app.py

import streamlit as st
#from langchain.chat_models import ChatOpenAI
from langchain.document_loaders import WebBaseLoader
from langchain.chains.summarize import load_summarize_chain
from bs4 import BeautifulSoup   #WebBaseLoader会需要 用到？
from langchain import HuggingFaceHub
import requests
import sys
from huggingface_hub import InferenceClient

import os
from dotenv import load_dotenv
load_dotenv()
hf_token = os.environ.get('HUGGINGFACEHUB_API_TOKEN')
repo_id=os.environ.get('repo_id')
#port = os.getenv('port')

#OPENAI_API_KEY = os.environ.get('OPENAI_API_KEY')

llm = HuggingFaceHub(repo_id=repo_id,  # for StarChat
                     huggingfacehub_api_token=hf_token,  #这个变量huggingfacehub_api_token名称似乎没有问题！
                     model_kwargs={"min_length": 512,  # for StarChat
                                   "max_new_tokens": 1024, "do_sample": True,  # for StarChat
                                   "temperature": 0.01,
                                   "top_k": 50,
                                   "top_p": 0.95, "eos_token_id": 49155})

#chain = load_summarize_chain(llm, chain_type="stuff")    #stuff模式容易导致出错：估计是超LLM的token限制所致
chain = load_summarize_chain(llm, chain_type="refine")

#llm = ChatOpenAI(temperature=0, model_name="gpt-3.5-turbo-16k")

#**************************************************************#
#url=st.text_input("Enter webiste URL to summarize (format: https://www.usinoip.com):")
#loader = WebBaseLoader("https://www.usinoip.com/")
#loader = WebBaseLoader(url)
#with st.spinner("AI Thinking...Please wait a while to Cheers!"):
#    docs = loader.load()
#    result=chain.run(docs)
#    print(result)
#    st.write("AI Summarization:")
#    st.write(result)
#**************************************************************#

#**************************************************************#
url=st.text_input("Enter webiste URL to summarize (format: https://www.usinoip.com):")

#loader = WebBaseLoader("https://www.usinoip.com/")

if url !="" and not url.strip().isspace() and not url == "" and not url.strip() == "" and not url.isspace():
    try:
        loader = WebBaseLoader(url)
        with st.spinner("AI Thinking...Please wait a while to Cheers!"):
            docs = loader.load()
            result=chain.run(docs)   #这个result的格式比较特殊，可以直接print，但不可以和其他字符串联合print输出
            print(url)
            #print("AI Summarization: "+result)   #这个会出错，原因见上方
            print("AI Summarization:")
            print(result)
            
            st.write("AI Summarization:")
            st.write(result)
    except Exception as e:
        st.write("Wrong URL or URL not parsable.")    
#**************************************************************#

#try:
#    loader = WebBaseLoader(url)
#    with st.spinner("AI Thinking...Please wait a while to Cheers!"):
#        docs = loader.load()
#        result=chain.run(docs)
#        print(result)
#        st.write("AI Summarization:")
#        st.write(result)
#except Exception as e:
#    st.write("Wrong URL")