I made explainer.js, a CLI tool to processes a file and output the code blocks along with explanation. So something that can be helpful if you have no clue what's going on in a file.
Built using Commander.js and Groq SDK!
Check it out @ https://github.com/aamfahim/explainer.js
Instant Response
For code review my partner @sych_andrii aka Andrii and I used discord call feature. Usually, for code reviews I prefer sync approach. Hopping on a video call and screen sharing and doing it live can help fast track the process as opposed to doing it asynchronously in a git issue. Although asynchronous can also be viable if there is a lot of details but that is rarely the case. So, if possible I prefer to do it synchronously.
Not our first rodeo
Both of us are very familiar with code review. Both of us have had reviewed each other's code. I think the review part was much easier since we knew each other before hand.
I'll do yours first, You do mine after
As the title suggests. My partner reviewed my code first. Since most of my issues were README.md related, he had a lot of questions. And I let him use with my tool. As he wanted, initial issue was he trying to use OpenAI API key, but Groq was throwing errors. So, we tried it with Groq generated API key and it worked. Then we tried OpenAI base URL which was working (technically) but response was indicating the base url provided was not correct. Then we tried different models and discovered we can only use Groq supported models.
These are the issues he filed on my repo
Issues filed on my repo
- https://github.com/aamfahim/explainer.js/issues/9 README.md was generated in a hurry so it had a lot of in accurate information and it was incorrect in some places as well.
- https://github.com/aamfahim/explainer.js/issues/10 Andrii tried to use gpt-4o model but I dont think Groq supports that. Then we tried a supported model which was for content safety which gave him the output on the issue. So, I had to mention the limitations in README.md as well.
- https://github.com/aamfahim/explainer.js/issues/11 The --outout was not implemented yet but it was in the README.md and help description.
- https://github.com/aamfahim/explainer.js/issues/12 -t flag was not being parsed properly
- https://github.com/aamfahim/explainer.js/issues/13 The description was not clear about the expected argument was file path.
WSL or Powershell?
So, after that it was my turn to review. First road block was the installation. Initially I was using my default terminal, WSL with Ubuntu as my distro. And I'll be honest I don't use python a lot and my version was pretty old. And I don't even remember installing python. So because of version I couldn't install dependencies adn I couldn't update my version for some reason. So I switched to cmd and I know I have python installed on my windows laptop. So, I had to use cmd but the issue was that after every command I had to mention --python version. And to do that I had to also pass the file path which was somewhere in the file directory. After getting the setup run it was pretty good. I didn't find any bugs but I did have feedback about the readme structure. Also I requested a .env support since in order to run the tool I had to copy paste the API key every time which was kinda tedious. Also he did it in python. What a madlad!
Issues filed on Andrii's repo
- https://github.com/SychAndrii/infusion/issues/8 I found README.md a bit hard to follow, so I added some recommendations.
- https://github.com/SychAndrii/infusion/issues/9 Requested a .env support so that I don't have to paste the api every time I use it.
- https://github.com/SychAndrii/infusion/issues/10 A specific part of help section needed a bit refactoring.
- https://github.com/SychAndrii/infusion/issues/11 If the file has unrelated content sometimes it would throw error sometimes it would not.
Was it all fixed?
I fixed all my issues. It didn't take a lot of time since the issues were pretty simple. Most to the work was going through commander.js docs. Did some refactoring in between as well.
What did you learn through the process of doing the testing and reviewing?
Lessons learned
Sometimes its easy to get lost in the sauce while developing. I was so focused on making the code more clean I put README.md as a second thought. I think along with the code, documentation should be also given the same importance. Also, using Pure JS in Node with ESM is NOT fun. Learned a lot things about pure JS that I never knew or forgot like, + is a shorthand for converting sting to appropriate data types and arguments is not available in arrow methods.And I wholeheartedly agree with this comment.
以上是Dealing with issues的详细内容。更多信息请关注PHP中文网其他相关文章!

我使用您的日常技术工具构建了功能性的多租户SaaS应用程序(一个Edtech应用程序),您可以做同样的事情。 首先,什么是多租户SaaS应用程序? 多租户SaaS应用程序可让您从唱歌中为多个客户提供服务

本文展示了与许可证确保的后端的前端集成,并使用Next.js构建功能性Edtech SaaS应用程序。 前端获取用户权限以控制UI的可见性并确保API要求遵守角色库

JavaScript是现代Web开发的核心语言,因其多样性和灵活性而广泛应用。1)前端开发:通过DOM操作和现代框架(如React、Vue.js、Angular)构建动态网页和单页面应用。2)服务器端开发:Node.js利用非阻塞I/O模型处理高并发和实时应用。3)移动和桌面应用开发:通过ReactNative和Electron实现跨平台开发,提高开发效率。

JavaScript的最新趋势包括TypeScript的崛起、现代框架和库的流行以及WebAssembly的应用。未来前景涵盖更强大的类型系统、服务器端JavaScript的发展、人工智能和机器学习的扩展以及物联网和边缘计算的潜力。

JavaScript是现代Web开发的基石,它的主要功能包括事件驱动编程、动态内容生成和异步编程。1)事件驱动编程允许网页根据用户操作动态变化。2)动态内容生成使得页面内容可以根据条件调整。3)异步编程确保用户界面不被阻塞。JavaScript广泛应用于网页交互、单页面应用和服务器端开发,极大地提升了用户体验和跨平台开发的灵活性。

Python更适合数据科学和机器学习,JavaScript更适合前端和全栈开发。 1.Python以简洁语法和丰富库生态着称,适用于数据分析和Web开发。 2.JavaScript是前端开发核心,Node.js支持服务器端编程,适用于全栈开发。

JavaScript不需要安装,因为它已内置于现代浏览器中。你只需文本编辑器和浏览器即可开始使用。1)在浏览器环境中,通过标签嵌入HTML文件中运行。2)在Node.js环境中,下载并安装Node.js后,通过命令行运行JavaScript文件。

如何在Quartz中提前发送任务通知在使用Quartz定时器进行任务调度时,任务的执行时间是由cron表达式设定的。现�...


热AI工具

Undresser.AI Undress
人工智能驱动的应用程序,用于创建逼真的裸体照片

AI Clothes Remover
用于从照片中去除衣服的在线人工智能工具。

Undress AI Tool
免费脱衣服图片

Clothoff.io
AI脱衣机

AI Hentai Generator
免费生成ai无尽的。

热门文章

热工具

螳螂BT
Mantis是一个易于部署的基于Web的缺陷跟踪工具,用于帮助产品缺陷跟踪。它需要PHP、MySQL和一个Web服务器。请查看我们的演示和托管服务。

ZendStudio 13.5.1 Mac
功能强大的PHP集成开发环境

SublimeText3汉化版
中文版,非常好用

PhpStorm Mac 版本
最新(2018.2.1 )专业的PHP集成开发工具

SecLists
SecLists是最终安全测试人员的伙伴。它是一个包含各种类型列表的集合,这些列表在安全评估过程中经常使用,都在一个地方。SecLists通过方便地提供安全测试人员可能需要的所有列表,帮助提高安全测试的效率和生产力。列表类型包括用户名、密码、URL、模糊测试有效载荷、敏感数据模式、Web shell等等。测试人员只需将此存储库拉到新的测试机上,他就可以访问到所需的每种类型的列表。