Відео

Hi, follow the link to my blog in the description. In the description you find the link to the source code at GitHub.

I am happy to read this - thanks for letting me know!

I would like to use copilot like an enhanced code styler. Eg. there are suggestions around too long code, deprecated functions, etc. And copilot can add some other things like “this part of code may result in infinite loop” or “this part may throw error you don’t handle”

@@roronosekai4886 Sounds like a nice feature idea. I hope we will get this some day.

Thanks for sharing this. I did not face this so far, but I am also not using every code Copilot suggests. When I write mappers form one data type to another one I made bugs, which I did not spot. Since Copilot is not just doing copy and paste, it is very strong in the write a mapper task. So this is some kind of opposite example. How works the Copilot chat feature for you?

I have no demo code ready for this. You can look in Spring documentation: docs.spring.io/spring-security/reference/servlet/authentication/logout.html

The number of kernel threads is still the same. But by using virtual threads you can create an "infinite" number of virtual threads, which are carried by the kernel threads. So the difference is, that I could start all virtual threads very early, because there is no limitation.

My demo is about doing a safe refactoring, because the logic is covered by Unit-Tests. In case you have production code and it is covered by Unit-Tests you can give it also to AI and request clean code refactorings. With the existing Unit-Tests you can check, if AI did a proper job. You can also add the Unit-Tests to the chat-context to get better results. So for me it is not about demo or production code, it is about have code under test coverage.

Fair point. I recommend to build the native image in the team's CICD pipeline. In that case you would just code on your windows computer and have no need for Docker.

Steht noch etwas mehr in den Logs der Konsole?

Problem gelöst, der API Key war am Anfang nicht richtig aktiviert, danke schön fürs Video @@elmarbrauch

Did you set proper redirect-uri in application.properties or -.yml file? That could be something to check. On the other hand on your SSO provider there might be also some config required to do the redirect. So I would check first those things.

@@elmarbrauch Thanks for your reply.. I will check them again.

Check this for more information about Spring and serverless : spring.io/serverless/

@@elmarbrauch why you call serverless..

@@47lokeshkumar74 Maybe I did not understand the original question. So I thought that might help.

Do you know also Copilot? There you have AI integrated in your IDE, see ua-cam.com/video/h4wXW8VQGdA/v-deo.html

@@elmarbrauch thanks, these AIs are evolving so fast. Can't wait to see how things will be in 2 years.

Are You using Spring Boot 3.1.X ? If you clone my GitHub project the login page with username and password input should be shown as presented in this video.

@@elmarbrauch Yes, It's working with (.formLogin(Customizer.withDefaults())) !! TNX!! My question is if I used older version of spring security does it influence the level of security of the project ?

I did not try this so far. Maybe in the OpenAI API documentation something is written about that.

It is the better choice. But many existing software has still the classic Spring MVC. Changing this in the mass will take some time. In my latest projects I use WebFlux. Also in older big Spring MVC projects I add it at least, so that I can use at least WebClient.

This can have several causes, because the error handling in my demo code is not very detailed. If you changed some code like switch to different GPT-model, revert this. Otherwise check my blog-article: agile-coding.blogspot.com/2023/01/gpt-3.html. It shows simpler code only about calling the API, this might help you to see, if your API key is correct or not.

@@elmarbrauch Thx, I fixed the problem. I notice that the response of the api is cut (it is not the complete response but it is partical) do you know if I can fix it? Thank you

@@DanChen-m9q This question was asked before. Search for max_tokens value and increase it to get longer answers.

Hi, you have to edit this class: github.com/elmar-brauch/openai-api/blob/master/src/main/java/de/bsi/openai/chatgpt/CompletionRequest.java It has 100 as constant value for max_tokens. If you increase 100 to a bigger value, you will get more often complete answers. (But you will have to pay more, because you consume more tokens)

Ich habe in letzter Zeit weniger deutsche Versionen meiner Videos aufgenommen - irgendwo muss es dann ja hin 😀

Hi. After downloading the Code from GitHub, you have to set the OpenAI API-Key in application.properties file. Then it should work.

@@elmarbrauch thank you so much for your great work. you saved me!!!

Hi. The process is called OpenID Connect, it is a standard. My provider in this video implements the standard and Spring Boot also supports this standard. Check the picture in my video or in the linked blog-article, it should explain the target of the redirect: agile-coding.blogspot.com/2022/10/openid.html

Hi. I wrote 3 articles with 3 videos about Spring Security. This one is the 2nd in the series of 3, so I recommend that you start with my video for beginners. It starts from the scratch based on Spring 3, see this blog article: agile-coding.blogspot.com/2020/12/spring-security-starter.html or the video here: ua-cam.com/video/E6P7V5vvkus/v-deo.html

Hi, you can increase logging level to see more details by adapting application.properties/.yml like this: logging: level.org.springframework: security: DEBUG web.client: DEBUG This might help you to better understand what worked as expected before your problem and maybe you see also more details about the problem beside the stacktraces.

It is not just the model to be changed. It is different URL and different request body. Here is a sample from API reference: curl api.openai.com/v1/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $OPENAI_API_KEY" \ -d '{ "model": "gpt-3.5-turbo", "messages": [{"role": "user", "content": "Hello!"}] }'

Hallo. Ich empfehle "Learning by coding", programmiere so viel wie möglich selbst. Also nicht nur anderen in Udemy Kursen beim Programmieren zuschauen und den Code verstehen, sondern direkt mit programmieren. Codewars würde ich Dir empfehlen, sobald Du in Java Hello World, Schleifen und If-Statements kannst. Ich habe meine Meinung dazu auch in einem Blog-Artikel aufgeschrieben: agile-coding.blogspot.com/2022/05/codewars.html

Sure, you can download my code at GitHub and use it. My code is anyways just a demo.

Das Video gibt es auch in deutscher Sprache mit englischem Akzent in meinem Kanal. 😄

Spring Boot defines the Jackson library. So it is not specified in my code. Open pom.xml in my linked GitHub project and compare it with your Maven or Gradle config.

Thi video is about a stateful Spring Boot backend with a session. If the react, angular or Vue.js frontend is a single page application with a stateless Spring Boot backend it is different. Maybe a topic for a future video.

I optimized the Git project, when I added DALL-E API. This class is now here: github.com/elmar-brauch/openai-api/blob/master/src/main/java/de/bsi/openai/chatgpt/CompletionRequest.java

To get bigger or the full answer, you have to increase the parameter max_tokens, which I set for demo reasons to 100. Open class CompletionRequest to adapt it to your needs.

@@elmarbrauch thank you veru much !

I am happy, that my video was helpful. I do not know yet, if I will make videos about the other image APIs.