PromptTemplate basics in LangChain - Step-by-Step Execution

Concept Flow - PromptTemplate basics

Define template string with placeholders

↓

Create PromptTemplate object

↓

Provide values for placeholders

↓

Call format() method

↓

Get final prompt string with values inserted

↓

Use prompt string in language model or output

This flow shows how you define a template with placeholders, create a PromptTemplate, fill in values, and get the final prompt string.

Execution Sample

LangChain

from langchain import PromptTemplate

template = "Hello, {name}! Today is {day}."
prompt = PromptTemplate(template=template, input_variables=["name", "day"])
final_prompt = prompt.format(name="Alice", day="Monday")
print(final_prompt)

This code creates a prompt template with placeholders for name and day, then fills them with values and prints the final prompt.

Execution Table

Step	Action	Input	Output	Notes
1	Define template string	"Hello, {name}! Today is {day}."	Template string stored	Placeholders {name} and {day} noted
2	Create PromptTemplate object	template string, input_variables=["name", "day"]	PromptTemplate instance created	Ready to format with values
3	Call format() with values	name="Alice", day="Monday"	Formatted string generated	Placeholders replaced with values
4	Print final prompt	Formatted string	Hello, Alice! Today is Monday.	Final prompt ready for use
5	End	-	-	Process complete

💡 All placeholders replaced, final prompt string generated and printed

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	Final
template	undefined	"Hello, {name}! Today is {day}."	"Hello, {name}! Today is {day}."	"Hello, {name}! Today is {day}."	"Hello, {name}! Today is {day}."
prompt	undefined	undefined	PromptTemplate object	PromptTemplate object	PromptTemplate object
final_prompt	undefined	undefined	undefined	"Hello, Alice! Today is Monday."	"Hello, Alice! Today is Monday."

Key Moments - 3 Insights

Why do we need to specify input_variables when creating PromptTemplate?

Input_variables tell PromptTemplate which placeholders to expect. Without them, format() won't know what values to replace. See execution_table step 2 where input_variables are set.

What happens if we call format() without providing all required variables?

Format() will raise an error because it can't replace all placeholders. In execution_table step 3, all variables are provided to avoid this error.

Can we reuse the same PromptTemplate object with different values?

Yes, you can call format() multiple times with different inputs to get different prompt strings. The prompt object stays the same as shown in variable_tracker.

Visual Quiz - 3 Questions

Look at the execution_table at step 3. What is the output after calling format()?

A "Hello, Alice! Today is Monday."

B "Hello, {name}! Today is {day}."

C An error because variables are missing

D "Hello, Bob! Today is Tuesday."

💡 Hint

Check the Output column in execution_table row with Step 3

According to variable_tracker, what is the value of 'final_prompt' after step 2?

A PromptTemplate object

B "Hello, Alice! Today is Monday."

C undefined

D "Hello, {name}! Today is {day}."

💡 Hint

Look at the 'final_prompt' row and 'After Step 2' column in variable_tracker

If you forget to include 'day' in input_variables when creating PromptTemplate, what will happen when calling format()?

A Format will succeed and ignore missing placeholders

B Format will raise an error due to missing variable

C PromptTemplate will automatically add missing variables

D The output will be empty string

💡 Hint

Refer to key_moments about input_variables and format() behavior

Concept Snapshot

PromptTemplate basics:
- Define a template string with placeholders like {name}
- Create PromptTemplate(template, input_variables=[...])
- Call format() with values to replace placeholders
- Result is a final prompt string
- Useful for dynamic prompt creation in LangChain

Full Transcript

PromptTemplate basics in LangChain involve creating a template string with placeholders such as {name} and {day}. You then create a PromptTemplate object by passing the template string and a list of input variables that correspond to the placeholders. When you want to generate the final prompt, you call the format() method on the PromptTemplate object, providing values for each placeholder. The format() method returns the prompt string with all placeholders replaced by the provided values. This process allows you to reuse the same template with different inputs easily. It is important to specify all input variables when creating the PromptTemplate to avoid errors during formatting.

Practice

1. What is the main purpose of PromptTemplate in langchain?

easy

A. To create message templates with placeholders for dynamic content

B. To execute machine learning models directly

C. To store data in a database

D. To visualize data in charts

Solution

1. Step 1: Understand the role of PromptTemplate

   PromptTemplate is designed to create text templates that include placeholders for variables.

2. Step 2: Identify its main use

   It helps organize prompts by allowing you to fill in placeholders later, making prompt reuse easier.

3. Final Answer:

   To create message templates with placeholders for dynamic content -> Option A

4. Quick Check:

   PromptTemplate = Templates with placeholders [OK]

Hint: PromptTemplate = template with blanks to fill [OK]

Common Mistakes:

• Thinking PromptTemplate runs models
• Confusing it with data storage
• Assuming it creates visual charts

2. Which of the following is the correct way to create a PromptTemplate with a placeholder named name?

easy

A. PromptTemplate.from_template("Hello, <name>!")

B. PromptTemplate.from_template("Hello, $name!")

C. PromptTemplate.from_template("Hello, {name}!")

D. PromptTemplate.from_template("Hello, %name%!")

Solution

1. Step 1: Recall placeholder syntax in PromptTemplate

   Langchain uses curly braces {} to mark placeholders in templates.

2. Step 2: Match the syntax to the options

   Only PromptTemplate.from_template("Hello, {name}!") uses curly braces correctly: {name}.

3. Final Answer:

   PromptTemplate.from_template("Hello, {name}!") -> Option C

4. Quick Check:

   Placeholders use curly braces {} [OK]

Hint: Placeholders always use curly braces {} [OK]

Common Mistakes:

• Using $ or % instead of {} for placeholders
• Using angle brackets <> which are invalid
• Confusing placeholder syntax with other languages

3. Given the code:

from langchain.prompts import PromptTemplate
prompt = PromptTemplate.from_template("Hello, {name}!")
result = prompt.format(name="Alice")
print(result)

What will be printed?

medium

A. Hello, {name}!

B. Hello, Alice!

C. Hello, name!

D. Error: missing argument

Solution

1. Step 1: Understand the template and format method

   The template has a placeholder {name}. The format method fills it with the value "Alice".

2. Step 2: Determine the output of print(result)

   After formatting, the placeholder is replaced, so the output is "Hello, Alice!".

3. Final Answer:

   Hello, Alice! -> Option B

4. Quick Check:

   format() replaces {name} with "Alice" [OK]

Hint: format() fills placeholders with given values [OK]

Common Mistakes:

• Printing template without formatting
• Confusing placeholder name with literal text
• Expecting an error without missing arguments

4. What is the error in this code snippet?

from langchain.prompts import PromptTemplate
prompt = PromptTemplate.from_template("Hi, {user}!")
result = prompt.format(name="Bob")
print(result)

medium

A. KeyError because 'user' placeholder is not provided

B. No error; output is 'Hi, Bob!'

C. SyntaxError due to wrong placeholder syntax

D. TypeError because format() is missing arguments

Solution

1. Step 1: Check placeholder and format argument names

   The template uses {user} but format() provides name="Bob" which does not match.

2. Step 2: Understand the error caused

   Since {user} is not given a value, a KeyError occurs during formatting.

3. Final Answer:

   KeyError because 'user' placeholder is not provided -> Option A

4. Quick Check:

   Placeholder name must match format() keys [OK]

Hint: Placeholder and format keys must match exactly [OK]

Common Mistakes:

• Assuming format() keys can differ from placeholders
• Thinking it's a syntax error
• Expecting no error when keys mismatch

5. You want to create a PromptTemplate that asks for a user's name and age, then formats a greeting. Which code correctly defines and uses this template?

hard

A. prompt = PromptTemplate.from_template("Hello {name}, you are {age} years old.")
result = prompt.format(user="Eva", years=30)
print(result)

B. prompt = PromptTemplate.from_template("Hello {name}, you are {age} years old.")
result = prompt.format(name="Eva")
print(result)

C. prompt = PromptTemplate.from_template("Hello {name}, you are {age} years old.")
result = prompt.format("Eva", 30)
print(result)

D. prompt = PromptTemplate.from_template("Hello {name}, you are {age} years old.")
result = prompt.format(name="Eva", age=30)
print(result)

Solution

1. Step 1: Check template placeholders

   The template has placeholders {name} and {age} which must be provided as keyword arguments in format().

2. Step 2: Verify format() usage

   prompt = PromptTemplate.from_template("Hello {name}, you are {age} years old.") result = prompt.format(name="Eva", age=30) print(result) correctly passes name="Eva" and age=30 matching placeholders. Others miss arguments or use wrong keys or positional args.

3. Final Answer:

   Correctly defines and uses placeholders with matching keys -> Option D

4. Quick Check:

   Match placeholders and format() keys exactly [OK]

Hint: Pass all placeholders as keyword args in format() [OK]

Common Mistakes:

• Missing required placeholder arguments
• Using wrong argument names in format()
• Passing positional args instead of keywords