Code-aware text splitting in LangChain - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ component_behavior

intermediate

What is the output of this LangChain code-aware text splitter?

Given the following Python code using LangChain's CodeTextSplitter, what will be the length of the resulting list chunks after splitting the code_text?

LangChain

from langchain.text_splitter import CodeTextSplitter

code_text = '''def add(a, b):\n    return a + b\n\nprint(add(2, 3))\n'''

splitter = CodeTextSplitter(chunk_size=20, chunk_overlap=0)
chunks = splitter.split_text(code_text)
print(len(chunks))

A 2

B 3

C 1

D 4

💡 Hint

Think about how the splitter respects code boundaries and chunk size.

The CodeTextSplitter splits the code into chunks of max 20 characters without overlap. The given code is split into 3 chunks because it respects code structure and line breaks.

📝 Syntax

intermediate

Which option correctly initializes a CodeTextSplitter with Python language support?

You want to create a CodeTextSplitter that understands Python code syntax. Which of the following code snippets correctly initializes it?

A splitter = CodeTextSplitter(lang='python')

B splitter = CodeTextSplitter(language='py')

C splitter = CodeTextSplitter(language='python')

D splitter = CodeTextSplitter(language='Python')

💡 Hint

Check the exact parameter name and accepted language string.

The correct parameter is 'language' with the value 'python' in lowercase. Other options use incorrect parameter names or values.

🔧 Debug

advanced

Why does this code-aware splitter raise an error?

Consider this code snippet using LangChain's CodeTextSplitter:

from langchain.text_splitter import CodeTextSplitter

splitter = CodeTextSplitter(chunk_size=50)
chunks = splitter.split_text(12345)

Why does this code raise an error?

LangChain

from langchain.text_splitter import CodeTextSplitter

splitter = CodeTextSplitter(chunk_size=50)
chunks = splitter.split_text(12345)

A Because split_text expects a string, but an integer was passed

B Because chunk_size must be a string, not an integer

C Because CodeTextSplitter requires a language parameter

D Because split_text is not a method of CodeTextSplitter

💡 Hint

Check the type of the argument passed to split_text.

The split_text method expects a string input to split. Passing an integer causes a TypeError.

❓ state_output

advanced

What is the content of the first chunk after splitting?

Given this Python code snippet using LangChain's CodeTextSplitter:

code = '''def greet(name):\n    print(f"Hello, {name}!")\n\ngreet('Alice')\n'''
splitter = CodeTextSplitter(chunk_size=30, chunk_overlap=5)
chunks = splitter.split_text(code)
print(chunks[0])

What will be printed as the first chunk?

LangChain

code = '''def greet(name):
    print(f"Hello, {name}!")

greet('Alice')
'''
splitter = CodeTextSplitter(chunk_size=30, chunk_overlap=5)
chunks = splitter.split_text(code)
print(chunks[0])

A def greet(name):\n print(f"Hello, {name}!")\n\n

B def greet(name):\n print(f"Hello, {name}!")\n\ngreet('Alice')

C def greet(name):\n print(f"Hello, {name}!")

D def greet(name):\n print(f"Hello, {name}!")\n\ngreet('Al'

💡 Hint

Remember chunk size and overlap affect how much text is included.

The first chunk includes the function definition and the newline after it, fitting within 30 characters with 5 characters overlap.

🧠 Conceptual

expert

Why use a code-aware text splitter instead of a simple character splitter?

Which of the following best explains the advantage of using a code-aware text splitter like LangChain's CodeTextSplitter over a simple character-based splitter?

A It automatically converts code to pseudocode for easier understanding

B It splits text faster by ignoring whitespace and comments

C It compresses code chunks to reduce size for storage

D It splits code respecting syntax and structure, preventing broken code blocks in chunks

💡 Hint

Think about how code structure affects splitting.

Code-aware splitters understand code syntax and avoid splitting inside code blocks, which keeps chunks meaningful and runnable.