jiahui
/
generate_articles_directory
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Browse Source

增加主要内容问法

master
majiahui@haimaqingfan.com 2 months ago
parent
011da31bf4
commit
4dbd5bf0e2
11 changed files with 258 additions and 71 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Unified View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 47
      .idea/deployment.xml
  2. 2
      .idea/generate_articles_directory.iml
  3. 2
      .idea/misc.xml
  4. 26
      articles_directory_predict.py
  5. 4
      articles_directory_predict_detection_word.py
  6. 38
      articles_directory_predict_qikan_word.py
  7. 24
      articles_directory_predict_qikan_word_online.py
  8. 38
      articles_directory_predict_qikan_word_table_formula_kongzhi.py
  9. 56
      articles_directory_predict_word.py
  10. 31
      articles_directory_predict_word_online.py
  11. 61
      articles_directory_predict_word_table_formula_kongzhi.py

47
.idea/deployment.xml
View File

@ -1,6 +1,6 @@
<?xml version="1.0" encoding="UTF-8"?> <?xml version="1.0" encoding="UTF-8"?>
<project version="4"> <project version="4">
<component name="PublishConfigData" autoUpload="Always" serverName="majiahui@192.168.31.74:22 (83)" remoteFilesAllowedToDisappearOnAutoupload="false" confirmBeforeUploading="false"> <component name="PublishConfigData" remoteFilesAllowedToDisappearOnAutoupload="false" confirmBeforeUploading="false">
<option name="confirmBeforeUploading" value="false" /> <option name="confirmBeforeUploading" value="false" />
<serverData> <serverData>
<paths name="majiahui@192.168.31.149:22"> <paths name="majiahui@192.168.31.149:22">
@ -10,6 +10,20 @@
</mappings> </mappings>
</serverdata> </serverdata>
</paths> </paths>
<paths name="majiahui@192.168.31.149:22 (2)">
<serverdata>
<mappings>
<mapping local="$PROJECT_DIR$" web="/" />
</mappings>
</serverdata>
</paths>
<paths name="majiahui@192.168.31.149:22 (3)">
<serverdata>
<mappings>
<mapping deploy="/home/majiahui/project/generate_articles_directory" local="$PROJECT_DIR$" />
</mappings>
</serverdata>
</paths>
<paths name="majiahui@192.168.31.74:22"> <paths name="majiahui@192.168.31.74:22">
<serverdata> <serverdata>
<mappings> <mappings>
@ -384,7 +398,7 @@
<paths name="majiahui@192.168.31.74:22 (58)"> <paths name="majiahui@192.168.31.74:22 (58)">
<serverdata> <serverdata>
<mappings> <mappings>
<mapping deploy="/home/majiahui/project/generate_articles_directory" local="$PROJECT_DIR$" /> <mapping local="$PROJECT_DIR$" web="/" />
</mappings> </mappings>
</serverdata> </serverdata>
</paths> </paths>
@ -580,6 +594,34 @@
<paths name="majiahui@192.168.31.74:22 (83)"> <paths name="majiahui@192.168.31.74:22 (83)">
<serverdata> <serverdata>
<mappings> <mappings>
<mapping local="$PROJECT_DIR$" web="/" />
</mappings>
</serverdata>
</paths>
<paths name="majiahui@192.168.31.74:22 (84)">
<serverdata>
<mappings>
<mapping local="$PROJECT_DIR$" web="/" />
</mappings>
</serverdata>
</paths>
<paths name="majiahui@192.168.31.74:22 (85)">
<serverdata>
<mappings>
<mapping local="$PROJECT_DIR$" web="/" />
</mappings>
</serverdata>
</paths>
<paths name="majiahui@192.168.31.74:22 (86)">
<serverdata>
<mappings>
<mapping local="$PROJECT_DIR$" web="/" />
</mappings>
</serverdata>
</paths>
<paths name="majiahui@192.168.31.74:22 (87)">
<serverdata>
<mappings>
<mapping deploy="/home/majiahui/project/generate_articles_directory" local="$PROJECT_DIR$" /> <mapping deploy="/home/majiahui/project/generate_articles_directory" local="$PROJECT_DIR$" />
</mappings> </mappings>
</serverdata> </serverdata>
@ -592,6 +634,5 @@
</serverdata> </serverdata>
</paths> </paths>
</serverData> </serverData>
<option name="myAutoUpload" value="ALWAYS" />
</component> </component>
</project> </project>

2
.idea/generate_articles_directory.iml
View File

@ -5,7 +5,7 @@
</component> </component>
<component name="NewModuleRootManager"> <component name="NewModuleRootManager">
<content url="file://$MODULE_DIR$" /> <content url="file://$MODULE_DIR$" />
<orderEntry type="jdk" jdkName="Remote Python 3.8.18 (sftp://majiahui@192.168.31.74:22/home/majiahui/miniconda3/envs/llama_paper/bin/python)" jdkType="Python SDK" /> <orderEntry type="jdk" jdkName="Python 3.8 (ldm)" jdkType="Python SDK" />
<orderEntry type="sourceFolder" forTests="false" /> <orderEntry type="sourceFolder" forTests="false" />
</component> </component>
<component name="TemplatesService"> <component name="TemplatesService">

2
.idea/misc.xml
View File

@ -1,4 +1,4 @@
<?xml version="1.0" encoding="UTF-8"?> <?xml version="1.0" encoding="UTF-8"?>
<project version="4"> <project version="4">
<component name="ProjectRootManager" version="2" project-jdk-name="Remote Python 3.8.18 (sftp://majiahui@192.168.31.74:22/home/majiahui/miniconda3/envs/llama_paper/bin/python)" project-jdk-type="Python SDK" /> <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.8 (ldm)" project-jdk-type="Python SDK" />
</project> </project>

26
articles_directory_predict.py
View File

@ -1,3 +1,4 @@
import os
from flask import Flask, jsonify from flask import Flask, jsonify
from flask import request from flask import request
import time import time
@ -20,6 +21,25 @@ def get_host_ip():
return ip return ip
class log:
def __init__(self):
pass
def log(*args, **kwargs):
format = '%Y/%m/%d-%H:%M:%S'
format_h = '%Y-%m-%d'
value = time.localtime(int(time.time()))
dt = time.strftime(format, value)
dt_log_file = time.strftime(format_h, value)
log_file = 'log_file/access-%s' % dt_log_file + ".log"
if not os.path.exists(log_file):
with open(os.path.join(log_file), 'w', encoding='utf-8') as f:
print(dt, *args, file=f, **kwargs)
else:
with open(os.path.join(log_file), 'a+', encoding='utf-8') as f:
print(dt, *args, file=f, **kwargs)
chatgpt_url_predict = "http://{}:12000/predict".format(str(get_host_ip())) chatgpt_url_predict = "http://{}:12000/predict".format(str(get_host_ip()))
chatgpt_url_search = "http://{}:12000/search".format(str(get_host_ip())) chatgpt_url_search = "http://{}:12000/search".format(str(get_host_ip()))
# prompt = "<|role|>user<|says|>{}<|end|>\n<|role|>assistant<|says|>" # prompt = "<|role|>user<|says|>{}<|end|>\n<|role|>assistant<|says|>"
@ -166,7 +186,7 @@ def get_multiple_urls(urls):
def articles_directory(): def articles_directory():
text = request.json["texts"] # 获取用户query中的文本 例如"I love you" text = request.json["texts"] # 获取用户query中的文本 例如"I love you"
nums = request.json["nums"] nums = request.json["nums"]
text = qwen_think_prompt.format(text) text = qwen_prompt.format(text)
nums = int(nums) nums = int(nums)
input_data = [] input_data = []
@ -176,6 +196,10 @@ def articles_directory():
results = get_multiple_urls(input_data) results = get_multiple_urls(input_data)
log.log('start at',
'text:{},results:{}'.format(
text, str(results)))
return jsonify(results) # 返回结果 return jsonify(results) # 返回结果
if __name__ == "__main__": if __name__ == "__main__":

4
articles_directory_predict_detection_word.py
View File

@ -93,7 +93,9 @@ def articles_directory():
"min_word": results[0], "min_word": results[0],
"max_word": results[1], "max_word": results[1],
} }
# log.log('text:{},'.format(prompt)) log.log('start at',
'text:{},results:{}'.format(
mulu, str(results)))
return jsonify(data) # 返回结果 return jsonify(data) # 返回结果
# @app.route("/detection_word", methods=["POST"]) # @app.route("/detection_word", methods=["POST"])

38
articles_directory_predict_qikan_word.py
View File

@ -52,28 +52,37 @@ prompt = {
"mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中", "mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
} }
prompt_main_content = {
"mulu_title_Level_2_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题",
"mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
}
nums_word_dict = { nums_word_dict = {
"3000": { "3000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["0","3"], "title_2_nums": ["0","3"],
"title_small_nums": ["3", "0","3"] "title_small_nums": ["3", "0","3"]
}, },
"5000": { "5000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "4", "title_1_nums": "4",
"title_2_nums": ["0","5"], "title_2_nums": ["0","5"],
"title_small_nums": ["4", "0","5"] "title_small_nums": ["4", "0","5"]
}, },
"8000": { "8000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "5", "title_1_nums": "5",
"title_2_nums": ["0", "6"], "title_2_nums": ["0", "6"],
"title_small_nums": ["5", "0", "6"] "title_small_nums": ["5", "0", "6"]
}, },
"10000": { "10000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["0", "5"], "title_2_nums": ["0", "5"],
"title_small_nums": ["6", "0", "5"], "title_small_nums": ["6", "0", "5"],
@ -125,7 +134,7 @@ def request_api_chatgpt(prompt):
"content": prompt, "content": prompt,
"model": "gpt-4-turbo", "model": "gpt-4-turbo",
"top_p": 0.95, "top_p": 0.95,
"temperature": 0.7 "temperature": 0.6
} }
response = requests.post( response = requests.post(
chatgpt_url_predict, chatgpt_url_predict,
@ -217,19 +226,26 @@ def get_multiple_urls(urls):
@app.route("/articles_directory", methods=["POST"]) @app.route("/articles_directory", methods=["POST"])
def articles_directory(): def articles_directory():
title = request.json["title"] # 获取用户query中的文本 例如"I love you" title = request.json["title"]
nums_catalogue = request.json["nums_catalogue"] nums_catalogue = request.json["nums_catalogue"]
nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you" nums_word = request.json["nums_word"]
main_content = request.json.get("main_content")
nums_catalogue = int(nums_catalogue) nums_catalogue = int(nums_catalogue)
keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"]) if not main_content:
prompt = nums_word_dict[nums_word]["prompt"]%keyword keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"])
print(prompt) prompt = nums_word_dict[nums_word]["prompt"] % keyword
else:
keyword = tuple([title, main_content] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt_main_content"] % keyword
input_data = [] input_data = []
for i in range(nums_catalogue): for i in range(nums_catalogue):
input_data.append([i, prompt]) input_data.append([i, prompt])
results = get_multiple_urls(input_data) results = get_multiple_urls(input_data)
log.log('text:{},'.format(prompt)) log.log('start at',
'text:{},results:{}'.format(
prompt, str(results)))
# log.log('text:{},'.format(prompt))
return jsonify(results) # 返回结果 return jsonify(results) # 返回结果
if __name__ == "__main__": if __name__ == "__main__":

24
articles_directory_predict_qikan_word_online.py
View File

@ -49,35 +49,42 @@ prompt = {
"mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中", "mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
} }
prompt_main_content = {
"mulu_title_Level_2_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题",
"mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
}
nums_word_dict = { nums_word_dict = {
"3000": { "3000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["0","3"], "title_2_nums": ["0","3"],
"title_small_nums": ["3", "0","3"] "title_small_nums": ["3", "0","3"]
}, },
"5000": { "5000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "4", "title_1_nums": "4",
"title_2_nums": ["0","5"], "title_2_nums": ["0","5"],
"title_small_nums": ["4", "0","5"] "title_small_nums": ["4", "0","5"]
}, },
"8000": { "8000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "5", "title_1_nums": "5",
"title_2_nums": ["0", "6"], "title_2_nums": ["0", "6"],
"title_small_nums": ["5", "0", "6"] "title_small_nums": ["5", "0", "6"]
}, },
"10000": { "10000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["0", "5"], "title_2_nums": ["0", "5"],
"title_small_nums": ["6", "0", "5"], "title_small_nums": ["6", "0", "5"],
} }
} }
app = Flask(__name__) app = Flask(__name__)
app.config["JSON_AS_ASCII"] = False app.config["JSON_AS_ASCII"] = False
@ -122,7 +129,7 @@ def request_api_chatgpt(prompt):
"content": prompt, "content": prompt,
"model": "gpt-4-turbo", "model": "gpt-4-turbo",
"top_p": 0.95, "top_p": 0.95,
"temperature": 0.7 "temperature": 0.6
} }
response = requests.post( response = requests.post(
chatgpt_url_predict, chatgpt_url_predict,
@ -217,16 +224,23 @@ def articles_directory():
title = request.json["title"] # 获取用户query中的文本 例如"I love you" title = request.json["title"] # 获取用户query中的文本 例如"I love you"
nums_catalogue = request.json["nums_catalogue"] nums_catalogue = request.json["nums_catalogue"]
nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you" nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you"
main_content = request.json.get("main_content")
nums_catalogue = int(nums_catalogue) nums_catalogue = int(nums_catalogue)
keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"]) if not main_content:
prompt = nums_word_dict[nums_word]["prompt"]%keyword keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt"] % keyword
else:
keyword = tuple([title, main_content] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt_main_content"] % keyword
print(prompt) print(prompt)
input_data = [] input_data = []
for i in range(nums_catalogue): for i in range(nums_catalogue):
input_data.append([i, prompt]) input_data.append([i, prompt])
results = get_multiple_urls(input_data) results = get_multiple_urls(input_data)
log.log('text:{},'.format(prompt)) log.log('start at',
'text:{},results:{}'.format(
prompt, str(results)))
return jsonify(results) # 返回结果 return jsonify(results) # 返回结果
if __name__ == "__main__": if __name__ == "__main__":

38
articles_directory_predict_qikan_word_table_formula_kongzhi.py
View File

@ -53,34 +53,43 @@ prompt = {
"mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中", "mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
} }
prompt_main_content = {
"mulu_title_Level_2_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题",
"mulu_title_Level_3_1": "任务：生成目录\n请帮我生成一个期刊论文的目录，论文题目是“%s”，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
}
nums_word_dict = { nums_word_dict = {
"3000": { "3000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["0","3"], "title_2_nums": ["0","3"],
"title_small_nums": ["3", "0","3"] "title_small_nums": ["3", "0","3"]
}, },
"5000": { "5000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "4", "title_1_nums": "4",
"title_2_nums": ["0","5"], "title_2_nums": ["0","5"],
"title_small_nums": ["4", "0","5"] "title_small_nums": ["4", "0","5"]
}, },
"8000": { "8000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "5", "title_1_nums": "5",
"title_2_nums": ["0", "6"], "title_2_nums": ["0", "6"],
"title_small_nums": ["5", "0", "6"] "title_small_nums": ["5", "0", "6"]
}, },
"10000": { "10000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["0", "5"], "title_2_nums": ["0", "5"],
"title_small_nums": ["6", "0", "5"], "title_small_nums": ["6", "0", "5"],
} }
} }
prompt_biaogekongzhi = { prompt_biaogekongzhi = {
"prompt_1":"背景:我是一名博士生，我想写一篇论文。论文题目《{}》，目录是“{}”，我现在需要论文中有{}，请根据这些信息告诉我", "prompt_1":"背景:我是一名博士生，我想写一篇论文。论文题目《{}》，目录是“{}”，我现在需要论文中有{}，请根据这些信息告诉我",
"prompt_1_form":"哪些小标题生成的内容需要包含表格", "prompt_1_form":"哪些小标题生成的内容需要包含表格",
@ -155,7 +164,7 @@ def request_api_chatgpt(prompt):
"content": prompt, "content": prompt,
"model": "gpt-4-turbo", "model": "gpt-4-turbo",
"top_p": 0.95, "top_p": 0.95,
"temperature": 0.9 "temperature": 0.6
} }
response = requests.post( response = requests.post(
chatgpt_url_predict, chatgpt_url_predict,
@ -428,21 +437,26 @@ def biaogegongshi_kongzhi_prompt(title, mulu_list, table_and_formulas_list):
prompt_biaogekongzhi["prompt3_multiple_tags"].format("&".join(lable_en), "&".join(lable_en)), prompt_biaogekongzhi["prompt3_multiple_tags"].format("&".join(lable_en), "&".join(lable_en)),
]) ])
prompt = "任务：公式和表格生成控制\n{}".format(prompt) prompt = "任务：公式和表格生成控制\n{}".format(prompt)
prompt = qwen_think_prompt.format(prompt) prompt = qwen_prompt.format(prompt)
prompt_list.append(prompt) prompt_list.append(prompt)
return prompt_list return prompt_list
@app.route("/articles_directory", methods=["POST"]) @app.route("/articles_directory", methods=["POST"])
def articles_directory(): def articles_directory():
title = request.json["title"] # 获取用户query中的文本 例如"I love you" title = request.json["title"]
nums_catalogue = request.json["nums_catalogue"] nums_catalogue = request.json["nums_catalogue"]
nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you" nums_word = request.json["nums_word"]
main_content = request.json.get("main_content")
table_and_formulas = request.json["table_and_formulas"] table_and_formulas = request.json["table_and_formulas"]
table_and_formulas_list = str(table_and_formulas).split(",") table_and_formulas_list = str(table_and_formulas).split(",")
nums_catalogue = int(nums_catalogue) nums_catalogue = int(nums_catalogue)
keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"]) if not main_content:
prompt = nums_word_dict[nums_word]["prompt"]%keyword keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt"] % keyword
else:
keyword = tuple([title, main_content] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt_main_content"] % keyword
print(prompt) print(prompt)
input_data = [] input_data = []
for i in range(nums_catalogue): for i in range(nums_catalogue):
@ -474,7 +488,9 @@ def articles_directory():
"catalog_list": mulu_list, "catalog_list": mulu_list,
"catalog_dict": results "catalog_dict": results
} }
# log.log('text:{},'.format(prompt)) log.log('start at',
'text:{},results:{}'.format(
prompt, str(results)))
return jsonify(data) # 返回结果 return jsonify(data) # 返回结果
if __name__ == "__main__": if __name__ == "__main__":

56
articles_directory_predict_word.py
View File

@ -53,51 +53,67 @@ prompt = {
"mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好" "mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好"
} }
prompt_main_content = {
"mulu_title_Level_2": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题和二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s个二级标题",
"mulu_title_Level_2_1": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题和二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题",
"mulu_title_Level_3": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s个二级标题；三级标题个数适中",
"mulu_title_Level_3_1": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
"mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好"
}
nums_word_dict = { nums_word_dict = {
"3000": { "3000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["2"], "title_2_nums": ["2"],
"title_small_nums": ["3", "2"] "title_small_nums": ["3", "2"]
}, },
"5000": { "5000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["2","4"], "title_2_nums": ["2","4"],
"title_small_nums": ["3", "2","4"] "title_small_nums": ["3", "2","4"]
}, },
"8000": { "8000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "4", "title_1_nums": "4",
"title_2_nums": ["2", "4"], "title_2_nums": ["2", "4"],
"title_small_nums": ["4", "2", "4"], "title_small_nums": ["4", "2", "4"],
}, },
"10000": { "10000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["2", "5"], "title_2_nums": ["2", "5"],
"title_small_nums": ["6", "2", "5"], "title_small_nums": ["6", "2", "5"],
}, },
"15000": { "15000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "5", "title_1_nums": "5",
"title_2_nums": ["2", "3"], "title_2_nums": ["2", "3"],
"title_small_nums": ["5", "2", "3"], "title_small_nums": ["5", "2", "3"],
}, },
"20000": { "20000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["2", "4"], "title_2_nums": ["2", "4"],
"title_small_nums": ["6", "2", "4"], "title_small_nums": ["6", "2", "4"],
}, },
"30000": { "30000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "7", "title_1_nums": "7",
"title_2_nums": ["3", "5"], "title_2_nums": ["3", "5"],
"title_small_nums": ["7", "3", "5"], "title_small_nums": ["7", "3", "5"],
}, },
"50000": { "50000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_2"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_2"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "8", "title_1_nums": "8",
"title_2_nums": ["3", "8"], "title_2_nums": ["3", "8"],
"title_small_nums": ["8", "3", "8"], "title_small_nums": ["8", "3", "8"],
@ -149,7 +165,7 @@ def request_api_chatgpt(prompt):
"content": prompt, "content": prompt,
"model": "gpt-4-turbo", "model": "gpt-4-turbo",
"top_p": 0.95, "top_p": 0.95,
"temperature": 0.7 "temperature": 0.6
} }
response = requests.post( response = requests.post(
chatgpt_url_predict, chatgpt_url_predict,
@ -241,19 +257,31 @@ def get_multiple_urls(urls):
@app.route("/articles_directory", methods=["POST"]) @app.route("/articles_directory", methods=["POST"])
def articles_directory(): def articles_directory():
title = request.json["title"] # 获取用户query中的文本 例如"I love you" title = request.json["title"]
nums_catalogue = request.json["nums_catalogue"] nums_catalogue = request.json["nums_catalogue"]
nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you" nums_word = request.json["nums_word"]
main_content = request.json.get("main_content")
print(title)
print(nums_catalogue)
print(nums_word)
print("main_content", main_content)
nums_catalogue = int(nums_catalogue) nums_catalogue = int(nums_catalogue)
keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"]) if not main_content:
prompt = nums_word_dict[nums_word]["prompt"]%keyword keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt"] % keyword
else:
keyword = tuple([title, main_content] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt_main_content"] % keyword
print(prompt) print(prompt)
input_data = [] input_data = []
for i in range(nums_catalogue): for i in range(nums_catalogue):
input_data.append([i, prompt]) input_data.append([i, prompt])
results = get_multiple_urls(input_data) results = get_multiple_urls(input_data)
log.log('text:{},'.format(prompt)) log.log('start at',
'text:{},results:{}'.format(
prompt, str(results)))
return jsonify(results) # 返回结果 return jsonify(results) # 返回结果
if __name__ == "__main__": if __name__ == "__main__":

31
articles_directory_predict_word_online.py
View File

@ -53,51 +53,67 @@ prompt = {
"mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好" "mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好"
} }
prompt_main_content = {
"mulu_title_Level_2": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题和二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s个二级标题",
"mulu_title_Level_2_1": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题和二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题",
"mulu_title_Level_3": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s个二级标题；三级标题个数适中",
"mulu_title_Level_3_1": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
"mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好"
}
nums_word_dict = { nums_word_dict = {
"3000": { "3000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_2"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["2"], "title_2_nums": ["2"],
"title_small_nums": ["3", "2"] "title_small_nums": ["3", "2"]
}, },
"5000": { "5000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["2","4"], "title_2_nums": ["2","4"],
"title_small_nums": ["3", "2","4"] "title_small_nums": ["3", "2","4"]
}, },
"8000": { "8000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "4", "title_1_nums": "4",
"title_2_nums": ["2", "4"], "title_2_nums": ["2", "4"],
"title_small_nums": ["4", "2", "4"], "title_small_nums": ["4", "2", "4"],
}, },
"10000": { "10000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["2", "5"], "title_2_nums": ["2", "5"],
"title_small_nums": ["6", "2", "5"], "title_small_nums": ["6", "2", "5"],
}, },
"15000": { "15000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "5", "title_1_nums": "5",
"title_2_nums": ["2", "3"], "title_2_nums": ["2", "3"],
"title_small_nums": ["5", "2", "3"], "title_small_nums": ["5", "2", "3"],
}, },
"20000": { "20000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["2", "4"], "title_2_nums": ["2", "4"],
"title_small_nums": ["6", "2", "4"], "title_small_nums": ["6", "2", "4"],
}, },
"30000": { "30000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "7", "title_1_nums": "7",
"title_2_nums": ["3", "5"], "title_2_nums": ["3", "5"],
"title_small_nums": ["7", "3", "5"], "title_small_nums": ["7", "3", "5"],
}, },
"50000": { "50000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_2"]), "prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_2"]),
"prompt_main_content": qwen_think_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "8", "title_1_nums": "8",
"title_2_nums": ["3", "8"], "title_2_nums": ["3", "8"],
"title_small_nums": ["8", "3", "8"], "title_small_nums": ["8", "3", "8"],
@ -148,7 +164,7 @@ def request_api_chatgpt(prompt):
"content": prompt, "content": prompt,
"model": "gpt-4-turbo", "model": "gpt-4-turbo",
"top_p": 0.95, "top_p": 0.95,
"temperature": 0.7 "temperature": 0.6
} }
response = requests.post( response = requests.post(
chatgpt_url_predict, chatgpt_url_predict,
@ -243,16 +259,23 @@ def articles_directory():
title = request.json["title"] # 获取用户query中的文本 例如"I love you" title = request.json["title"] # 获取用户query中的文本 例如"I love you"
nums_catalogue = request.json["nums_catalogue"] nums_catalogue = request.json["nums_catalogue"]
nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you" nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you"
main_content = request.json.get("main_content")
nums_catalogue = int(nums_catalogue) nums_catalogue = int(nums_catalogue)
keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"]) if not main_content:
prompt = nums_word_dict[nums_word]["prompt"]%keyword keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt"] % keyword
else:
keyword = tuple([title, main_content] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt_main_content"] % keyword
print(prompt) print(prompt)
input_data = [] input_data = []
for i in range(nums_catalogue): for i in range(nums_catalogue):
input_data.append([i, prompt]) input_data.append([i, prompt])
results = get_multiple_urls(input_data) results = get_multiple_urls(input_data)
log.log('text:{},'.format(prompt)) log.log('start at',
'text:{},results:{}'.format(
prompt, str(results)))
return jsonify(results) # 返回结果 return jsonify(results) # 返回结果
if __name__ == "__main__": if __name__ == "__main__":

61
articles_directory_predict_word_table_formula_kongzhi.py
View File

@ -42,10 +42,10 @@ def get_host_ip():
return ip return ip
# chatgpt_url_predict = "http://{}:12000/predict".format(str(get_host_ip())) chatgpt_url_predict = "http://{}:12000/predict".format(str(get_host_ip()))
# chatgpt_url_search = "http://{}:12000/search".format(str(get_host_ip())) chatgpt_url_search = "http://{}:12000/search".format(str(get_host_ip()))
chatgpt_url_predict = "http://{}:12000/predict".format("192.168.31.149") # chatgpt_url_predict = "http://{}:12000/predict".format("192.168.31.149")
chatgpt_url_search = "http://{}:12000/search".format("192.168.31.149") # chatgpt_url_search = "http://{}:12000/search".format("192.168.31.149")
qwen_prompt = "<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n" qwen_prompt = "<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n"
qwen_think_prompt = "<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n" qwen_think_prompt = "<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n"
@ -58,6 +58,14 @@ prompt = {
"mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好" "mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好"
} }
prompt_main_content = {
"mulu_title_Level_2": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题和二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s个二级标题",
"mulu_title_Level_2_1": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题和二级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题",
"mulu_title_Level_3": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s个二级标题；三级标题个数适中",
"mulu_title_Level_3_1": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数适中",
"mulu_title_Level_3_2": "任务：生成目录\n为论文题目“%s”生成中文目录，生成目录需要参考一些关键内容：“%s”，要求只有一级标题，二级标题和三级标题，一级标题使用中文数字 例如一、xxx；二级标题使用阿拉伯数字 例如1.1 xxx；三级标题使用阿拉伯数字 例如1.1.2 xxx；一级标题生成%s个；每个一级标题包含%s-%s个二级标题；三级标题个数越多越好"
}
prompt_biaogekongzhi = { prompt_biaogekongzhi = {
"prompt_1":"背景:我是一名博士生，我想写一篇论文。论文题目《{}》，目录是“{}”，我现在需要论文中有{}，请根据这些信息告诉我", "prompt_1":"背景:我是一名博士生，我想写一篇论文。论文题目《{}》，目录是“{}”，我现在需要论文中有{}，请根据这些信息告诉我",
"prompt_1_form":"哪些小标题生成的内容需要包含表格", "prompt_1_form":"哪些小标题生成的内容需要包含表格",
@ -89,49 +97,57 @@ id_to_lable = {
nums_word_dict = { nums_word_dict = {
"3000": { "3000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["2"], "title_2_nums": ["2"],
"title_small_nums": ["3", "2"] "title_small_nums": ["3", "2"]
}, },
"5000": { "5000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "3", "title_1_nums": "3",
"title_2_nums": ["2","4"], "title_2_nums": ["2","4"],
"title_small_nums": ["3", "2","4"] "title_small_nums": ["3", "2","4"]
}, },
"8000": { "8000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "4", "title_1_nums": "4",
"title_2_nums": ["2", "4"], "title_2_nums": ["2", "4"],
"title_small_nums": ["4", "2", "4"], "title_small_nums": ["4", "2", "4"],
}, },
"10000": { "10000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_2_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_2_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_2_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["2", "5"], "title_2_nums": ["2", "5"],
"title_small_nums": ["6", "2", "5"], "title_small_nums": ["6", "2", "5"],
}, },
"15000": { "15000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "5", "title_1_nums": "5",
"title_2_nums": ["2", "3"], "title_2_nums": ["2", "3"],
"title_small_nums": ["5", "2", "3"], "title_small_nums": ["5", "2", "3"],
}, },
"20000": { "20000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "6", "title_1_nums": "6",
"title_2_nums": ["2", "4"], "title_2_nums": ["2", "4"],
"title_small_nums": ["6", "2", "4"], "title_small_nums": ["6", "2", "4"],
}, },
"30000": { "30000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_1"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_1"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "7", "title_1_nums": "7",
"title_2_nums": ["3", "5"], "title_2_nums": ["3", "5"],
"title_small_nums": ["7", "3", "5"], "title_small_nums": ["7", "3", "5"],
}, },
"50000": { "50000": {
"prompt": qwen_think_prompt.format(prompt["mulu_title_Level_3_2"]), "prompt": qwen_prompt.format(prompt["mulu_title_Level_3_2"]),
"prompt_main_content": qwen_prompt.format(prompt_main_content["mulu_title_Level_3_1"]),
"title_1_nums": "8", "title_1_nums": "8",
"title_2_nums": ["3", "8"], "title_2_nums": ["3", "8"],
"title_small_nums": ["8", "3", "8"], "title_small_nums": ["8", "3", "8"],
@ -183,7 +199,7 @@ def request_api_chatgpt(prompt):
"content": prompt, "content": prompt,
"model": "gpt-4-turbo", "model": "gpt-4-turbo",
"top_p": 0.95, "top_p": 0.95,
"temperature": 0.7 "temperature": 0.6
} }
response = requests.post( response = requests.post(
chatgpt_url_predict, chatgpt_url_predict,
@ -456,21 +472,26 @@ def biaogegongshi_kongzhi_prompt(title, mulu_list, table_and_formulas_list):
prompt_biaogekongzhi["prompt3_multiple_tags"].format("&".join(lable_en), "&".join(lable_en)), prompt_biaogekongzhi["prompt3_multiple_tags"].format("&".join(lable_en), "&".join(lable_en)),
]) ])
prompt = "任务：公式和表格生成控制\n{}".format(prompt) prompt = "任务：公式和表格生成控制\n{}".format(prompt)
prompt = qwen_think_prompt.format(prompt) prompt = qwen_prompt.format(prompt)
prompt_list.append(prompt) prompt_list.append(prompt)
return prompt_list return prompt_list
@app.route("/articles_directory", methods=["POST"]) @app.route("/articles_directory", methods=["POST"])
def articles_directory(): def articles_directory():
title = request.json["title"] # 获取用户query中的文本 例如"I love you" title = request.json["title"]
nums_catalogue = request.json["nums_catalogue"] nums_catalogue = request.json["nums_catalogue"]
nums_word = request.json["nums_word"] # 获取用户query中的文本 例如"I love you" nums_word = request.json["nums_word"]
main_content = request.json.get("main_content")
table_and_formulas = request.json["table_and_formulas"] table_and_formulas = request.json["table_and_formulas"]
table_and_formulas_list = str(table_and_formulas).split(",") table_and_formulas_list = str(table_and_formulas).split(",")
nums_catalogue = int(nums_catalogue) nums_catalogue = int(nums_catalogue)
keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"]) if not main_content:
prompt = nums_word_dict[nums_word]["prompt"]%keyword keyword = tuple([title] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt"] % keyword
else:
keyword = tuple([title, main_content] + nums_word_dict[nums_word]["title_small_nums"])
prompt = nums_word_dict[nums_word]["prompt_main_content"] % keyword
print(prompt) print(prompt)
input_data = [] input_data = []
for i in range(nums_catalogue): for i in range(nums_catalogue):
@ -492,7 +513,9 @@ def articles_directory():
"catalog_list": mulu_list, "catalog_list": mulu_list,
"catalog_dict": results "catalog_dict": results
} }
# log.log('text:{},'.format(prompt)) log.log('start at',
'text:{},results:{}'.format(
prompt, str(results)))
return jsonify(data) # 返回结果 return jsonify(data) # 返回结果
if __name__ == "__main__": if __name__ == "__main__":

Write Preview
Loading…
Cancel
Save