Update app.py

app.py

@@ -1,97 +1,77 @@
-import os, re, validators, streamlit as st
-from youtube_transcript_api import YouTubeTranscriptApi, TranscriptsDisabled, VideoUnavailable
-from langchain.prompts import PromptTemplate
-from langchain.chains.summarize import load_summarize_chain
+# app.py  ─  Streamlit + LangChain + Groq
+import os, asyncio, streamlit as st
+from dotenv import load_dotenv
+from langchain.schema import SystemMessage, HumanMessage, AIMessage
 from langchain_groq import ChatGroq
-from langchain.schema import Document
-from langchain_community.document_loaders import UnstructuredURLLoader
-from langchain.document_loaders import PyPDFLoader
 
-# ─────────────────────────  STREAMLIT CONFIG  ──────────────────────────
-st.set_page_config(page_title="LangChain Summarizer", page_icon="🦜")
-st.title("🦜 LangChain: Summarize YT / Webpage / PDF")
+# ─────────────────────────  bootstrap event‑loop  ─────────────────────
+try:
+    asyncio.get_running_loop()
+except RuntimeError:
+    asyncio.set_event_loop(asyncio.new_event_loop())
+if os.name == "nt":
+    import asyncio as _asyncio
+    asyncio.set_event_loop_policy(_asyncio.WindowsSelectorEventLoopPolicy())
 
-# ────────────────────────────  API KEY INPUT  ──────────────────────────
-with st.sidebar:
-    st.header("API keys")
-    groq_api_key = st.text_input("Groq API Key", type="password")
-    if groq_api_key:
-        os.environ["GROQ_API_KEY"] = groq_api_key          # for libraries
+# ───────────────────────────  UI  /  SETTINGS  ────────────────────────
+st.set_page_config("Groq Chatbot", "🤖")
+st.title("🤖  Groq‑powered Advanced Chatbot")
+st.caption("DeepSeek‑R1‑Distill‑Llama‑70B • LangChain • Streamlit")
 
-# ─────────────────────  PLACEHOLDERS / FILE & URL INPUT  ───────────────
-generic_url   = st.text_input("Paste a YouTube / web URL here:")
-uploaded_file = st.file_uploader("…or upload a PDF", type=["pdf"])
+with st.sidebar:
+    st.header("🔑  Groq API Key")
+    groq_key = st.text_input("Paste your key here", type="password")
+    st.divider()
+    temperature = st.slider("Temperature", 0.0, 1.2, 0.7, 0.1)
+    top_p       = st.slider("Top‑p",       0.0, 1.0, 1.0, 0.05)
+    st.markdown("*All values remain local to your browser.*")
 
-# ──────────────────────────  UTILITY FUNCTIONS  ────────────────────────
-def get_video_id(url: str) -> str | None:
-    m = re.search(r"(?:v=|\/)([0-9A-Za-z_-]{11})", url)
-    return m.group(1) if m else None
+user_q = st.chat_input("Type your message…")
 
-SUMMARY_PROMPT = PromptTemplate(
-    template="Provide a concise summary (~300 words):\n\nContent:\n{text}",
-    input_variables=["text"],
-)
+# ──────────────────────────  LLM (lazy init)  ─────────────────────────
+MODEL_NAME = "deepseek-r1-distill-llama-70b"
 
-def build_llm() -> ChatGroq:
-    """Instantiate ChatGroq once and cache it in session_state."""
+def get_llm():
     if "llm" not in st.session_state:
+        key = groq_key or os.getenv("GROQ_API_KEY")
+        if not key:
+            raise ValueError("Add your Groq key in the sidebar.")
+        os.environ["GROQ_API_KEY"] = key                  # for the client
         st.session_state.llm = ChatGroq(
-            model="llama3-70b-8192",
-            groq_api_key=os.environ["GROQ_API_KEY"],
+            model        = MODEL_NAME,
+            groq_api_key = key,
+            temperature  = temperature,
+            top_p        = top_p,
         )
-    return st.session_state.llm
+    # refresh sampling params if the sliders changed
+    llm = st.session_state.llm
+    llm.temperature = temperature
+    llm.top_p       = top_p
+    return llm
 
-def summarize(docs):
-    llm   = build_llm()
-    chain = load_summarize_chain(llm, chain_type="stuff", prompt=SUMMARY_PROMPT)
-    return chain({"input_documents": docs})["output_text"]
+# ─────────────────────────  conversation memory  ──────────────────────
+if "history" not in st.session_state:
+    st.session_state.history = [
+        SystemMessage(content="You are an advanced, helpful assistant.")
+    ]
 
-# ─────────────────────────────  MAIN ACTION  ───────────────────────────
-if st.button("Summarize"):
-    if not groq_api_key:
-        st.error("Please enter your Groq API key in the sidebar.")
-    elif not generic_url and not uploaded_file:
-        st.error("Provide a URL or upload a PDF, then press Summarize.")
-    else:
-        try:
-            with st.spinner("Fetching and summarizing…"):
+# ────────────────────────────  main loop  ─────────────────────────────
+if user_q:
+    st.session_state.history.append(HumanMessage(content=user_q))
 
-                # ---------- PDF ----------
-                if uploaded_file:
-                    tmp_path = f"/tmp/{uploaded_file.name}"
-                    with open(tmp_path, "wb") as f:
-                        f.write(uploaded_file.read())
-                    docs = PyPDFLoader(tmp_path).load()
-                    st.success(summarize(docs))
-                    os.remove(tmp_path)
+    try:
+        with st.chat_message("assistant", avatar="🤖"):
+            with st.spinner("Thinking…"):
+                answer = get_llm().invoke(st.session_state.history).content
+            st.markdown(answer)
 
-                # ---------- YouTube ----------
-                elif "youtube" in generic_url or "youtu.be" in generic_url:
-                    vid = get_video_id(generic_url)
-                    if not vid:
-                        st.error("Couldn’t extract a YouTube video ID 🤔")
-                    else:
-                        transcript = YouTubeTranscriptApi.get_transcript(vid)
-                        text = " ".join(t["text"] for t in transcript)
-                        st.success(summarize([Document(page_content=text)]))
+        st.session_state.history.append(AIMessage(content=answer))
 
-                # ---------- Plain Webpage ----------
-                else:
-                    if not validators.url(generic_url):
-                        st.error("That doesn’t look like a valid URL.")
-                    else:
-                        docs = UnstructuredURLLoader(
-                            urls=[generic_url],
-                            ssl_verify=False,
-                            headers={
-                                "User-Agent":
-                                "Mozilla/5.0 (X11; Linux) AppleWebKit/537.36 "
-                                "(KHTML, like Gecko) Chrome/121.0 Safari/537.36"
-                            },
-                        ).load()
-                        st.success(summarize(docs))
+    except Exception as err:
+        st.error(f"**Error:** {err}")
 
-        except (TranscriptsDisabled, VideoUnavailable) as yt_err:
-            st.error(str(yt_err))
-        except Exception as e:
-            st.exception(e)
+# ────────────────────────  display chat history  ──────────────────────
+for msg in st.session_state.history[1:]:              # skip system message
+    role = "user" if isinstance(msg, HumanMessage) else "assistant"
+    with st.chat_message(role):
+        st.markdown(msg.content)