写下面这句话，要求意思接近但是改动幅度比较大，字数只能多不能少:\n{sentence}\nAssistant:" + else: - text = "You are a helpful assistant.\n\nUser:下面词不做任何变化:\n{}\nAssistant:".format(sentence) + text = f"下面词不做任何变化:\n{sentence}" if_change = False return text, if_change @@ -361,20 +388,42 @@ def main(texts: dict): ## ==================================================================== + # input_data = [] + # for i in range(len(text_sentence)): + # input_data.append([chatgpt_url, {"texts": text_sentence[i]}]) + # + # with concurrent.futures.ThreadPoolExecutor() as executor: + # # 使用submit方法将任务提交给线程池，并获取Future对象 + # futures = [executor.submit(dialog_line_parse, i[0], i[1]) for i in input_data] + # + # # 使用as_completed获取已完成的任务，并获取返回值 + # results = [future.result() for future in concurrent.futures.as_completed(futures)] + # + # generated_text_list = [] + # for dan in results: + # generated_text_list.append(dan["data"]) + + ## ============================================================================================== + + # urls = [ + # [0, 'http://114.116.25.228:12000/predict', {'texts': '任务：改写句子\n改写下面这句话，要求意思接近但是改动幅度比较大，字数只能多不能少:\n中华苏维埃共和国时期的法律制度建设是中国共产党'}], + # [1, 'http://114.116.25.228:12000/predict', {'texts': '任务：改写句子\n改写下面这句话，要求意思接近但是改动幅度比较大，字数只能多不能少:\n革命道路的实践中,无数革命先辈用鲜血和生命铸就了以坚定信念、求真务实、一心为民、清正廉洁、艰苦奋斗、争创一流、无私奉献等为主要内涵的苏区精神,孕育了'}], + # [2, 'http://114.116.25.228:12000/predict', {'texts': '任务：改写句子\n改写下面这句话，要求意思接近但是改动幅度比较大，字数只能多不能少:\n年前的浙江枫桥干部群众创造了“依靠群众就地化解矛盾”的“枫桥经验”,并根据形势变化不断赋予新的内涵,已成为基层治理的'}], + # [3, 'http://114.116.25.228:12000/predict', {'texts': '任务：改写句子\n改写下面这句话，要求意思接近但是改动幅度比较大，字数只能多不能少:\n新时代,习近平总书记反复强调要坚持和发展新时代“枫桥经验”,加强和创新基层社会治理,切实把矛盾化解在基层,维护好社会稳定。'}], + # [4, 'http://114.116.25.228:12000/predict', {'texts': '任务：改写句子\n改写下面这句话，要求意思接近但是改动幅度比较大，字数只能多不能少:\n一是新时代“枫桥经验”对'}] + # ] + input_data = [] for i in range(len(text_sentence)): - input_data.append([url, {"texts": text_sentence[i]}]) + input_data.append([i, chatgpt_url, {"texts": text_sentence[i]}]) - with concurrent.futures.ThreadPoolExecutor() as executor: - # 使用submit方法将任务提交给线程池，并获取Future对象 - futures = [executor.submit(dialog_line_parse, i[0], i[1]) for i in input_data] + results = get_multiple_urls(input_data) - # 使用as_completed获取已完成的任务，并获取返回值 - results = [future.result() for future in concurrent.futures.as_completed(futures)] + generated_text_list = [""] * len(input_data) + for url, result in results: + # print(f"Result for {url}: {result}") + generated_text_list[url[0]] = result["data"] - generated_text_list = [] - for dan in results: - generated_text_list.append(dan["data"]) for i in range(len(generated_text_list)): # if len(text_list[i][0]) > 7: @@ -511,4 +560,4 @@ if __name__ == "__main__": '%(asctime)s - %(pathname)s[line:%(lineno)d] - %(levelname)s: %(message)s' # 日志格式 ) - app.run(host="0.0.0.0", port=14002, threaded=True, debug=False) + app.run(host="0.0.0.0", port=14008, threaded=True, debug=False) diff --git a/redis_check_uuid_mistral.py b/redis_check_uuid_mistral.py index 01c9015..0c7f833 100644 --- a/redis_check_uuid_mistral.py +++ b/redis_check_uuid_mistral.py @@ -89,4 +89,4 @@ def handle_query(): if __name__ == "__main__": - app.run(debug=False, host='0.0.0.0', port=14003) + app.run(debug=False, host='0.0.0.0', port=14009)