Install Langtang community using pip

Nestle_HR_Assistant_LangChain_Shareable.ipynb

@@ -0,0 +1,274 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "8980ecb3",
+   "metadata": {},
+   "source": [
+    "\n",
+    "# Nestlé HR Assistant (LangChain, Modern) — Shareable Gradio URL\n",
+    "\n",
+    "This notebook sets up a Retrieval-Augmented Generation (RAG) chatbot over your Nestlé HR policy PDF.\n",
+    "\n",
+    "**What you get**\n",
+    "- `%pip install` with correct modules (fixes `ModuleNotFoundError: langchain_community...`).\n",
+    "- PDF loading & chunking with `PyPDFLoader` + `RecursiveCharacterTextSplitter`.\n",
+    "- Embeddings + FAISS vector store using OpenAI (or change model as needed).\n",
+    "- `ChatOpenAI` + `RetrievalQA` chain that answers **only** from the policy.\n",
+    "- Gradio UI with `share=True` → **Public URL** for screenshots.\n",
+    "\n",
+    "> Run cells in order. Last cell prints the Public URL.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "73dc5e11",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# Install or upgrade required libraries\n",
+    "%pip -q install -U langchain langchain-openai langchain-community faiss-cpu pypdf gradio tiktoken\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f1008770",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "import os, glob, shutil, sys\n",
+    "\n",
+    "import langchain\n",
+    "import langchain_openai\n",
+    "import langchain_community\n",
+    "import gradio as gr\n",
+    "\n",
+    "from langchain_community.document_loaders import PyPDFLoader\n",
+    "from langchain.text_splitter import RecursiveCharacterTextSplitter\n",
+    "from langchain_openai import OpenAIEmbeddings, ChatOpenAI\n",
+    "from langchain_community.vectorstores import FAISS\n",
+    "from langchain.prompts import PromptTemplate\n",
+    "from langchain.chains import RetrievalQA\n",
+    "\n",
+    "print(\"Python:\", sys.version)\n",
+    "print(\"langchain:\", langchain.__version__)\n",
+    "print(\"langchain-openai:\", getattr(langchain_openai, \"__version__\", \"n/a\"))\n",
+    "print(\"langchain-community:\", getattr(langchain_community, \"__version__\", \"n/a\"))\n",
+    "print(\"gradio:\", gr.__version__)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e22527ac",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# --- Configuration ---\n",
+    "# IMPORTANT: Set your OpenAI API key here or in the environment before running.\n",
+    "# os.environ[\"OPENAI_API_KEY\"] = \"sk-...\"  # <-- uncomment and paste if needed\n",
+    "\n",
+    "MODEL_NAME = os.getenv(\"OPENAI_MODEL\", \"gpt-4o-mini\")  # you can change to \"gpt-3.5-turbo\"\n",
+    "if not os.getenv(\"OPENAI_API_KEY\"):\n",
+    "    raise SystemExit(\"Missing OPENAI_API_KEY. Set it above or in your environment and re-run.\")\n",
+    "\n",
+    "RETRIEVE_K = 5\n",
+    "CHUNK_SIZE = 900\n",
+    "CHUNK_OVERLAP = 150\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5d82453a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# Find your Nestlé HR policy PDF; prefers your uploaded filename.\n",
+    "candidates = [\n",
+    "    \"1728286846_the_nestle_hr_policy_pdf_2012 (1).pdf\",\n",
+    "    \"the_nestle_hr_policy_pdf_2012.pdf\",\n",
+    "]\n",
+    "\n",
+    "PDF_PATH = next((p for p in candidates if os.path.exists(p)), None)\n",
+    "\n",
+    "# If running where uploads live under /mnt/data, copy it locally\n",
+    "if PDF_PATH is None:\n",
+    "    mnt = \"/mnt/data/1728286846_the_nestle_hr_policy_pdf_2012 (1).pdf\"\n",
+    "    if os.path.exists(mnt):\n",
+    "        try:\n",
+    "            shutil.copy(mnt, os.path.basename(mnt))\n",
+    "            PDF_PATH = os.path.basename(mnt)\n",
+    "        except Exception:\n",
+    "            PDF_PATH = mnt\n",
+    "\n",
+    "# Fallback: glob\n",
+    "if PDF_PATH is None:\n",
+    "    hits = glob.glob(\"*nestle*hr*policy*.pdf\") + glob.glob(\"*HR*Policy*.pdf\")\n",
+    "    PDF_PATH = hits[0] if hits else None\n",
+    "\n",
+    "if PDF_PATH is None or not os.path.exists(PDF_PATH):\n",
+    "    raise SystemExit(\"Policy PDF not found. Put it next to this notebook or under /mnt/data/\")\n",
+    "else:\n",
+    "    print(\"Using PDF:\", PDF_PATH)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2d018f1b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "loader = PyPDFLoader(PDF_PATH)\n",
+    "pages = loader.load()\n",
+    "\n",
+    "splitter = RecursiveCharacterTextSplitter(\n",
+    "    chunk_size=CHUNK_SIZE,\n",
+    "    chunk_overlap=CHUNK_OVERLAP,\n",
+    "    separators=[\"\\n\\n\", \"\\n\", \" \", \"\"],\n",
+    ")\n",
+    "docs = splitter.split_documents(pages)\n",
+    "print(f\"Loaded pages: {len(pages)} | Chunks: {len(docs)}\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f93d6b4c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "emb = OpenAIEmbeddings()\n",
+    "vs = FAISS.from_documents(docs, emb)\n",
+    "retriever = vs.as_retriever(search_kwargs={\"k\": RETRIEVE_K})\n",
+    "print(\"Vector store ready (FAISS).\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9dd28a74",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "system_rules = (\n",
+    "    \"You are an assistant answering questions about the Nestlé HR Policy. \"\n",
+    "    \"Use ONLY the provided context. If the answer is not present, say: \"\n",
+    "    \"'I don’t know based on the provided policy.' Be concise and factual.\"\n",
+    ")\n",
+    "\n",
+    "prompt = PromptTemplate(\n",
+    "    input_variables=[\"context\", \"question\"],\n",
+    "    template=(\n",
+    "        \"{rules}\\n\\n\"\n",
+    "        \"Context:\\n{context}\\n\\n\"\n",
+    "        \"Question: {question}\\n\\n\"\n",
+    "        \"Answer:\"\n",
+    "    ).format(rules=system_rules, context=\"{context}\", question=\"{question}\")\n",
+    ")\n",
+    "\n",
+    "llm = ChatOpenAI(model=MODEL_NAME, temperature=0)\n",
+    "qa = RetrievalQA.from_chain_type(\n",
+    "    llm=llm,\n",
+    "    chain_type=\"stuff\",\n",
+    "    retriever=retriever,\n",
+    "    chain_type_kwargs={\"prompt\": prompt, \"document_variable_name\": \"context\"},\n",
+    "    return_source_documents=True,\n",
+    ")\n",
+    "\n",
+    "def format_sources(source_documents):\n",
+    "    pages = []\n",
+    "    for d in source_documents or []:\n",
+    "        p = d.metadata.get(\"page\", None)\n",
+    "        if isinstance(p, int):\n",
+    "            pages.append(p + 1)  # 1-based\n",
+    "    if not pages:\n",
+    "        return \"\"\n",
+    "    uniq = sorted(set(pages))\n",
+    "    return \"**Source:** Nestlé Human Resources Policy (pp. \" + \", \".join(map(str, uniq)) + \").\"\n",
+    "\n",
+    "def snippets(source_documents, max_chars=280):\n",
+    "    out = []\n",
+    "    for d in (source_documents or [])[:3]:\n",
+    "        txt = (d.page_content or \"\").strip().replace(\"\\n\", \" \")\n",
+    "        out.append(\"• \" + (txt[:max_chars] + (\"…\" if len(txt) > max_chars else \"\")))\n",
+    "    return \"\\n\".join(out)\n",
+    "\n",
+    "def ask(query: str):\n",
+    "    if not query.strip():\n",
+    "        return \"Please enter a question.\", \"\"\n",
+    "    res = qa({\"query\": query})\n",
+    "    ans = (res.get(\"result\") or \"\").strip() or \"I don’t know based on the provided policy.\"\n",
+    "    srcs = res.get(\"source_documents\") or []\n",
+    "    return ans + (\"\\n\\n\" + format_sources(srcs) if srcs else \"\"), (\"**Top snippets:**\\n\" + snippets(srcs)) if srcs else \"\"\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4eaaa3c8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# Quick smoke test\n",
+    "test_q = \"What does the policy say about training and learning?\"\n",
+    "a, s = ask(test_q)\n",
+    "print(\"Q:\", test_q)\n",
+    "print(\"A:\", a[:500], \"...\" if len(a) > 500 else \"\")\n",
+    "print(s)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7cf1a435",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# Gradio UI with share=True to get a Public URL for screenshots\n",
+    "dark_css = (\n",
+    "    \".gradio-container {max-width: 900px !important}\\n\"\n",
+    "    \"body {background: #0b0f14 !important}\\n\"\n",
+    "    \".prose, .gr-markdown, label, .gr-button {color: #e6edf3 !important}\\n\"\n",
+    "    \"textarea, input, .gr-box {background:#0f1620 !important; color:#e6edf3 !important}\\n\"\n",
+    ")\n",
+    "\n",
+    "with gr.Blocks(title=\"Nestlé HR Policy Assistant\", css=dark_css, theme=gr.themes.Soft()) as demo:\n",
+    "    gr.Markdown(\"# Nestlé HR Policy Assistant (LangChain, Modern)\")\n",
+    "    gr.Markdown(\"Ask questions grounded in the HR Policy PDF. Answers include page citations.\")\n",
+    "    q = gr.Textbox(label=\"Your question\", placeholder=\"Type your question and press Enter\")\n",
+    "    ask_btn = gr.Button(\"Ask\")\n",
+    "    ans = gr.Markdown(label=\"Answer\")\n",
+    "    snips = gr.Markdown(label=\"Sources (snippets)\")\n",
+    "    with gr.Row():\n",
+    "        for ex in [\n",
+    "            \"What is Nestlé’s approach to Total Rewards?\",\n",
+    "            \"How does Nestlé handle hiring decisions?\",\n",
+    "            \"What does the policy say about training and learning?\",\n",
+    "            \"How are performance and promotions managed?\",\n",
+    "        ]:\n",
+    "            gr.Button(ex).click(lambda x=ex: x, outputs=q)\n",
+    "    ask_btn.click(ask, inputs=q, outputs=[ans, snips])\n",
+    "    q.submit(ask, inputs=q, outputs=[ans, snips])\n",
+    "\n",
+    "print(\"Launching... The following includes a Public URL you can open for screenshots.\")\n",
+    "demo.launch(share=True, server_name=\"0.0.0.0\")\n"
+   ]
+  }
+ ],
+ "metadata": {},
+ "nbformat": 4,
+ "nbformat_minor": 5
+}