ChatGPT – 程式碼編寫

ChatGPT 可以作為多功能助手，幫助開發者完成各種編碼任務，例如生成程式碼片段、修復bug、程式碼最佳化、快速原型設計以及程式碼翻譯等。本章將透過使用 OpenAI API 的 Python 例項如您展示 ChatGPT 如何提升您的編碼體驗。

使用 ChatGPT 自動生成程式碼

我們可以輕鬆地使用 ChatGPT 在任何程式語言中建立程式碼片段。讓我們來看一個例子，我們使用 OpenAI API 生成一個 Python 程式碼片段來檢查給定的數字是否為阿姆斯特朗數：

示例

import openai

# Set your OpenAI API key
openai.api_key = 'your-api-key-goes-here'

# Provide a prompt for code generation
prompt = "Generate Python code to check if the number is an Armstrong number or not."

# Make a request to the OpenAI API for code completion
response = openai.Completion.create(
   engine="gpt-3.5-turbo-instruct",
   prompt=prompt,
   max_tokens=200
)

# Extract and print the generated code from the API response
generated_code = response['choices'][0]['text']
print(generated_code)

輸出

上述程式碼片段將給出以下 Python 程式碼片段，您可以使用它來檢查給定的數字是否為阿姆斯特朗數。

num = int(input("Enter a number: "))
sum = 0
temp = num

while temp > 0:
   digit = temp % 10
   sum += digit ** 3
   temp //= 10

if num == sum:
   print(num, "is an Armstrong number")
else:
   print(num, "is not an Armstrong number")

使用 ChatGPT 修復 Bug

ChatGPT 可以幫助我們識別和修復程式碼中的 bug。它還可以提供改進程式碼，使其免於錯誤的見解。為了更清楚地說明，讓我們來看下面的例子：

import openai

# Set your OpenAI API key
openai.api_key = 'your-api-key-goes-here'

# Example code with a bug
code_with_bug = "for j in range(5): print(i)"

# Provide a prompt to fix the bug in the code
prompt = f"Fix the bug in the following Python code:\n{code_with_bug}"

# Make a request to the OpenAI API for bug fixing
response = openai.Completion.create(
   engine="gpt-3.5-turbo-instruct",
   prompt=prompt,
   max_tokens=150
)

# Extract and print the fixed code from the API response
fixed_code = response['choices'][0]['text']
print(fixed_code)

執行上述程式碼片段後，ChatGPT 將提供以下見解：

The bug in the above code is that the variable used in the loop, "j", is not 
being used at all in the loop body. Instead, the variable "i" is being 
printed, which is not declared or initialized anywhere in the code snippet. 

To fix this bug, we can either replace "i" with "j" in the loop body, or 
change the initial variable in the loop definition from "j" to "i" to ensure 
that the correct variable is being used in the loop. 

Corrected code:

for i in range(5):
   print(i)

使用 ChatGPT 進行程式碼最佳化

我們已經看到 ChatGPT 可以非常快速地生成程式碼片段，並且還可以識別和修復程式碼中的 bug。開發過程的另一個關鍵部分是程式碼最佳化，它可以幫助我們提高任何應用程式的效能和速度。

示例

以下程式碼展示了 ChatGPT 如何幫助我們最佳化程式碼：

import openai

# Set your OpenAI API key
openai.api_key = 'your-api-key-goes-here'

# Example code to optimize
code_to_optimize = "sum_of_squares = sum(i**2 for i in range(1000000))"  

# Provide a prompt to optimize the code
prompt = f"Optimize the following Python code:\n{code_to_optimize}"

# Make a request to the OpenAI API for code optimization
response = openai.Completion.create(
   engine="gpt-3.5-turbo-instruct",
   prompt=prompt,
   max_tokens=200
)

# Extract and print the optimized code from the API response
optimized_code = response['choices'][0]['text']
print(optimized_code)

輸出

以下是 ChatGPT 給出的最佳化上述程式碼的建議：

One possible way to optimize this code is to use the built-in function "sum" 
with a list comprehension instead of a generator expression. 
This can be done as follows:

sum_of_squares = sum([i**2 for i in range(1000000)])

Using list comprehension instead of a generator expression can reduce the 
time complexity of this code from O(n) to O(1). This is because list 
comprehension creates a list of all the squared values beforehand, 
while generator expressions calculate each squared value individually
when needed. This makes the computation for summing the squared 
values much faster.

Another optimization technique could be to use the built-in function "pow" 
instead of the exponential operator "**". This can be done as follows:

sum_of_squares = sum(pow(i,2) for i in range(1000000))

The "pow" function is optimized for calculating powers and may
improve the performance of the code.

使用 ChatGPT 進行快速原型設計

快速原型設計是開發過程中的另一個關鍵步驟。它允許開發人員快速測試和迭代想法。ChatGPT 憑藉其生成程式碼片段的能力，已成為快速原型設計的寶貴工具。

示例

在這個例子中，我們將探討 ChatGPT 如何幫助建立 Python 程式碼片段，用於從 Web API 獲取資料並列印前 3 個結果。

import openai

# Set your OpenAI API key
openai.api_key = 'your-api-key-goes-here'

# Provide a prompt for rapid prototyping
prompt = "Create a Python code snippet to fetch data from a web API and print the first 3 results."

# Make a request to the OpenAI API for code completion
response = openai.Completion.create(
   engine="gpt-3.5-turbo-instruct",
   prompt=prompt,
   max_tokens=250
)

# Extract and print the prototyped code from the API response
prototyped_code = response['choices'][0]['text']
print(prototyped_code)

輸出

讓我們看看 ChatGPT 的響應：

import requests 

# Define the URL of the web API
url = "https://example.com/api"

# Send a GET request and store the response 
response = requests.get(url)

# Convert the JSON response to a Python dictionary 
data = response.json()

# Loop through the first 3 items in the response 
for i in range(3):
   # Print the title and description of each item 
   print("Title:", data["results"][i]["title"])
   print("Description:", data["results"][i]["description"])
# Output: 
# Title: Example Title 1
# Description: This is the first example result.
# Title: Example Title 2
# Description: This is the second example result.
# Title: Example Title 3
# Description: This is the third example result.

使用 ChatGPT 進行程式碼翻譯和遷移

在處理各種專案時，程式碼翻譯和遷移是一個常見的挑戰。ChatGPT 可以透過生成程式碼翻譯來簡化此過程，從而允許開發人員將程式碼片段適應不同的語言或框架。

示例

在這個例子中，我們將探討 ChatGPT 如何幫助將 Python 程式碼片段翻譯成 JavaScript。

import openai

# Set your OpenAI API key
openai.api_key = 'your-api-key-goes-here'

# Example Python code for translation
original_code = "print('Hello, World!')"

# Provide a prompt to translate the code to JavaScript
prompt = f"Translate the following Python code to JavaScript:\n{original_code}"

# Make a request to the OpenAI API for code translation
response = openai.Completion.create(
   engine="gpt-3.5-turbo-instruct",
   prompt=prompt,
   max_tokens=150
)

# Extract and print the translated code from the API response
translated_code = response['choices'][0]['text']
print(translated_code)

輸出

讓我們檢視下面的程式碼翻譯：

console.log('Hello, World!');

結論

本章展示了 ChatGPT 如何幫助您進行編碼。我們學習瞭如何生成程式碼、修復 bug、最佳化程式碼、快速進行程式碼原型設計，甚至在程式語言之間進行程式碼翻譯。

列印頁面