Build SmartCart v1 : Setup & API Integration

Business Scenario

Welcome!

Previously, you created model through the GroqCloud API, measured token usage, explored textspliting into tokens, checked context-window limits, tested Temperature, Top-P, and Max Tokens & built a reusable recommender script for different use cases

Today, your manager has assigned you a task to set up LangChain for SmartCart AI and connect it to the Qwen model using GroqCloud. You will install LangChain, store the API key safely, and write a Python script that sends a shopping prompt to the model and displays both the input and the response.

Git Pull

git pull origin branchName

Pre-Lab Preparation

Topic : Build Chatbot V1

1) Orchestration Framework

2) Langchain

3) Prompt Templates

Dataset :

Task 1: Install Langchain

!pip install -q langchain langchain-groq 
from importlib.metadata import version 
print("langchain      :", version("langchain")) 
print("langchain-groq :", version("langchain-groq"))

Task 2: Environment Setup and API Key Authentication

Create the project files

1

%%writefile requirements.txt 
langchain==1.0.3 
langchain-groq==1.0.1 
%%writefile .gitignore 
.env 
__pycache__/

Load the API key into the environment and check it​

2

import os 
from google.colab import userdata 
 
os.environ["GROQ_API_KEY"] = userdata.get("GROQ_API_KEY") 
 
key = os.environ["GROQ_API_KEY"] 
print("Key loaded:", key[:4] + "..." + key[-4:])   # never print the full key 
print("Key length:", len(key))

Use fake key to see authentication failure

3

from langchain_groq import ChatGroq 
bad_llm = ChatGroq(model="qwen/qwen3.8-27b", api_key="gsk_fake_key_123") 
try: 
    bad_llm.invoke("hello") 
except Exception as error: 
    print(type(error).__name__, "-", error)

Output

Task 3 : Make a Single LLM API Call with LangChain

Create a chat model

1

from langchain_groq import ChatGroq 
 
llm = ChatGroq( 
    model="qwen/qwen3.8-27b", 
    temperature=0.7, 
    max_tokens=300, 
    model_kwargs={"top_p": 0.95},   # top_p is passed through model_kwargs 
    reasoning_effort="none",        # thinking off, same as Labs 7 and 8 
) 
 
response = llm.invoke("Recommend 3 gym earbuds under Rs. 2,000, one line each.") 
print(response.content)

Output

Task 4 : Understand the Request Structure

from langchain_core.messages import HumanMessage 
 
request = [ 
    HumanMessage(content="Recommend 3 gym earbuds under Rs. 2,000, one line each.") 
] 
 
for message in request: 
    print("role   :", message.type) 
    print("content:", message.content) 
 
response = llm.invoke(request)

Output

Task 5 : Understand the Response Structure

print("Type          :", type(response).__name__) 
print("Content       :", response.content[:60], "...") 
print("Token usage   :", response.usage_metadata) 
print("Model name    :", response.response_metadata["model_name"]) 
print("Finish reason :", response.response_metadata["finish_reason"])

Output

Task 6 : Control Length with Max Tokens

%%writefile smartcart_v1.py 
import os 
from langchain_groq import ChatGroq 
from langchain_core.messages import HumanMessage

Save the script

1

%%writefile smartcart_v1.py 
import os 
from langchain_groq import ChatGroq 
from langchain_core.messages import HumanMessage 
 
# 1. Authentication: stop early with a clear message if the key is missing 
if not os.environ.get("GROQ_API_KEY"): 
    raise SystemExit("GROQ_API_KEY not found. Add it to Colab Secrets (Lab 9, Task 2).") 
 
# 2. Model: created once, reused for every request 
llm = ChatGroq( 
    model="qwen/qwen3.8-27b", 
    temperature=0.7, 
    max_tokens=300, 
    model_kwargs={"top_p": 0.95}, 
    reasoning_effort="none", 
) 
 
 
def ask_smartcart(shopper_request): 
    """Send one shopping request and return the AIMessage.""" 
    prompt = ( 
        "You are SmartCart's shopping assistant. Recommend 3 products " 
        "for the request below, one line each, with an approximate " 
        f"price in rupees.\nRequest: {shopper_request}" 
    ) 
    request = [HumanMessage(content=prompt)]   # 3. Request structure 
    return llm.invoke(request)                 # 4. Single API call 
 
 
if __name__ == "__main__": 
    shopper_request = input("What are you shopping for? ") 
    response = ask_smartcart(shopper_request) 
 
    
%%writefile smartcart_v1.py 
import os 
from langchain_groq import ChatGroq 
from langchain_core.messages import HumanMessage 
 
# 1. Authentication: stop early with a clear message if the key is missing 
if not os.environ.get("GROQ_API_KEY"): 
    raise SystemExit("GROQ_API_KEY not found. Add it to Colab Secrets (Lab 9, Task 2).") 
 
# 2. Model: created once, reused for every request 
llm = ChatGroq( 
    model="qwen/qwen3.8-27b", 
    temperature=0.7, 
    max_tokens=300, 
    model_kwargs={"top_p": 0.95}, 
    reasoning_effort="none", 
) 
 
 
def ask_smartcart(shopper_request): 
    """Send one shopping request and return the AIMessage.""" 
    prompt = ( 
        "You are SmartCart's shopping assistant. Recommend 3 products " 
        "for the request below, one line each, with an approximate " 
        f"price in rupees.\nRequest: {shopper_request}" 
    ) 
    request = [HumanMessage(content=prompt)]   # 3. Request structure 
    return llm.invoke(request)                 # 4. Single API call 
 
 
if __name__ == "__main__": 
    shopper_request = input("What are you shopping for? ") 
    response = ask_smartcart(shopper_request) 
 
    # 5. Response structure 
    usage = response.usage_metadata 
    print("\n" + response.content) 
    print(f"\n[model: {response.response_metadata['model_name']}" 
          f" | tokens in: {usage['input_tokens']}" 
          f" | out: {usage['output_tokens']}" 
          f" | finish: {response.response_metadata['finish_reason']}]")

Run the script and type shopping request

2

%run smartcart_v1.py

Output

Insert your request in the box : office ke liye formal shirt chahiye under 1500 rupees

Import the script from a normal cell and test it on other prompts.

3

from smartcart_v1 import ask_smartcart 
 
tests = [ 
    "gym earbuds under Rs. 2,000", 
    "smartphone with good camera and fast charging under 5000 rupees", 
    "healthy snacks for office under 500 rupees", 
] 
 
for shopper_request in tests: 
    response = ask_smartcart(shopper_request) 
    usage = response.usage_metadata 
    print(f"{shopper_request[:40]:<40} | out tokens: {usage['output_tokens']:>3}" 
          f" | finish: {response.response_metadata['finish_reason']}")

Output

   Git Push

git push origin branchName

 

Great job!

You have successfully installed LangChain and the GroqCloud connector, secured the API key, managed dependencies in requirements.txt, and explored API authentication errors. You made LLM calls using ChatGroq and invoke(), examined message-based request and AIMessage response, and built smartcart_v1.py as the foundation for SmartCart AI v1.

Checkpoint

Next-Lab Preparation

Topic : Build Chatbot V1 : Baseline

1) Conversational Loop

2) Handling User Input

3) Basic Error Handling & Testing

GenAI - 9

By Content ITV

Made with Slides.com

GenAI - 9

  • 323

More from Content ITV