Debugging with AI
Practical Exercise 4
Hands-On
Use AI to identify, understand, and fix bugs efficiently
Practical Objectives
1
Use AI to analyze error messages and stack traces
2
Apply chain-of-thought debugging prompts
3
Debug logic errors with AI assistance
4
Write tests to prevent regression
Task 1: Analyze This Buggy Code
This shopping cart has bugs. Find and fix them:
class ShoppingCart:
def __init__(self):
self.items = []
def add_item(self, name, price, quantity):
self.items.append({"name": name, "price": price, "qty": quantity})
def remove_item(self, name):
for item in self.items:
if item["name"] == name:
self.items.remove(item) # Bug 1: Modifying list while iterating
def calculate_total(self):
total = 0
for item in self.items:
total += item["price"] * item["quantity"] # Bug 2: Wrong key name
return total
def apply_discount(self, percent):
self.total = self.calculate_total()
return self.total - (self.total * percent) # Bug 3: percent should be /100
Task 2: Use AI to Debug
Use this chain-of-thought debugging prompt:
Debugging Prompt TemplateI have a bug in my code. Help me debug it step by step:
1. Code:
[paste the buggy code]
2. Expected behavior:
Shopping cart should correctly add/remove items and calculate totals with discounts.
3. Actual behavior:
- remove_item doesn't remove all matching items
- calculate_total throws KeyError
- apply_discount gives wrong values
4. Debug step by step:
- What is each function supposed to do?
- Trace through with sample data
- Identify where actual differs from expected
- Propose fixes with explanations
Task 3: Stack Trace Analysis
Ask AI to analyze this error:
Traceback (most recent call last):
File "cart.py", line 42, in <module>
total = cart.calculate_total()
File "cart.py", line 15, in calculate_total
total += item["price"] * item["quantity"]
KeyError: 'quantity'
Error Analysis PromptAnalyze this Python error:
[paste stack trace]
1. What does this error mean?
2. What line caused it?
3. What's the root cause?
4. How do I fix it?
5. How do I prevent similar errors?
Task 4: Debug Logic Errors
This function has a logic bug - no error, but wrong output:
def find_duplicates(lst):
"""Return list of duplicate values."""
seen = []
duplicates = []
for item in lst:
if item in seen:
duplicates.append(item) # Bug: adds multiple times
seen.append(item)
return duplicates
# find_duplicates([1, 2, 2, 2, 3]) returns [2, 2] instead of [2]
Debugging Strategy
Ask AI: "Trace through find_duplicates([1, 2, 2, 2, 3]) step by step. Show the state of 'seen' and 'duplicates' after each iteration."
Task 5: Write Regression Tests
After fixing bugs, write tests to prevent regression:
Test Generation PromptGenerate pytest tests for the ShoppingCart class:
Test cases needed:
1. add_item adds items correctly
2. remove_item removes only the specified item
3. remove_item handles item not found
4. calculate_total returns correct sum
5. apply_discount with 10% reduces total by 10%
6. apply_discount with 0% returns original total
7. Empty cart has total of 0
Use descriptive test names and assertions.
Task 6: Async/Race Condition Bug
Debug this async code with AI:
async def fetch_user_data(user_ids):
results = []
for user_id in user_ids:
data = await fetch_user(user_id)
results.append(data) # Potential: Order not guaranteed with concurrent calls
return results
async def update_balance(user_id, amount):
balance = await get_balance(user_id)
new_balance = balance + amount # Bug: Race condition between read and write
await set_balance(user_id, new_balance)
Ask AI: "What race conditions exist in update_balance? How would concurrent calls with user_id=1 cause issues?"
Deliverables Checklist
Bug Fixes
- Fixed remove_item iteration bug
- Fixed KeyError in calculate_total
- Fixed discount calculation
Understanding
- Can explain each bug's root cause
- Used chain-of-thought debugging
- Traced execution step by step
Tests
- Tests for all fixed bugs
- Edge cases covered
- All tests passing
Advanced
- Identified race condition
- Proposed fix with locking
- Documented the solution
Debugging Complete!
You've mastered AI-assisted debugging
Next: Practical 5 - AI-Assisted Code Review
