Shulou(Shulou.com)11/24 Report--

What is the hottest thing in the tech world in 2023? There is no doubt that the answer is AIGC, with the development of this technology, a new way of programming has emerged, that is, using AI to assist programming, with the blessing of AI, developers' efficiency and productivity can be greatly improved. Today we will introduce a great artificial intelligence programming tool Amazon CodeWhisperer, believe me, after using it, you can at least double your work efficiency.

Do you encounter the following problems in your daily coding work?

1. Come into contact with a brand new Python module, do not know how to open hello world

two。 Some method of the module, forgetting the parameters and return values, repeatedly switching manuals will interrupt the train of thought.

3. Don't want to write notes.

4. Wrote a piece of code, but not robust, worried about loopholes that are difficult to find

5.

If you are suffering from these problems, then Amazon CodeWhisperer can solve the above problems very well.

Let's follow in my footsteps and start the practice of CodeWhisperer!

CodeWhisperer initialization

IDE supported by CodeWhisperer includes Visual Studio (VS) Code (IDE used in this blog) and JetBrains IDE (IntelliJ, PyCharm, CLion, GoLand, WebStorm, Rider, PhpStorm, RubyMine and DataGrip). The installation process only takes a few minutes. We will not explain it in detail here. You can refer to the official documentation and video tutorials prepared by the government.

In this blog, we use VSCode to complete a crawler project practice. You can directly retrieve [CodeWhisperer] in the VSCode plug-in and install it directly (there is nearly 200W of installation, so keep up with everyone's pace). After installation, the CodeWhisperer plug-in icon will appear in the VSCode sidebar, as shown below.

The most important thing in the initialization process is the link to the account. Click [Connect to] above. Link, and then follow the steps to log in to the account. After a series of operations, the browser displays the green status prompt screen in the following figure, and the account interfacing has been completed.

When you return to VSCode, you will find that the Amazon CodeWhisperper help manual has been opened. It is recommended to read it, which has sorted out the basic instructions for using the plug-in.

At the same time, the developer tool (DEVELOPER TOOLS) in the lower left corner of VSCode has also displayed a link to Builder ID. At this point, all our pre-work has been completed, and we can start using the tools.

CodeWhisperper usage

Simple logic example

When the CodeWhisperer plug-in is installed, Auto-Suggesions (automatic recommendation mode) will be enabled by default. When we write comments or code, CodeWhisperer will automatically provide us with code hints. This process is like a super version of syntax hints. If you have not installed the plug-in, you can take a look at the dynamic picture first and feel it as a whole.

CodeWhisperer code hint, in VSCode, you can use the left and right arrow keys to select, use Tab to confirm, other keys do not accept the prompt code. The first experience, as described by the plug-in name (Whisperer: whisperer), is like the editor gently telling us about the code to be typed in. If you agree, you can take it at one click, disagree, and just ignore it.

Code translation example

With this feature, many simple programs can be written based on smart prompts, but it's certainly not enough. We need CodeWhisperer to do something more challenging, and in the world of Python crawlers, we often need to reimplement a piece of front-end JS code in Python.

Enter your requirements directly into VSCode, and then CodeWhisperper will give us a complete implementation example, and you can see that the code rewriting is completely correct.

Function example

In addition to translating the code, when writing crawler cases, if you forget the details of the use of certain functions, you can ask CodeWhisperer to give examples directly to quickly awaken your memory.

The following figure shows the sample code for using the input note # to implement the hmac_sha256 function to get the function using Python.

If you press the wrong key while using CodeWhisperer, that is, you do not use Tab confirmation code, you can press the shortcut key at any time and the Alt+C,CodeWhisperer code prompt will appear again, and you can continue to speed up your development efficiency.

Mixed exampl

If all the above logic is defined as single logic, then the following requirement is an example of compound logic. In the actual coding, you will encounter the need to mix the two. The requirements are as follows:

# using Python to implement Base64 + AES string

Enter the above comments in IDE, and then enter, and CodeWhisperer will automatically generate the subsequent code. Some of the logic can be completed at one time, and the results are as follows. In the core logic, CodeWhisperer performs very well, the answers are output in seconds, and are directly available.

Example of algorithm template

In the process of writing business logic, common algorithms are sometimes used, such as fast typesetting, stacking, hashing, etc., many of these algorithms are similar to template code. When writing, if you can generate the code with one click, it can greatly improve the coding speed, including algorithm rewriting speed. In CodeWhisperer, it can be quickly generated through comments.

Using Python to generate Quick Row Code

After experiencing 5 examples, the use of CodeWhisperer is very simple, and there is no need to switch the editing window. The question and answer process of AIGC is completed directly in the code file of IDE. The response speed is very fast and the code accuracy is very high.

Individual cases are no longer difficult for CodeWhisperer, so let's try to complete a complete Python crawler case (let's see if we can write only comments to see if we can complete a qualified crawler collection program).

CodeWhisperer project actual combat

The target collection site is the author's blog, no infringement problem.

Write comments on the crawler basic framework, including the following:

1. Acquisition module used by the program

two。 Address of target collection station

3. Collected target label

4. Data is stored to file

5. Send the collected data to the specified mailbox

Step 1: enter the following comments to get the response content of the web page

#

After writing the comments, enter directly, CodeWhisperer directly generated the corresponding code for us, in the process, you only need to press enter and Tab key.

If you think the above code is a little simple, you can continue to improve it on the basis of the code, such as improving the request header, adding referer and host parameters, and adding comments directly where you need to modify it.

Step 2: write the target data acquisition function

In this step, we use etree in the lxml module to extract the target tag, continue to enter our logic into the comments, and then enter.

# use etree module to extract all article tags in web page responses

# use etree module to extract all article tags in web page responses

Print (article_tags)

At this time, the target blog tag has been initially extracted, and the title and hyperlink address should be extracted again. The comments here need to be written as clearly as possible. If you have some knowledge of Python crawler technology stack, you will find that CodeWhisperer has a speed-up effect on code writing.

Step 3: save the data to the csv file

The logic of writing to a file is a common operation. Enter the function requirements directly. It is believed that CodeWhisperer will be generated directly. The reference notes for input are as follows:

The write_to_csv () function call is placed in the above loop to complete this step. Run the code again, the corresponding file will be generated in the crawler directory, and the file will be opened to get the target data, as shown below.

A simple crawler has been written here, but our work is not finished yet, so we should continue to optimize this program.

Step 4: send the collected data to the mailbox 163

Enter comments: # write an email sending function to send the newly generated file codewhisperper_demo.csv to the specified mailbox, CodeWhisperer will instantly help us generate a reference function, the actual effect is shown in the following figure.

You can send email with one click by changing the account number and password in the above code to your own real data.

When you open your inbox, you can see that the email has been sent successfully. CodeWhisperer provides us with an implementation of email delivery, which you can modify to meet your personalized needs.

Summary of CodeWhisperer usage

At the beginning of the blog, I put forward several headache problems in the coding process, especially the second, programmers always cut out to find code examples and documentation manuals in the coding process, thus interrupting coding ideas, while CodeWhisperer solved the problem perfectly. In the coding process, Amazon CodeWhisperer automatically provides coding suggestions, agrees to use, disagrees to give up, and really realizes immersive programming.

In addition, CodeWhisperer can also generate new business code based on code and comments, especially when we Python engineers come into contact with a new module, he can quickly produce sample code, and the generated code is very similar to our coding style, even coding style and naming rules can be learned perfectly. During the coding process, CodeWhisperer also automatically provides comments for our code, allowing us to devote more energy to the business logic.

Use CodeWhisperer, and after a few minutes, you will feel the improvement in coding efficiency.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.