Upload 2 files

app.py

@@ -1,162 +1,305 @@
-#================imports==============
-
-import uuid
-import requests
-import os
-os.environ["USER_AGENT"] = "RAG-App/1.0"
-from typing import Dict, List, Any
-from dotenv import load_dotenv
-
-from bs4 import BeautifulSoup
-from langchain_core.globals import set_llm_cache
-from langchain_core.caches import InMemoryCache
-
-from langchain_community.document_loaders import WebBaseLoader
-from langchain_text_splitters import RecursiveCharacterTextSplitter
-
-from langchain_huggingface import HuggingFaceEmbeddings
-from langchain_community.vectorstores import Weaviate
-from langchain_community.vectorstores import FAISS
-from langchain_groq import ChatGroq
-from langchain_core.prompts import ChatPromptTemplate,MessagesPlaceholder
-
-from langchain_classic.chains.combine_documents import create_stuff_documents_chain
-from langchain_classic.chains import create_retrieval_chain
-
-from langchain_core.runnables.history import RunnableWithMessageHistory
-from langchain_community.chat_message_histories import ChatMessageHistory
-from langchain_core.chat_history import BaseChatMessageHistory
-
-#================== CONFIG==================
-
-load_dotenv()
-
-set_llm_cache(InMemoryCache())
-api_key=os.environ["GROQ_API_KEY"]
-#os.environ["HF_API_KEY"]
-print("api chargée:" if api_key else "y'a probleme!!")
-
-#========== charger et decouper documents=================
-
-urls=[
-    "https://fr.wikipedia.org/wiki/%C3%89levage",
-    "https://fr.wikipedia.org/wiki/La_P%C3%AAche"
-]
-
-loader = WebBaseLoader(urls,
-                      requests_kwargs={
-                          "headers":{
-                              "User-Agent":"RAG-App/1.0"
-                          }
-                      }
-)
-docs =loader.load()
-
-splitter = RecursiveCharacterTextSplitter(chunk_size= 1000, chunk_overlap=200)
-chunks= splitter.split_documents(docs)
-
-#============embeding et indexation vers faiss_db================
-
-embeddings= HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
-
-faiss_db=FAISS.from_documents(
-    documents=chunks,
-    embedding=embeddings
-    )
-
-retriever=faiss_db.as_retriever(search_type="similarity", search_kwargs={"k":3})
-
-#=============== LLM et Prompt=================
-
-llm = ChatGroq(
-    model="llama-3.3-70b-versatile",
-    temperature=0.0,
-    max_tokens=1200
-    )
-
-prompt = ChatPromptTemplate.from_messages([
-    ("system", """Tu es un assistant expert en dans le domaine de l'elevage et la pêche. Réponds clairement.
-Si tu ne connais pas, n'invente pas. Garde un ton amical.
-
-Contexte :
-{context}"""),
-    MessagesPlaceholder(variable_name="chat_history"),
-    ("human", "{input}"),
-])
-
-#============= CHAINE DE RECUPERATION=======
-
-stuff_chain= create_stuff_documents_chain(llm, prompt)
-rag_chain=create_retrieval_chain(retriever, stuff_chain)
-
-import gradio as gr
-
-store = {}
-
-def get_session_history(session_id:str)->BaseChatMessageHistory:
-    if session_id not in store:
-        store[session_id] = ChatMessageHistory()
-    return store[session_id]
-
-# ====== CHAIN AVEC MÉMOIRE ======
-
-convers_chain = RunnableWithMessageHistory(
-    rag_chain,
-    get_session_history,
-    input_messages_key="input",
-    history_messages_key="chat_history",
-    output_messages_key="answer"
-)
-
-# =============FONCTION CHAT ================
-
-SESSION_ID = str(uuid.uuid4())  # session globale
-
-def chat_fn(message, history):
-    result = convers_chain.invoke(
-        {"input": message},
-        config={"configurable": {"session_id": SESSION_ID}}
-    )
-    return result.get("answer", str(result))
-
-# def chat_fn(message, history, request: gr.Request):
-#     session_id = request.session_hash
-    
-#     result = convers_chain.invoke(
-#         {"input": message},
-#         config={"configurable": {"session_id": session_id}}
-#     )
-#     return result.get("answer", str(result))
-
-#session_id_state = gr.State(value=None)
-
-# def chat_fn(message, history, session_id_state):
-#     if session_id_state is None:
-#         session_id_state = str(uuid.uuid4())
-
-#     result = convers_chain.invoke(
-#         {"input": message},
-#         config={"configurable": {"session_id": session_id_state}}
-#     )
-
-#     response = result.get("answer", str(result))
-#     return response, session_id_state
-
-
-# ================= GRADIO ====================
-
-demo = gr.ChatInterface(
-    fn=chat_fn,
-    title="🤖 RAG:Specialist en Science Animale 👌",
-    description="Posez vos questions sur l'élévage et la pêche",
-    examples=[
-        "C'est quoi la pêche ?",
-        "Explique l'élévage",
-        "Quelle est la différence entre l'élévage et pêche ?"
-    ]
-)
-
-
-# ===================LANCEMENT ================
-
-demo.launch()
+#================imports==============
+
+import uuid
+import requests
+import os
+os.environ["USER_AGENT"] = "RAG-App/1.0"
+from typing import Dict, List, Any
+from dotenv import load_dotenv
+
+from bs4 import BeautifulSoup
+from langchain_core.globals import set_llm_cache
+from langchain_core.caches import InMemoryCache
+
+from langchain_community.document_loaders import WebBaseLoader
+from langchain_text_splitters import RecursiveCharacterTextSplitter
+
+from langchain_huggingface import HuggingFaceEmbeddings
+from langchain_community.vectorstores import FAISS
+from langchain_groq import ChatGroq
+from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
+
+from langchain.chains.combine_documents import create_stuff_documents_chain
+from langchain.chains import create_retrieval_chain
+
+from langchain_core.runnables.history import RunnableWithMessageHistory
+from langchain_community.chat_message_histories import ChatMessageHistory
+from langchain_core.chat_history import BaseChatMessageHistory
+
+import gradio as gr
+
+#================== CONFIG==================
+
+load_dotenv()
+
+set_llm_cache(InMemoryCache())
+api_key = os.environ.get("GROQ_API_KEY")
+if not api_key:
+    raise ValueError("❌ GROQ_API_KEY non trouvée!")
+print("✅ API chargée avec succès")
+
+#========== charger et découper documents=================
+
+print("📥 Chargement des documents...")
+urls = [
+    "https://fr.wikipedia.org/wiki/%C3%89levage",
+    "https://fr.wikipedia.org/wiki/La_P%C3%AAche"
+]
+
+try:
+    loader = WebBaseLoader(
+        urls,
+        requests_kwargs={"headers": {"User-Agent": "RAG-App/1.0"}}
+    )
+    docs = loader.load()
+    print(f"✅ {len(docs)} documents chargés")
+except Exception as e:
+    print(f"⚠️ Erreur de chargement: {e}")
+    from langchain_core.documents import Document
+    docs = [
+        Document(page_content="L'élevage est l'ensemble des activités qui assurent la multiplication et l'entretien des animaux domestiques pour la production de biens et services."),
+        Document(page_content="La pêche est l'activité consistant à capturer des animaux aquatiques dans leur milieu naturel.")
+    ]
+
+splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
+chunks = splitter.split_documents(docs)
+print(f"✅ {len(chunks)} segments créés")
+
+#============embedding et indexation================
+
+print("🔧 Création des embeddings...")
+embeddings = HuggingFaceEmbeddings(
+    model_name="sentence-transformers/all-MiniLM-L6-v2",
+    model_kwargs={'device': 'cpu'}
+)
+
+faiss_db = FAISS.from_documents(documents=chunks, embedding=embeddings)
+print("✅ Base FAISS créée")
+
+retriever = faiss_db.as_retriever(search_type="similarity", search_kwargs={"k": 3})
+
+#=============== LLM et Prompt=================
+
+print("🤖 Initialisation du LLM...")
+llm = ChatGroq(
+    model="llama-3.3-70b-versatile",
+    temperature=0.0,
+    max_tokens=1200
+)
+
+prompt = ChatPromptTemplate.from_messages([
+    ("system", """Tu es un assistant expert en élevage et pêche. 
+Réponds de manière claire et concise en français.
+Si tu ne connais pas la réponse, dis-le honnêtement.
+Contexte : {context}"""),
+    MessagesPlaceholder(variable_name="chat_history"),
+    ("human", "{input}"),
+])
+
+#============= CHAINE DE RÉCUPÉRATION================
+
+stuff_chain = create_stuff_documents_chain(llm, prompt)
+rag_chain = create_retrieval_chain(retriever, stuff_chain)
+
+# ====== GESTION DE L'HISTORIQUE ======
+
+store = {}
+
+def get_session_history(session_id: str) -> BaseChatMessageHistory:
+    if session_id not in store:
+        store[session_id] = ChatMessageHistory()
+    return store[session_id]
+
+convers_chain = RunnableWithMessageHistory(
+    rag_chain,
+    get_session_history,
+    input_messages_key="input",
+    history_messages_key="chat_history",
+    output_messages_key="answer"
+)
+
+# ================= CSS POUR LE STYLE ====================
+
+custom_css = """
+.sidebar {
+    background: #202123 !important;
+    min-height: 100vh;
+}
+.main-area {
+    background: #343541 !important;
+}
+.chatbot-container {
+    height: calc(100vh - 200px) !important;
+}
+"""
+
+# ================= INTERFACE GRADIO (Gradio 6.0 compatible) ====================
+
+with gr.Blocks() as demo:
+    with gr.Row(equal_height=True):
+        # Colonne gauche : Historique
+        with gr.Column(scale=1, min_width=250, elem_classes="sidebar"):
+            gr.Markdown("## 📚 Historique")
+            new_chat_btn = gr.Button("➕ Nouvelle conversation", variant="secondary")
+            history_radio = gr.Radio(
+                choices=[],
+                label="Conversations",
+                interactive=True
+            )
+            clear_btn = gr.Button("🗑️ Effacer", variant="stop", size="sm")
+        
+        # Colonne droite : Chat
+        with gr.Column(scale=3, elem_classes="main-area"):
+            gr.Markdown("# 🤖 Assistant Élevage & Pêche")
+            
+            chatbot = gr.Chatbot(
+                label="",
+                height=500,
+                show_label=False,
+                avatar_images=(None, "🐟")
+            )
+            
+            with gr.Row():
+                msg_input = gr.Textbox(
+                    placeholder="Posez votre question sur l'élevage ou la pêche...",
+                    show_label=False,
+                    scale=9,
+                    container=False
+                )
+                send_btn = gr.Button("📤", variant="primary", scale=1)
+            
+            gr.Examples(
+                examples=["C'est quoi l'élevage ?", "Explique la pêche", "Différence entre élevage et pêche ?"],
+                inputs=msg_input
+            )
+    
+    # États
+    conversations_state = gr.State([])
+    current_session_id = gr.State(str(uuid.uuid4()))
+    
+    # ================= FONCTIONS =================
+    
+    def create_new_chat(conversations):
+        """Nouvelle conversation"""
+        new_id = str(uuid.uuid4())
+        conversations.append({"id": new_id, "title": "Nouveau chat", "messages": []})
+        choices = [c["title"] for c in conversations]
+        return conversations, new_id, [], gr.update(choices=choices, value=None)
+    
+    def load_chat(selected_title, conversations):
+        """Charger une conversation"""
+        if not selected_title or not conversations:
+            return [], ""
+        
+        for conv in conversations:
+            if conv["title"] == selected_title:
+                history = []
+                user_msg = None
+                for msg in conv["messages"]:
+                    if msg["role"] == "user":
+                        user_msg = msg["content"]
+                    else:
+                        if user_msg:
+                            history.append([user_msg, msg["content"]])
+                            user_msg = None
+                return history, conv["id"]
+        return [], ""
+    
+    def send_message(message, chat_history, conversations, session_id):
+        """Envoyer un message"""
+        if not message or not message.strip():
+            return "", chat_history, conversations, session_id, gr.update()
+        
+        # Gérer la session
+        if not session_id:
+            session_id = str(uuid.uuid4())
+        
+        # Mettre à jour la conversation dans l'historique
+        conv_exists = False
+        for conv in conversations:
+            if conv["id"] == session_id:
+                conv_exists = True
+                conv["messages"].append({"role": "user", "content": message})
+                if conv["title"] == "Nouveau chat":
+                    conv["title"] = message[:40] + "..."
+                break
+        
+        if not conv_exists:
+            conversations.append({
+                "id": session_id,
+                "title": message[:40] + "...",
+                "messages": [{"role": "user", "content": message}]
+            })
+        
+        # Ajouter le message au chatbot
+        chat_history.append([message, None])
+        
+        try:
+            result = convers_chain.invoke(
+                {"input": message},
+                config={"configurable": {"session_id": session_id}}
+            )
+            response = result.get("answer", "Désolé, je n'ai pas compris.")
+        except Exception as e:
+            response = f"❌ Erreur: {str(e)}"
+        
+        # Sauvegarder la réponse
+        for conv in conversations:
+            if conv["id"] == session_id:
+                conv["messages"].append({"role": "assistant", "content": response})
+                break
+        
+        # Mettre à jour le chatbot
+        chat_history[-1] = [message, response]
+        
+        # Mettre à jour la liste
+        choices = [c["title"] for c in conversations]
+        
+        return "", chat_history, conversations, session_id, gr.update(choices=choices, value=conversations[-1]["title"] if conversations else None)
+    
+    def clear_all():
+        """Tout effacer"""
+        store.clear()
+        return [], [], [], gr.update(choices=[])
+    
+    # ================= ÉVÉNEMENTS =================
+    
+    msg_input.submit(
+        send_message,
+        inputs=[msg_input, chatbot, conversations_state, current_session_id],
+        outputs=[msg_input, chatbot, conversations_state, current_session_id, history_radio]
+    )
+    
+    send_btn.click(
+        send_message,
+        inputs=[msg_input, chatbot, conversations_state, current_session_id],
+        outputs=[msg_input, chatbot, conversations_state, current_session_id, history_radio]
+    )
+    
+    new_chat_btn.click(
+        create_new_chat,
+        inputs=[conversations_state],
+        outputs=[conversations_state, current_session_id, chatbot, history_radio]
+    )
+    
+    history_radio.change(
+        load_chat,
+        inputs=[history_radio, conversations_state],
+        outputs=[chatbot, current_session_id]
+    )
+    
+    clear_btn.click(
+        clear_all,
+        outputs=[chatbot, conversations_state, current_session_id, history_radio]
+    )
+
+# ================= LANCEMENT ====================
+
+if __name__ == "__main__":
+    print("🚀 Lancement de l'application...")
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        css=custom_css,
+        theme="soft"
+    )

requirements.txt

@@ -1,14 +1,16 @@
-#====== installer=====
-
- langchain
- langchain-huggingface
- langchain-core
- langchain-classic
- langchain-community
- langchain-groq
- sentence-transformers
- python-dotenv
- faiss-cpu
- beautifulsoup4
- gradio>=5.0.0
- torch
+langchain>=0.3.0
+langchain-core>=0.3.0
+langchain-community>=0.3.0
+langchain-groq>=0.2.0
+langchain-huggingface>=0.1.0
+langchain-text-splitters>=0.3.0
+sentence-transformers>=3.0.0
+huggingface-hub>=0.20.0
+torch>=2.0.0
+faiss-cpu>=1.8.0
+beautifulsoup4>=4.12.0
+lxml>=5.0.0
+gradio>=5.0.0
+python-dotenv>=1.0.0
+requests>=2.31.0
+numpy>=1.24.0