News for hackers. Stuff that matters.

aditikothari

Wow thats an interesting use case. Check out the readme - it explains how you could build your custom agents.
This is amazing!!
Thank you!
Absolutely, debugging is one of the things that can benefit very well from the knowledge graph. Our debugging agent has the ability to look up a function's "neighbours" i.e. functions that call it and function that it calls. This way it will be able to do more comprehensive debugging much better
Hey thats great! Which agent did you try?

Show HN: AI Agents for engineering use cases like debugging, LLD,testing etc.

https://github.com/potpie-ai/potpie

Hi I am Aditi and I co-founded Potpie AI with my college mate Dhiren. We are building an open-source infrastructure to create custom agents for engineering use-cases like debugging, system design, integration testing, PR review etc. The agents are powered by a knowledge graph built on your code base to provide better context and memory, leading to better planning and execution. Currently we...

( Read More... | 14 comments | 19 points | on HN )
This is part of our roadmap for August.
Hey there!

You can use momentum's open-source auditor to generate integration tests and visualise your code at every git push or pull request. The test plan, test data setup, and test code generation followed by debugging is completely done by momentum in an automated way.

We decided to use visualization for test generation because it helps identify interdependencies at a glance and proactively fix anything that could go wrong.

hey there!

We are launching momentum.sh a behavioral code auditor that analysesthe necessary code behavior and tests it at every git push to ensure the code is ready for production. It does this by performing these steps

Blast radius detection - this will be a list of endpoints that could potentially be affected by the changes you made in your code. This will also be a starting point to decide what parts of your code need to be tested before shipping to production.

Dependency Visualization: Visualize code dependencies and relationships.

Behaviour identification: Automatically identify behaviors written in your code and generate a plan to test their functionality

Code generation to test functionality: Generate code to test all behaviors identified and run it in your local environment through momentum

Run code in local dev environment: No need for yet another yaml, our cli works with your existing dev environment to run code.

Debug code: Based on the stack trace of the run, momentum can diagnose and propose a solution.

We currently support python but will add Java support by the end of July and Javascript in the following month.

Any feedback is highly appreciated! Thanks in advance for giving momentum a try!

Re:Is Apache AGE a better alternative to neo4j on Wednesday July 03, 2024 @07:27PM (#40871037)
Got it. In your experience is maintaining AGE straightforward or not given that it's Postgres extension and not a managed db?

Is Apache AGE a better alternative to neo4j

I’m using a neo4j for tracking simple relationships and querying a node’s neighbours recursively. However, I’m unsure if I need a full-fledged graph database like neo4j, which offers a lot more complex graph querying capabilities. Apache AGE seems like a potential alternative, as it can be used as an extension on an existing PostgreSQL database, which I have, allows the use of both SQL and...

( Read More... | 2 comments | 2 points | on HN )
Re:Show HN: FiddleCube – Generate Q&A to test your LLM on Tuesday June 25, 2024 @03:48PM (#40792773)
This is super cool!