AGENTIC AI 002 : Agent to Parse Text and provide answers

!pip install langchain langchain-core langchain-community langchain-google-genai faiss-cpu langchain-text-splitters

from langchain_community.vectorstores import FAISS

from langchain_text_splitters import RecursiveCharacterTextSplitter

from langchain_core.documents import Document

from langchain_core.prompts import ChatPromptTemplate

from langchain_core.output_parsers import StrOutputParser

import os

from langchain_google_genai import ChatGoogleGenerativeAI, GoogleGenerativeAIEmbeddings

from google.colab import userdata

# --- Setup ---

os.environ["GOOGLE_API_KEY"] = userdata.get('geminiapikey')

documents = [

    "Our company offers 21 days of paid annual leave. Employees may work from home up to three days a week with manager approval. Health insurance covers the employee, spouse, and two children.",

    "Standard working hours are 9:00–18:00 with a 1-hour break; flexible start between 8:00–10:00 if 8 hours are met.",

    "The company observes 12 public holidays per calendar year, announced each December.",

    "Employees receive 12 paid sick days per year; a doctor’s note is required after 2+ consecutive days.",

    "Parental leave: 16 weeks paid maternity and 4 weeks paid paternity; additional unpaid leave per local law.",

    "Up to 7 unused annual leave days may carry over; cash encashment is not offered.",

    "Overtime must be pre-approved and is compensated at 1.5× or as time off in lieu within 60 days.",

    "Hybrid work: up to 3 WFH days/week with manager approval; availability required during core hours 10:00–16:00.",

    "Company provides a laptop; eligible roles receive a ₹2,000/month home-office stipend.",

    "Expenses must be submitted within 30 days via the expense portal; economy airfare and mid-tier hotels; receipts required for amounts over ₹1,000.",

    "Medical insurance covers employee, spouse, and up to two children; optional parental cover at employee cost.",

    "Payroll runs on the last business day of the month; reimbursements paid with the next payroll.",

    "Notice period is 30 days for voluntary resignation; the company may provide pay in lieu of notice.",

    "The company enforces a zero-tolerance policy for harassment or discrimination; violations may lead to termination.",

    "Employees must follow the IT acceptable-use policy; confidential data cannot be shared outside approved tools."

]

# --- Convert to docs ---

docs = [Document(page_content=d) for d in documents]

# --- Split ---

splitter = RecursiveCharacterTextSplitter(chunk_size=200, chunk_overlap=30)

chunks = splitter.split_documents(docs)

# --- Embeddings ---

#embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")

embeddings = GoogleGenerativeAIEmbeddings(model="gemini-embedding-2-preview")

# --- Vector store ---

vectorstore = FAISS.from_documents(chunks, embeddings)

retriever = vectorstore.as_retriever(search_kwargs={"k": 6})

# --- LLM ---

#llm = ChatGoogleGenerativeAI(model="gemini-1.5-flash", temperature=0)

llm = ChatGoogleGenerativeAI(model="gemini-3-flash-preview", temperature=0)

# --- Prompt ---

prompt = ChatPromptTemplate.from_template("""

You are an HR assistant. Answer ONLY from the context.

Rules:

- Do not use outside knowledge

- If answer is not found, say "Not mentioned"

- Answer in a respectful way

- Answer in complete sentences

Context:

{context}

Question:

{question}

""")

# --- Helper ---

def format_docs(docs):

    return "\n\n".join(doc.page_content for doc in docs)

# --- RAG Chain (LCEL) ---

rag_chain = (

    {

 

      "context": retriever | format_docs,

        "question": lambda x: x

    }

    | prompt

    | llm

    | StrOutputParser()

)

# --- Query ---

response = rag_chain.invoke("What is the sick leave policy?")

print(response)