e-mail

app.py

@@ -1,5 +1,9 @@
 import os
+import json
+import smtplib
 import streamlit as st
+from email.mime.multipart import MIMEMultipart
+from email.mime.text import MIMEText
 from dotenv import load_dotenv
 from langchain_groq import ChatGroq
 from langchain_huggingface import HuggingFaceEmbeddings
@@ -11,76 +15,84 @@ from langchain_core.output_parsers import StrOutputParser
 from langchain_core.runnables import RunnableLambda
 from requests.exceptions import RequestException, Timeout
 
-# Установка конфигурации страницы
+# Streamlit page configuration
 st.set_page_config(page_title="Legal Chatbot", page_icon="🤖")
 
-# Загрузка переменных окружения
+# Load environment variables
 if os.path.exists(".env"):
     load_dotenv(verbose=True)
 
-# Загрузка API-ключей
+# Load API keys
 try:
     GROQ_API_KEY = st.secrets["GROQ_API_KEY"]
     USER_AGENT = st.secrets["USER_AGENT"]
     OPENAI_API_KEY = st.secrets["OPENAI_API_KEY"]
+    EMAIL_SENDER = st.secrets["EMAIL_SENDER"]
+    EMAIL_PASSWORD = st.secrets["EMAIL_PASSWORD"]
 except FileNotFoundError:
     GROQ_API_KEY = os.getenv("GROQ_API_KEY")
     USER_AGENT = os.getenv("USER_AGENT")
     OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+    EMAIL_SENDER = os.getenv("EMAIL_SENDER")
+    EMAIL_PASSWORD = os.getenv("EMAIL_PASSWORD")
 
-# Проверка API-ключей
-if not all([GROQ_API_KEY, USER_AGENT, OPENAI_API_KEY]):
-    st.error("Ошибка: Не все переменные окружения заданы.")
+# Check if API keys are set
+if not all([GROQ_API_KEY, USER_AGENT, OPENAI_API_KEY, EMAIL_SENDER, EMAIL_PASSWORD]):
+    st.error("Error: Missing required environment variables.")
     st.stop()
 
-# Инициализация LLM
+# Initialize LLM
 try:
     llm = ChatGroq(model_name="llama-3.3-70b-versatile", temperature=0.6, api_key=GROQ_API_KEY)
 except Exception as e:
-    st.error(f"Ошибка инициализации LLM: {e}")
+    st.error(f"LLM initialization failed: {e}")
     st.stop()
 
-# Инициализация эмбеддингов
+# Initialize embeddings model
 embeddings_model = HuggingFaceEmbeddings(model_name="intfloat/multilingual-e5-large-instruct")
 
-# Список страниц для анализа
+# List of website pages for knowledge base
 urls = [
-    "https://status.law",  
+    "https://status.law",
     "https://status.law/about",
     "https://status.law/careers",
     "https://status.law/challenging-sanctions",
-    "https://status.law/contact", 
-    "https://status.law/cross-border-banking-legal-issues", 
-    "https://status.law/extradition-defense", 
-    "https://status.law/international-prosecution-protection", 
-    "https://status.law/interpol-red-notice-removal",  
-    "https://status.law/practice-areas",  
+    "https://status.law/contact",
+    "https://status.law/cross-border-banking-legal-issues",
+    "https://status.law/extradition-defense",
+    "https://status.law/international-prosecution-protection",
+    "https://status.law/interpol-red-notice-removal",
+    "https://status.law/practice-areas",
     "https://status.law/reputation-protection",
     "https://status.law/faq"
 ]
 
-# Путь к файлу векторного хранилища
-VECTOR_STORE_PATH = "vector_store"
+# Path to knowledge base storage
+VECTOR_STORE_PATH = "storage/vector_store"
+HISTORY_PATH = "storage/chat_history.json"
 
-# Функция для создания базы знаний
+# Function to build knowledge base
 def build_knowledge_base():
     documents = []
     for url in urls:
         try:
             loader = WebBaseLoader(url)
             documents.extend(loader.load())
+            st.write(f"[INFO] Loaded content from {url}")
         except (RequestException, Timeout) as e:
-            st.write(f"[ERROR] Ошибка загрузки страницы {url}: {e}")
-    
+            st.write(f"[ERROR] Failed to load {url}: {e}")
+
     text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=100)
     chunks = text_splitter.split_documents(documents)
-    
+    st.write(f"[INFO] Split into {len(chunks)} chunks")
+
     vector_store = FAISS.from_documents(chunks, embeddings_model)
     vector_store.save_local(VECTOR_STORE_PATH)
-    
+
+    st.write("[INFO] Knowledge base successfully created and saved")
     return vector_store
 
-# Функция для загрузки базы знаний
+# Function to load existing knowledge base
 def load_knowledge_base():
     if os.path.exists(VECTOR_STORE_PATH):
         return FAISS.load_local(
@@ -90,66 +102,107 @@ def load_knowledge_base():
         )
     return None
 
-# Загружаем базу, если её нет в `st.session_state`
-if "vector_store" not in st.session_state or st.session_state.vector_store is None:
+# Function to load chat history
+def load_history():
+    if os.path.exists(HISTORY_PATH):
+        with open(HISTORY_PATH, "r", encoding="utf-8") as file:
+            return json.load(file)
+    return []
+
+# Function to save chat history
+def save_history(history):
+    with open(HISTORY_PATH, "w", encoding="utf-8") as file:
+        json.dump(history, file, ensure_ascii=False, indent=4)
+
+# Function to send chat history via email
+def send_email(recipient_email, subject, message):
+    try:
+        msg = MIMEMultipart()
+        msg["From"] = EMAIL_SENDER
+        msg["To"] = recipient_email
+        msg["Subject"] = subject
+        msg.attach(MIMEText(message, "plain"))
+
+        server = smtplib.SMTP("smtp.gmail.com", 587)
+        server.starttls()
+        server.login(EMAIL_SENDER, EMAIL_PASSWORD)
+        server.send_message(msg)
+        server.quit()
+        
+        return True
+    except Exception as e:
+        st.error(f"Email sending error: {e}")
+        return False
+
+# Load or create knowledge base
+if "vector_store" not in st.session_state:
     st.session_state.vector_store = load_knowledge_base()
 
-vector_store = st.session_state.vector_store
-
-# Если база знаний отсутствует, предлагаем её создать
-if vector_store is None:
-    st.write("База знаний не найдена. Нажмите кнопку, чтобы создать её.")
-    if st.button("Создать базу знаний"):
-        with st.spinner("Создание базы знаний..."):
+# Prompt to create knowledge base if missing
+if st.session_state.vector_store is None:
+    st.write("Knowledge base not found. Click the button to generate it.")
+    if st.button("Generate Knowledge Base"):
+        with st.spinner("Building knowledge base..."):
             st.session_state.vector_store = build_knowledge_base()
-            st.success("База знаний успешно создана!")
-            st.rerun()  # Перезапуск приложения
+            st.success("Knowledge base successfully created!")
+            st.rerun()
 else:
-    st.write("База знаний загружена. Вы можете задать вопрос.")
-
-    # Промпт для бота
-    template = """
-    You are a helpful legal assistant that answers questions based on information from status.law.
-    Answer accurately and concisely.
-    Question: {question}
-    Only use the provided context to answer the question.
-    Context: {context}
-    """
-    prompt = PromptTemplate.from_template(template)
-
-    # Инициализация цепочки обработки запроса
-    if "chain" not in st.session_state:
-        st.session_state.chain = (
-            RunnableLambda(lambda x: {"context": x["context"], "question": x["question"]}) 
-            | prompt
-            | llm
-            | StrOutputParser()
-        )
-    chain = st.session_state.chain
-
-    # Поле для ввода вопроса
-    user_input = st.text_input("Введите ваш вопрос:")
-    if st.button("Отправить") and user_input:
-        if st.session_state.vector_store:  # Проверяем, что база знаний загружена
-            retrieved_docs = st.session_state.vector_store.similarity_search(user_input)
-            context_text = "\n\n".join([doc.page_content for doc in retrieved_docs])
-
-            # Генерация ответа
-            response = chain.invoke({"question": user_input, "context": context_text})
-
-            # Сохранение истории сообщений
-            if "message_history" not in st.session_state:
-                st.session_state.message_history = []
-            st.session_state.message_history.append({"question": user_input, "answer": response})
-
-            # Вывод ответа
-            st.write(response)
-        else:
-            st.error("Ошибка: база знаний не загружена.")
-
-    # Вывод истории сообщений
-    if "message_history" in st.session_state:
-        st.write("### История сообщений")
-        for msg in st.session_state.message_history:
-            st.write(f"**User:** {msg['question']}")
-            st.write(f"**Bot:** {msg['answer']}")
+    st.write("Knowledge base loaded. You can ask questions.")
+
+# Chatbot prompt template
+template = """
+You are a helpful legal assistant answering questions based on information from status.law.
+Answer accurately and concisely.
+Question: {question}
+Only use the provided context to answer the question.
+Context: {context}
+"""
+prompt = PromptTemplate.from_template(template)
+
+# Initialize processing chain
+if "chain" not in st.session_state:
+    st.session_state.chain = (
+        RunnableLambda(lambda x: {"context": x["context"], "question": x["question"]})
+        | prompt
+        | llm
+        | StrOutputParser()
+    )
+
+chain = st.session_state.chain
+
+# Load chat history
+if "message_history" not in st.session_state:
+    st.session_state.message_history = load_history()
+
+# Chat input
+user_input = st.text_input("Enter your question:")
+if st.button("Send") and user_input:
+    vector_store = st.session_state.vector_store
+    retrieved_docs = vector_store.similarity_search(user_input)
+    context_text = "\n\n".join([doc.page_content for doc in retrieved_docs])
+
+    response = chain.invoke({"question": user_input, "context": context_text})
+
+    # Save to session and persist history
+    st.session_state.message_history.append({"question": user_input, "answer": response})
+    save_history(st.session_state.message_history)
+
+    st.write(response)
+
+# Display chat history
+if st.session_state.message_history:
+    st.write("### Chat History")
+    for msg in st.session_state.message_history:
+        st.write(f"**User:** {msg['question']}")
+        st.write(f"**Bot:** {msg['answer']}")
+
+# Email history feature
+recipient_email = st.text_input("Enter email to receive chat history:")
+if st.button("Send History via Email"):
+    if st.session_state.message_history:
+        history_text = "\n\n".join([f"User: {msg['question']}\nBot: {msg['answer']}" for msg in st.session_state.message_history])
+        success = send_email(recipient_email, "Chat History", history_text)
+        if success:
+            st.success(f"Chat history sent to {recipient_email}!")
+    else:
+        st.warning("Chat history is empty.")

requirements.txt

@@ -6,3 +6,5 @@ langchain-groq
 python-dotenv
 beautifulsoup4
 faiss-cpu
+requests
+smtplib