Exercise to practice Prompt Templates and Chains (LC<0.3)

exercise_langchain02_intro_chains.md

Practice: Building Chains

You've learned how to combine a PromptTemplate and an LLM into a chain using the | operator, then run it with .invoke().

Iteration 1 — Environment Setup

• Create a notebook (e.g. locally on VS Code, or on Google Colab).
• Use langchain-core<0.3
• You'll also need an OpenAI API key

Iteration 2 — Single-input chain

Create a chain that generates a short joke about an animal.

1. Build a PromptTemplate with one input variable, animal, and a template like "Tell me a short joke about a {animal}".
2. Create the chain using prompt | llm.
3. Invoke it with animal="penguin" and print the result.

Iteration 3 — Multi-input chain

Now create a chain that recommends a recipe based on two inputs: ingredient and cuisine.

1. Build a PromptTemplate with two input variables, ingredient and cuisine.
2. Create the chain and invoke it with values of your choice.
3. Print the result.

Bonus iterations

Bonus 1: Add a third input variable, diet (e.g. "vegetarian", "gluten-free"), to the recipe chain, and update the prompt to take it into account.

Bonus 2: Write a small Python function run_recipe_chain(ingredient, cuisine, diet) that builds the prompt, creates the chain, invokes it, and returns the result — so you can call it multiple times with different arguments without repeating code.

Solution — Iteration 1

prompt = PromptTemplate(
    input_variables=["animal"],
    template="Tell me a short joke about a {animal}"
)

chain = prompt | llm

result = chain.invoke({"animal": "penguin"})
print(result)

Solution — Iteration 2

prompt = PromptTemplate(
    input_variables=["ingredient", "cuisine"],
    template="Suggest a {cuisine} recipe that uses {ingredient} as a main ingredient."
)

chain = prompt | llm

result = chain.invoke({"ingredient": "chicken", "cuisine": "Thai"})
print(result)

Solution — Bonus 1

prompt = PromptTemplate(
    input_variables=["ingredient", "cuisine", "diet"],
    template="Suggest a {cuisine} recipe that uses {ingredient} as a main ingredient. It must be {diet}."
)

chain = prompt | llm

result = chain.invoke({"ingredient": "mushrooms", "cuisine": "Italian", "diet": "vegetarian"})
print(result)

Solution — Bonus 2

def run_recipe_chain(ingredient, cuisine, diet):
    prompt = PromptTemplate(
        input_variables=["ingredient", "cuisine", "diet"],
        template="Suggest a {cuisine} recipe that uses {ingredient} as a main ingredient. It must be {diet}."
    )
    chain = prompt | llm
    return chain.invoke({"ingredient": ingredient, "cuisine": cuisine, "diet": diet})

print(run_recipe_chain("tofu", "Japanese", "vegan"))
print(run_recipe_chain("lentils", "Indian", "gluten-free"))