itamarcode

CodiumAI [0], helps me generate meaningful test suites and code suggestions. This is exactly what I need to enable me to code fast with confidence. Really, just try it, and let me know if you are not positively surprised. [0] https://codium.ai/

In a short time we saw real use by developers and companies, and we received great feedback (and 750 stars). We improved, added, and... decided to launch in Product Hunt today (25th July)

We will be happy if you try and give feedback ⬆ Thank you

https://www.producthunt.com/posts/pr-agent-by-codiumai

It isn't meant to replace a human reviewer. It is meant to help making the review process better (and faster).

E.g., PR-Agent makes it easy to have a proper PR description for any PR. It makes it easy get an initial set of clarification questions and a quick review (from the AI) before a human review. Etc..

I think that "/improve" can improve if in addition the developer can give a hint or request what to improve. Would be great to see such contribution to the open source

co-creator here, happy to answer questions.

No matter if you are using GitHub, GitLab, or Bitbucket, you can use this open-source AI-empowered tool to

/describe --> generate a description for the PR. how many times you were lazy to write a description. no more

/review --> generate a general review for the PR

/ask --> ask anything about the PR

/improve --> generate inline code suggestions

/reflect --> ask the PR-Agent to ask you tough questions, and then accordingly create an analysis

we aim to add more capabilities, including more complicated reasoning. we would love to get contributions from the community, you can see we respond quickly to new issues and PRs.

also, please take a look at the roadmap and make suggestions

I think it is a good usage of AI. It saves time, and even if there is a mistake (and there are) or something missing, it is easy to fix

the specific PR-Agent repo: https://github.com/Codium-ai/pr-agent

"Making pull requests less painful with an AI agent"

includes various AI-empowered functions: /review /describe /improve /ask /reflect

Is the open-source standalone or requires using other services? E.g. do I need to use tutim.io in any way?

Can I, for example, write the code components headers/structure/definitions, without detailed implementation, and still generate useful tests?

Good points. Developers spend a lot of time maintaining and changing code. Practically speaking, even a large proportion of "new" code is legacy code.

I wonder, AI (/CodiumAI) may "dramatically reduce the time it takes to capture the existing behavior of the code and start setting up a safety net so you can work with it" -- But can it help me create tests for newly written behavior and functionality?

"dolly-v2-12b is not a state-of-the-art generative language model and, though quantitative benchmarking is ongoing, is not designed to perform competitively with more modern model architectures or models subject to larger pretraining corpuses." from: https://huggingface.co/databricks/dolly-v2-12b

Shimon, are there any risks in activating Datree on the prod env?

Does an AI-empowered code level testing tool really require its own domain specific model?

It allows you to generate a test suite inside the IDE. And then interactively guide it to generate more or adjust tests.

Links: VS Code: https://marketplace.visualstudio.com/items?itemName=Codium.c... JetBrains: https://plugins.jetbrains.com/plugin/21206-codiumai--meaning...

While in GitHub Copilot X you need to join a waiting list, you can already today generate meaningful test suites right inside your IDE. Check out: https://www.codium.ai

https://plugins.jetbrains.com/plugin/21206-codiumai--meaning...

https://marketplace.visualstudio.com/items?itemName=Codium.c...

You can also consider using CodiumAI to create tests right inside your IDE. https://www.codium.ai

https://plugins.jetbrains.com/plugin/21206-codiumai--meaning...

https://marketplace.visualstudio.com/items?itemName=Codium.c...

Here is the official announcement by the CEO (me :) https://www.codium.ai/blog/codiumai-powered-by-testgpt-accou...

Available on VS Code and JetBrains marketplace, for Python, JS & TS:

https://marketplace.visualstudio.com/items?itemName=Codium.c... https://plugins.jetbrains.com/plugin/21206-codiumai--meaning...

Thank you for the shoutout alexrustic

Limited compared to what? Anyway, the mentioned technologies are definitely under heavy research and will advance LLMs capabilities meaningfully. Bigger context and Chain-of-Thoughts are the leading ones in my opinion

This is cool. This technique should work nicely quite often. Although, it will also generate many false-positives alerts.

At Codium.ai, we are trying to tackle this problem. We are developing a new code integrity product that intends to do something quite similar. Codium will mark problematic parts of the code (a.k.a bugs), via auto-generated tests, that isn't inline with the developer's intent. The tricky part is to have high accuracy. We don't want to annoy any folks with false positives.

We would love to get your feedback about what we are working on! We are developers with ML background, excited about exploiting ML for software development, so developers can code fast with confidence.