[Feature]: Support Pydantic 1.x API

[barracuda156]

🎯 Feature Summary

Pydantic 1.x is supported by upstream (it is not obsoleted by 2.x) and has an advantage of being pure-python, not pulling in a whole another compiler and llvm.

🔍 Problem Statement

Rust is broken on some non-mainstream platforms, and pydantic 2.x requires rust.
Rust is a huge thing to build and have, and it is desirable not to force it upon users.

💡 Proposed Solution

Support Pydantic 1.x API.

📝 Example Usage

No response

🏷️ Feature Category

Other

⭐ Priority Level

Medium - Nice to have enhancement

👥 Who Would Benefit?

• New users getting started with Cognix
• Daily Cognix users
• Developers working on large projects
• Teams collaborating on code
• Users of specific AI models
• Power users with complex workflows
• Users on specific platforms (Windows/Mac/Linux)

🔄 Alternatives Considered

No response

⚙️ Technical Considerations

No response

🤝 Implementation

• I'd like to implement this feature myself
• I can help with testing and feedback
• I can provide domain expertise
• I can help with documentation
• I prefer to let the maintainers handle it

📄 Additional Context

No response

✅ Checklist

• I have searched existing issues and discussions to avoid duplicates
• This feature aligns with Cognix's core mission (AI development assistant with persistent memory)
• I have provided enough detail for the maintainers to understand the request
• I understand this is a request, not a guarantee of implementation

[cognix-dev]

Hi @barracuda156

Thank you for the detailed feature request and for taking the time to use Cognix!

This is a valid concern - Rust dependencies can indeed be problematic on certain platforms, and supporting pure-Python environments makes sense for broader accessibility.

A few questions to better understand the requirements:

1. Which specific platforms are you encountering Rust build issues on?
2. Are there particular error messages or build failures you're seeing?
3. Would backward compatibility with Pydantic 1.x be sufficient, or do you need it as the primary/only option?

Since you've indicated willingness to help implement this, I'd be happy to discuss the technical approach:

• We could potentially add conditional imports to support both versions
• The main changes would likely be in the data validation and serialization layers
• We'd need to ensure feature parity across both versions

If you'd like to work on this, feel free to:

1. Fork the repo and create a feature branch
2. Start with a small proof-of-concept for the core validation logic
3. Submit a draft PR when ready for initial feedback

This aligns well with Cognix's goal of broad accessibility, so I'm supportive of adding this support.

Let me know your thoughts on the technical approach, and thanks again for the contribution!

[barracuda156]

@cognix-dev Thank you for responding! To address the points above:

Which specific platforms are you encountering Rust build issues on?

Well, myself I am primarily interested in using this on legacy macOS on PowerPC, however, AFAIK, Rust is also broken on some modern OS on non-mainstream archs. For example, OpenBSD on ppc does not have Rust. We won’t fix Rust here, so it suffices to say that there are some edge cases where it is desirable to avoid Rust.

Are there particular error messages or build failures you're seeing?

With pure-Python packages nothing will fail during the “build”, they just won’t work, failing at runtime, if some unsupported modules or functions are used in the code (say, importing pedantic_core will fail, since it does not exist, etc.).

Would backward compatibility with Pydantic 1.x be sufficient, or do you need it as the primary/only option?

It makes sense to prioritize mainstream cases, so if there is something in pydantic 2.x which makes it desirable over 1.x, rather not completely move to 1.x (I won’t expect that to happen anyway, but just in case).
So I am only asking for compatibility.

We'd need to ensure feature parity across both versions

This would be great, but TBH I think it is okay if some select features are not available with pydantic 1.x – certainly it is better to have core functionality working than to have nothing at all on rustless platforms.

[cognix-dev]

@barracuda156 Thank you for the detailed information!
I plan to keep Pydantic 2.x as the default on mainstream platforms,
but I’ll work on providing a Pydantic 1.x adapter and other Rust-free paths
for environments where pydantic_core isn’t available.
Your feedback is invaluable and really helps improve portability — thank you!