app.py

from openai import OpenAI
import streamlit as st
from langchain_openai import ChatOpenAI
from tools import sentiment_analysis_util
import numpy as np
import pandas as pd
from dotenv import load_dotenv
import os

st.set_page_config(page_title="LangChain Agent", layout="wide")
load_dotenv()
OPENAI_API_KEY = os.environ["OPENAI_API_KEY"]

from langchain_core.runnables import RunnableConfig

st.title("💬 News Search")
st.image('el_pic.png')

#@st.cache_resource
if "messages" not in st.session_state:
    st.session_state["messages"] = [{"role":"system", "content":"""💬 How can I help you?"""}]

# Display all previous messages
for msg in st.session_state.messages:
    st.chat_message(msg["role"]).write(msg["content"])

#initialize_session_state()


sideb=st.sidebar
with st.sidebar:
    prompt=st.text_input("Enter topic for sentiment analysis: ")

check1=sideb.button(f"analyze {prompt}")

if check1:
    # Add user message to chat history
    st.session_state.messages.append({"role": "user", "content": prompt})
    # Display user message in chat message container
    with st.chat_message("user"):
        st.markdown(prompt)

    # ========================== Sentiment analysis
    #Perform sentiment analysis on the cryptocurrency news & predict dominant sentiment along with plotting the sentiment breakdown chart
    # Downloading from reddit

    # Downloading from alpaca
    if len(prompt.split(' '))<3:
        st.write('I am analyzing Google News ...')
        news_articles = sentiment_analysis_util.fetch_news(str(prompt))
    st.write('Now, I am analyzing Reddit ...')
    reddit_news_articles=sentiment_analysis_util.fetch_reddit_news(prompt)
    # Fetch news articles
    tavily_news_articles = sentiment_analysis_util.fetch_tavily_news(prompt)

    # Handle empty results
    if not tavily_news_articles:
        print("No news articles found. Try adjusting your search terms.")
    else:
        # Process the articles
        for url in tavily_news_articles:
            try:
                # Your existing article processing code
                st.write(f"Article URL: {url}")
                # ... rest of your processing ...
            except Exception as e:
                st.error(f"Error processing article {url}: {e}")
                continue
    analysis_results = []

    #Perform sentiment analysis for each product review
    if len(prompt.split(' '))<3:
        for article in news_articles:
            if prompt.lower()[0:6] in article['News_Article'].lower():
                sentiment_analysis_result = sentiment_analysis_util.analyze_sentiment(article['News_Article'])

                # Display sentiment analysis results
                #print(f'News Article: {sentiment_analysis_result["News_Article"]} : Sentiment: {sentiment_analysis_result["Sentiment"]}', '\n')

                result = {
                            'News_Article': sentiment_analysis_result["News_Article"],
                            'Sentiment': sentiment_analysis_result["Sentiment"][0]['label'],
                            'Index': sentiment_analysis_result["Sentiment"][0]['score'],
                            'URL': article['URL']
                        }
                
                analysis_results.append(result)

    articles_url=[]
    for article in reddit_news_articles:
        if prompt.lower()[0:6] in article.lower():
            sentiment_analysis_result_reddit = sentiment_analysis_util.analyze_sentiment(article)

            # Display sentiment analysis results
            #print(f'News Article: {sentiment_analysis_result_reddit["News_Article"]} : Sentiment: {sentiment_analysis_result_reddit["Sentiment"]}', '\n')

            result = {
                        'News_Article': sentiment_analysis_result_reddit["News_Article"],
                        'Index':np.round(sentiment_analysis_result_reddit["Sentiment"][0]['score'],2)
                    }
            analysis_results.append(np.append(result,np.append(article.split('URL:')[-1:], ((article.split('Date: ')[-1:])[0][0:10]))))
    
    for article in tavily_news_articles:
        if prompt.lower()[0:5] in article:
            sentiment_analysis_result_tavily = sentiment_analysis_util.analyze_sentiment(article)

            # Display sentiment analysis results
            #print(f'News Article: {sentiment_analysis_result_tavily["News_Article"]} : Sentiment: {sentiment_analysis_result_tavily["Sentiment"]}', '\n')

            result = {
                        'News_Article': sentiment_analysis_result_tavily["News_Article"],
                        'Index':np.round(sentiment_analysis_result_tavily["Sentiment"][0]['score'],2)
                    }
            analysis_results.append(np.append(result,np.append(article.split('URL:')[-1:], ((article.split('Date: ')[-1:])[0][0:10]))))
            print('is_present tavily 2',analysis_results)
    # #LinkedIn and Twitter previously downloaded from phantombuster
    # st.write('Teď analyzuji data z LinkedInu a Twitteru ...')
    # df=pd.read_csv('./data/LinkedIn_transurban_phantombuster.csv',index_col='postTimestamp',parse_dates=True,infer_datetime_format=True)
    # df=df.sort_index(ascending=False)
    # df=df.dropna()
    # from tools import sentiment_analysis_util
    # for linkedin_news in df['postContent']:
    #     print(linkedin_news)
    #     news_article={
    #         'News_Article': linkedin_news,
    #         'URL': df.loc[df['postContent']==linkedin_news]['postUrl'][0],
    #         'date': df.loc[df['postContent']==linkedin_news].index[0]}
    #     if prompt.lower()[0:6] in linkedin_news.lower():
    #         sentiment_analysis_result = sentiment_analysis_util.analyze_sentiment(news_article)

    #         news_article["Sentiment"]=sentiment_analysis_result["Sentiment"][0]['label']
    #         news_article["Index"]=sentiment_analysis_result["Sentiment"][0]['score']

    #         analysis_results.append(news_article)
    
    # count=0
    # df=pd.read_csv('./data/Twitter_transurban_phantombuster.csv',index_col='tweetDate',parse_dates=True,infer_datetime_format=True)
    # df=df.sort_index(ascending=False)
    # df=df.dropna()
    # from tools import sentiment_analysis_util
    # for twitter_news in df['text']:
    #     print(twitter_news)
    #     news_article={
    #         'News_Article': twitter_news,
    #         'URL': df['tweetLink'][count],
    #         'date': df.iloc[count:count+1,:].index[0]}
    #     if prompt.lower()[0:6] in twitter_news.lower():
    #         sentiment_analysis_result = sentiment_analysis_util.analyze_sentiment(news_article)
    #         news_article["Sentiment"]=sentiment_analysis_result["Sentiment"][0]['label']
    #         news_article["Index"]=sentiment_analysis_result["Sentiment"][0]['score']

    #         analysis_results.append(news_article)
    #     count+=1
    

    #Generate summarized message rationalize dominant sentiment
    #st.write(analysis_results)
    summary = sentiment_analysis_util.generate_summary_of_sentiment(analysis_results) #, dominant_sentiment)
    st.chat_message("assistant").write((summary))
    st.session_state.messages.append({"role": "assistant", "content": summary})
    #answers=np.append(res["messages"][-1].content,summary)

client = ChatOpenAI(model="gpt-4o",api_key=OPENAI_API_KEY)

if "openai_model" not in st.session_state:
    st.session_state["openai_model"] = "gpt-4o"

if prompt := st.chat_input("Any other questions? "):
    # Add user message to chat history
    st.session_state.messages.append({"role": "user", "content": prompt})
    # Display user message in chat message container
    with st.chat_message("user"):
        st.markdown(prompt)
    # Display assistant response in chat message container
    with st.chat_message("assistant"):
        stream = client.chat.completions.create(
            model=st.session_state["openai_model"],
            messages=[
                {"role": m["role"], "content": m["content"]}
                for m in st.session_state.messages
            ],
            stream=True,
        )
        response = st.write_stream(stream)
    st.session_state.messages.append({"role": "assistant", "content": response})