Bug Detector
Ask Nova to identify and fix bugs in Python code.
System Prompt Template
You are an expert {programming language} developer who specializes in debugging and code quality.
Follow these instructions:
{enumerated instructions}
User Prompt Template
## Code ##
{code to be debugged}
Example
Amazon Nova Pro System Prompt
You are an expert Python developer who specializes in debugging and code quality.
Follow these instructions:
1. Read and understand the code provided in ## Code ##
2. List any bugs, potential issues, or code smells in ## Issues ##:
- Actual bugs that will cause errors
- Potential runtime issues
- Style/maintainability concerns
3. Explain the impact of each issue in ## Impact ##
4. Provide the fixed code in ## Fixed Code ## using ```python``` code blocks
5. Add comments explaining the fixes
Amazon Nova Pro User Prompt
## Code ##
def calculate_average(numbers):
total = 0
for num in numbers:
total += num
return total/len(numbers)
def process_data(data_list):
results = []
for item in data_list:
if type(item) in [int, float]:
avg = calculate_average(item) # Bug: passing single number instead of list
results.append(avg)
elif type(item) == list:
if len(item) == 0: # Bug: no handling of empty list
results.append(item)
else:
results.append(calculate_average(item))
return results
# Example usage with bugs
data = [
[1, 2, 3],
4,
[],
[5, 6]
]
print(process_data(data)) # Will raise TypeError
Amazon Nova Pro Sample Response
Response
Issues
- Actual Bug:
-
In
process_data, whenitemis an integer or float,calculate_average(item)is called. This is incorrect becausecalculate_averageexpects a list, not a single number. -
Actual Bug:
-
In
process_data, empty lists are appended to the results without calculating an average, which may not be the intended behavior. -
Potential Runtime Issue:
-
Division by zero is possible in
calculate_averageif an empty list is passed, though this is somewhat handled by the second bug above. -
Style/Maintainability Concern:
-
The type checking using
type(item) in [int, float]andtype(item) == listis not Pythonic. It would be better to useisinstance. -
Style/Maintainability Concern:
- The function
calculate_averagedoes not handle the case where the input list is empty, which should be addressed to avoid division by zero.
Impact
- The first bug causes a
TypeErrorbecausecalculate_averageis called with a single number instead of a list. - The second bug means that empty lists are not handled correctly, leading to potentially incorrect results or unexpected behavior.
- Without proper handling of empty lists, a division by zero error could occur in
calculate_average. - Using
isinstanceinstead oftypechecks makes the code more Pythonic and easier to maintain. - Not handling empty lists in
calculate_averagecan lead to runtime errors and makes the function less robust.
Fixed Code
def calculate_average(numbers):
if not numbers: # Check for empty list
return 0 # or some other appropriate value or behavior
total = 0
for num in numbers:
total += num
return total / len(numbers)
def process_data(data_list):
results = []
for item in data_list:
if isinstance(item, (int, float)): # Use isinstance for type checking
results.append(item) # Single numbers are appended as is
elif isinstance(item, list): # Use isinstance for type checking
avg = calculate_average(item)
results.append(avg)
return results
# Example usage
data = [
[1, 2, 3],
4,
[],
[5, 6]
]
print(process_data(data)) # Should now work correctly
Comments on Fixes
- Added a check for empty lists in
calculate_averageto return 0, avoiding division by zero. - Changed type checks in
process_datato useisinstancefor better practice. - Ensured that single numbers are appended directly to the results list without attempting to calculate an average.
- Handled empty lists by calculating their average as 0 in
calculate_average.
API Request
import boto3
import json
bedrock_runtime = boto3.client(
service_name="bedrock-runtime",
region_name="us-west-2"
)
response = bedrock_runtime.converse(
modelId='us.amazon.nova-pro-v1:0',
system = [
{
"text": """You are an expert Python developer who specializes in debugging and code quality.
Follow these instructions:
1. Read and understand the code provided in ## Code ##
2. List any bugs, potential issues, or code smells in ## Issues ##:
- Actual bugs that will cause errors
- Potential runtime issues
- Style/maintainability concerns
3. Explain the impact of each issue in ## Impact ##
4. Provide the fixed code in ## Fixed Code ## using ```python``` code blocks
5. Add comments explaining the fixes
"""
}
],
messages = [
{
"role": "user",
"content": [
{
"text": """## Code ##
def calculate_average(numbers):
total = 0
for num in numbers:
total += num
return total/len(numbers)
def process_data(data_list):
results = []
for item in data_list:
if type(item) in [int, float]:
avg = calculate_average(item) # Bug: passing single number instead of list
results.append(avg)
elif type(item) == list:
if len(item) == 0: # Bug: no handling of empty list
results.append(item)
else:
results.append(calculate_average(item))
return results
# Example usage with bugs
data = [
[1, 2, 3],
4,
[],
[5, 6]
]
print(process_data(data)) # Will raise TypeError"""
}
]
}
],
inferenceConfig={
"temperature": 0.1,
"topP": .99,
"maxTokens": 1024
}
)
print(json.dumps(response, indent=2))
aws bedrock-runtime converse \
--model-id "us.amazon.nova-pro-v1:0" \
--system '[
{
"text": "You are an expert Python developer who specializes in debugging and code quality.\nFollow these instructions:\n1. Read and understand the code provided in ## Code ##\n2. List any bugs, potential issues, or code smells in ## Issues ##:\n - Actual bugs that will cause errors\n - Potential runtime issues\n - Style/maintainability concerns\n3. Explain the impact of each issue in ## Impact ##\n4. Provide the fixed code in ## Fixed Code ## using ```python``` code blocks\n5. Add comments explaining the fixes"
}
]' \
--messages '[
{
"role": "user",
"content": [
{
"text": "## Code ##\ndef calculate_average(numbers):\n total = 0\n for num in numbers:\n total += num\n return total/len(numbers)\n\ndef process_data(data_list):\n results = []\n for item in data_list:\n if type(item) in [int, float]:\n avg = calculate_average(item) # Bug: passing single number instead of list\n results.append(avg)\n elif type(item) == list:\n if len(item) == 0: # Bug: no handling of empty list\n results.append(item)\n else:\n results.append(calculate_average(item))\n return results\n\n# Example usage with bugs\ndata = [\n [1, 2, 3],\n 4,\n [],\n [5, 6]\n]\nprint(process_data(data)) # Will raise TypeError"
}
]
}
]' \
--inference-config '{
"temperature": 0.1,
"topP": 0.99,
"maxTokens": 512
}' \
--region us-west-2
{
"modelId": "us.amazon.nova-pro-v1:0",
"system": [
{
"text": "You are an expert Python developer who specializes in debugging and code quality.\nFollow these instructions:\n1. Read and understand the code provided in ## Code ##\n2. List any bugs, potential issues, or code smells in ## Issues ##:\n - Actual bugs that will cause errors\n - Potential runtime issues\n - Style/maintainability concerns\n3. Explain the impact of each issue in ## Impact ##\n4. Provide the fixed code in ## Fixed Code ## using ```python``` code blocks\n5. Add comments explaining the fixes"
}
],
"messages": [
{
"role": "user",
"content": [
{
"text": "## Code ##\ndef calculate_average(numbers):\n total = 0\n for num in numbers:\n total += num\n return total/len(numbers)\n\ndef process_data(data_list):\n results = []\n for item in data_list:\n if type(item) in [int, float]:\n avg = calculate_average(item) # Bug: passing single number instead of list\n results.append(avg)\n elif type(item) == list:\n if len(item) == 0: # Bug: no handling of empty list\n results.append(item)\n else:\n results.append(calculate_average(item))\n return results\n\n# Example usage with bugs\ndata = [\n [1, 2, 3],\n 4,\n [],\n [5, 6]\n]\nprint(process_data(data)) # Will raise TypeError"
}
]
}
],
"inferenceConfig": {
"temperature": 0.1,
"topP": 0.99,
"maxTokens": 512
}
}