Add 2 plugins
相当于将“批量总结PDF文档”插件拆成了两部分，目的在于使用廉价的模型干粗活，再将关键的最终总结交给GPT-4，降低使用成本 批量总结PDF文档_初步：初步总结PDF，每个PDF输出一个md文档 批量总结Markdown文档_进阶：将所有md文档高度凝练并汇总至一个md文档，可直接使用“批量总结PDF文档_初步”的输出结果作为输入
This commit is contained in:
leike0813
parent
ac3d4cf073
commit
68a49d3758
@ -603,7 +603,35 @@ def get_crazy_functions():
|
||||
except:
|
||||
print(trimmed_format_exc())
|
||||
print('Load function plugin failed')
|
||||
try:
|
||||
from crazy_functions.批量总结PDF文档_初步 import 批量总结PDF文档_初步
|
||||
function_plugins.update({
|
||||
"批量总结PDF文档_初步": {
|
||||
"Group": "学术",
|
||||
"Color": "stop",
|
||||
"AsButton": False,
|
||||
"Info": "批量总结PDF文档的内容(仅做初步提炼) | 输入参数为路径",
|
||||
"Function": HotReload(批量总结PDF文档_初步)
|
||||
}
|
||||
})
|
||||
except:
|
||||
print(trimmed_format_exc())
|
||||
print('Load function plugin failed')
|
||||
|
||||
try:
|
||||
from crazy_functions.批量总结Markdown文档_进阶 import 批量总结Markdown文档_进阶
|
||||
function_plugins.update({
|
||||
"批量总结Markdown文档_进阶": {
|
||||
"Group": "学术",
|
||||
"Color": "stop",
|
||||
"AsButton": False,
|
||||
"Info": "批量总结Markdown文档的内容(在初步提炼的基础上进一步总结) | 输入参数为路径",
|
||||
"Function": HotReload(批量总结Markdown文档_进阶)
|
||||
}
|
||||
})
|
||||
except:
|
||||
print(trimmed_format_exc())
|
||||
print('Load function plugin failed')
|
||||
# try:
|
||||
# from crazy_functions.chatglm微调工具 import 微调数据集生成
|
||||
# function_plugins.update({
|
||||
|
||||
127
crazy_functions/批量总结Markdown文档_进阶.py
Normal file
127
crazy_functions/批量总结Markdown文档_进阶.py
Normal file
@ -0,0 +1,127 @@
|
||||
import logging, os
|
||||
from toolbox import update_ui, promote_file_to_downloadzone, gen_time_str, get_log_folder
|
||||
from toolbox import CatchException, report_exception, trimmed_format_exc
|
||||
from toolbox import write_history_to_file, promote_file_to_downloadzone
|
||||
from .crazy_utils import request_gpt_model_in_new_thread_with_ui_alive
|
||||
from .crazy_utils import request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency
|
||||
from .crazy_utils import input_clipping
|
||||
|
||||
|
||||
def 总结Markdown(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt):
|
||||
file_write_buffer = []
|
||||
SUMMARY_WORD_LIMIT = 800
|
||||
meta_inputs_array = []
|
||||
meta_inputs_show_user_array = []
|
||||
meta_sys_prompt_array = []
|
||||
inputs_array = []
|
||||
inputs_show_user_array = []
|
||||
sys_prompt_array = []
|
||||
file_name_array = []
|
||||
for idx, file_name in enumerate(file_manifest):
|
||||
print('begin analysis on:', file_name)
|
||||
file_name_array.append(f'# {idx}.{os.path.basename(file_name)}')
|
||||
|
||||
with open(file_name, 'r', encoding='utf-8', errors='replace') as f:
|
||||
file_content = f.read()
|
||||
|
||||
_ = file_content.split('## metadata')
|
||||
if len(_) >= 2:
|
||||
file_meta = _[-2]
|
||||
file_content = _[-1]
|
||||
else:
|
||||
file_meta = file_name
|
||||
|
||||
meta_inputs_array.append(
|
||||
"我需要你从一段文本中识别并提取出这篇文章的1.标题、2.作者、3.作者单位、4.关键词。"
|
||||
"其中，1.标题和4.关键词需要给出中文和英文的双语结果，2.作者和3.作者单位按原文语言给出。"
|
||||
"以下是需要你识别的文本: " + file_meta
|
||||
)
|
||||
meta_inputs_show_user_array.append(
|
||||
'开始分析元数据：' + file_name
|
||||
)
|
||||
meta_sys_prompt_array.append("As an academic professional, you need to extract basic informations of the paper from its metadata")
|
||||
|
||||
inputs_array.append(
|
||||
"我需要你根据我提供的文本总结一份Markdown文档，分为四个部分：1.研究背景，2.文章主要内容，3.主要创新点，4.结论。"
|
||||
+ f"各部分的题目采用二级标题前缀(## ),内容可适当的分为若干条，总字数不超过{SUMMARY_WORD_LIMIT}个中文字符."
|
||||
+ "以下是需要你处理的文本: " + file_content)
|
||||
inputs_show_user_array.append('开始总结：' + file_name)
|
||||
sys_prompt_array.append(f"As an academic professional, you need to summarize the text with less than {SUMMARY_WORD_LIMIT} Chinese characters")
|
||||
|
||||
gpt_meta_response_collection = yield from request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(
|
||||
inputs_array=meta_inputs_array,
|
||||
inputs_show_user_array=meta_inputs_show_user_array,
|
||||
llm_kwargs=llm_kwargs,
|
||||
chatbot=chatbot,
|
||||
history_array=[[""] for _ in range(len(inputs_array))],
|
||||
sys_prompt_array=meta_sys_prompt_array,
|
||||
# max_workers=5, # OpenAI所允许的最大并行过载
|
||||
scroller_max_len=80
|
||||
)
|
||||
|
||||
gpt_response_collection = yield from request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(
|
||||
inputs_array=inputs_array,
|
||||
inputs_show_user_array=inputs_show_user_array,
|
||||
llm_kwargs=llm_kwargs,
|
||||
chatbot=chatbot,
|
||||
history_array=[[""] for _ in range(len(inputs_array))],
|
||||
sys_prompt_array=sys_prompt_array,
|
||||
# max_workers=5, # OpenAI所允许的最大并行过载
|
||||
scroller_max_len=80
|
||||
)
|
||||
try:
|
||||
for idx, (gpt_say_meta, gpt_say) in enumerate(zip(gpt_meta_response_collection[1::2], gpt_response_collection[1::2])):
|
||||
file_write_buffer.append(file_name_array[idx])
|
||||
file_write_buffer.append("## 元数据\n\n" + gpt_say_meta)
|
||||
file_write_buffer.append(gpt_say)
|
||||
except:
|
||||
logging.error(trimmed_format_exc())
|
||||
|
||||
res = write_history_to_file(file_write_buffer, file_basename="result.md", auto_caption=False)
|
||||
promote_file_to_downloadzone(res, chatbot=chatbot)
|
||||
yield from update_ui(chatbot=chatbot, history=gpt_response_collection) # 刷新界面
|
||||
|
||||
|
||||
@CatchException
|
||||
def 批量总结Markdown文档_进阶(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
|
||||
import glob, os
|
||||
|
||||
# 基本信息：功能、贡献者
|
||||
chatbot.append([
|
||||
"函数插件功能？",
|
||||
"批量总结Markdown文档。函数插件贡献者: ValeriaWong，Eralien，Joshua Reed"])
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
|
||||
# 尝试导入依赖，如果缺少依赖，则给出安装建议
|
||||
try:
|
||||
import fitz
|
||||
except:
|
||||
report_exception(chatbot, history,
|
||||
a = f"解析项目: {txt}",
|
||||
b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade pymupdf```。")
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
return
|
||||
|
||||
# 清空历史，以免输入溢出
|
||||
history = []
|
||||
|
||||
# 检测输入参数，如没有给定输入参数，直接退出
|
||||
if os.path.exists(txt):
|
||||
project_folder = txt
|
||||
else:
|
||||
if txt == "": txt = '空空如也的输入栏'
|
||||
report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到本地项目或无权访问: {txt}")
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
return
|
||||
|
||||
# 搜索需要处理的文件清单
|
||||
file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.md', recursive=True)]
|
||||
|
||||
# 如果没找到任何文件
|
||||
if len(file_manifest) == 0:
|
||||
report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.md文件: {txt}")
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
return
|
||||
|
||||
# 开始正式执行任务
|
||||
yield from 总结Markdown(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt)
|
||||
131
crazy_functions/批量总结PDF文档_初步.py
Normal file
131
crazy_functions/批量总结PDF文档_初步.py
Normal file
@ -0,0 +1,131 @@
|
||||
import zipfile
|
||||
import os
|
||||
from toolbox import update_ui, promote_file_to_downloadzone, gen_time_str, get_log_folder
|
||||
from toolbox import CatchException, report_exception
|
||||
from toolbox import write_history_to_file, promote_file_to_downloadzone
|
||||
from .crazy_utils import request_gpt_model_in_new_thread_with_ui_alive
|
||||
from .crazy_utils import read_and_clean_pdf_text
|
||||
from .crazy_utils import input_clipping
|
||||
pj = os.path.join
|
||||
|
||||
|
||||
def move_file_to_zip(file_path, zip_file):
|
||||
zip_file.write(file_path, os.path.basename(file_path))
|
||||
os.remove(file_path)
|
||||
|
||||
|
||||
def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt):
|
||||
zip_file_path = pj(get_log_folder(), 'result.zip')
|
||||
with zipfile.ZipFile(zip_file_path, 'w') as zip_file:
|
||||
for file_name in file_manifest:
|
||||
file_write_buffer = []
|
||||
print('begin analysis on:', file_name)
|
||||
############################## <第 0 步，切割PDF> ##################################
|
||||
# 递归地切割PDF文件，每一块（尽量是完整的一个section，比如introduction，experiment等，必要时再进行切割）
|
||||
# 的长度必须小于 2500 个 Token
|
||||
file_content, page_one = read_and_clean_pdf_text(file_name) # （尝试）按照章节切割PDF
|
||||
file_content = file_content.encode('utf-8', 'ignore').decode() # avoid reading non-utf8 chars
|
||||
page_one = str(page_one).encode('utf-8', 'ignore').decode() # avoid reading non-utf8 chars
|
||||
|
||||
TOKEN_LIMIT_PER_FRAGMENT = 2500
|
||||
|
||||
from .crazy_utils import breakdown_txt_to_satisfy_token_limit_for_pdf
|
||||
from request_llms.bridge_all import model_info
|
||||
enc = model_info["gpt-3.5-turbo"]['tokenizer']
|
||||
def get_token_num(txt): return len(enc.encode(txt, disallowed_special=()))
|
||||
paper_fragments = breakdown_txt_to_satisfy_token_limit_for_pdf(
|
||||
txt=file_content, get_token_fn=get_token_num, limit=TOKEN_LIMIT_PER_FRAGMENT)
|
||||
page_one_fragments = breakdown_txt_to_satisfy_token_limit_for_pdf(
|
||||
txt=str(page_one), get_token_fn=get_token_num, limit=TOKEN_LIMIT_PER_FRAGMENT//4)
|
||||
# 为了更好的效果，我们剥离Introduction之后的部分（如果有）
|
||||
paper_meta = page_one_fragments[0].split('introduction')[0].split('Introduction')[0].split('INTRODUCTION')[0]
|
||||
|
||||
############################## <第 1 步，从摘要中提取高价值信息，放到history中> ##################################
|
||||
final_results = []
|
||||
final_results.append("## metadata\n\n" + paper_meta + "\n\n## metadata")
|
||||
|
||||
############################## <第 2 步，迭代地历遍整个文章，提取精炼信息> ##################################
|
||||
i_say_show_user = f'首先你在中文语境下通读整篇论文。'; gpt_say = "[Local Message] 收到。" # 用户提示
|
||||
chatbot.append([i_say_show_user, gpt_say]); yield from update_ui(chatbot=chatbot, history=[]) # 更新UI
|
||||
|
||||
iteration_results = []
|
||||
last_iteration_result = paper_meta # 初始值是摘要
|
||||
MAX_WORD_TOTAL = 4096 * 0.7
|
||||
n_fragment = len(paper_fragments)
|
||||
if n_fragment >= 20: print('文章极长，不能达到预期效果')
|
||||
for i in range(n_fragment):
|
||||
NUM_OF_WORD = MAX_WORD_TOTAL // n_fragment
|
||||
i_say = f"Read this section, recapitulate the content of this section in Chinese with less than {NUM_OF_WORD} Chinese characters: {paper_fragments[i]}"
|
||||
i_say_show_user = f"[{i+1}/{n_fragment}] Read this section, recapitulate the content of this section with less than {NUM_OF_WORD} Chinese characters: {paper_fragments[i][:200]}"
|
||||
gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(i_say, i_say_show_user, # i_say=真正给chatgpt的提问， i_say_show_user=给用户看的提问
|
||||
llm_kwargs, chatbot,
|
||||
history=["The main idea of the previous section is?", last_iteration_result], # 迭代上一次的结果
|
||||
sys_prompt="Extract the main idea of this section with Chinese." # 提示
|
||||
)
|
||||
iteration_results.append(gpt_say)
|
||||
last_iteration_result = gpt_say
|
||||
|
||||
############################## <第 3 步，整理history，提取总结> ##################################
|
||||
final_results.extend(iteration_results)
|
||||
file_write_buffer.extend(final_results)
|
||||
|
||||
############################## <第 4 步，设置一个token上限> ##################################
|
||||
_, final_results = input_clipping("", final_results, max_token_limit=3200)
|
||||
yield from update_ui(chatbot=chatbot, history=final_results) # 注意这里的历史记录被替代了
|
||||
|
||||
res = write_history_to_file(
|
||||
file_write_buffer,
|
||||
file_basename=os.path.splitext(os.path.basename(file_name))[0] + '.md',
|
||||
auto_caption=False
|
||||
)
|
||||
if len(file_manifest) == 1:
|
||||
promote_file_to_downloadzone(res, chatbot=chatbot)
|
||||
return
|
||||
move_file_to_zip(res, zip_file)
|
||||
|
||||
promote_file_to_downloadzone(zip_file_path, chatbot=chatbot)
|
||||
|
||||
|
||||
@CatchException
|
||||
def 批量总结PDF文档_初步(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
|
||||
import glob, os
|
||||
|
||||
# 基本信息：功能、贡献者
|
||||
chatbot.append([
|
||||
"函数插件功能？",
|
||||
"批量总结PDF文档。函数插件贡献者: ValeriaWong，Eralien，Joshua Reed"])
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
|
||||
# 尝试导入依赖，如果缺少依赖，则给出安装建议
|
||||
try:
|
||||
import fitz
|
||||
except:
|
||||
report_exception(chatbot, history,
|
||||
a = f"解析项目: {txt}",
|
||||
b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade pymupdf```。")
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
return
|
||||
|
||||
# 清空历史，以免输入溢出
|
||||
history = []
|
||||
|
||||
# 检测输入参数，如没有给定输入参数，直接退出
|
||||
if os.path.exists(txt):
|
||||
project_folder = txt
|
||||
else:
|
||||
if txt == "": txt = '空空如也的输入栏'
|
||||
report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到本地项目或无权访问: {txt}")
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
return
|
||||
|
||||
# 搜索需要处理的文件清单
|
||||
file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.pdf', recursive=True)]
|
||||
|
||||
# 如果没找到任何文件
|
||||
if len(file_manifest) == 0:
|
||||
report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.tex或.pdf文件: {txt}")
|
||||
yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
|
||||
return
|
||||
|
||||
# 开始正式执行任务
|
||||
yield from 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt)
|
||||
Loading…
x
Reference in New Issue
Block a user