Bug Reports & Feature Requests
We value your feedback and contributions to make ProxAI better. This guide will help you effectively report bugs and suggest new features.
Reporting Bugs
Before Reporting
- Check existing issues on GitHub to see if your bug has already been reported
- Update to the latest version to check if your issue has been fixed
- Try to isolate the problem to create a reliable reproduction case
How to Submit a Bug Report
- Go to our GitHub Issues page
- Click “Bug Report” template
- Fill out all the required information:
- Clear, descriptive title
- Steps to reproduce
- Expected behavior
- Actual behavior
- Screenshots (if applicable)
- Environment details (OS, Python version, etc.)
- Any additional context
What Makes a Good Bug Report
A helpful bug report includes:
**Bug**: Authentication fails when using special characters in password
**Steps to Reproduce**:
1. Install proxai v1.2.3
2. Create client with `px.Client(api_key="my-key")`
3. Attempt to authenticate with password containing `!@#$`
**Expected**: Authentication succeeds
**Actual**: Authentication fails with error `Invalid character in password`
**Environment**:
- ProxAI version: 1.2.3
- Python version: 3.10.4
- OS: macOS 13.1Requesting Features
Before Requesting
- Search existing issues to see if someone has already suggested a similar feature
- Check the roadmap to see if it’s already planned
- Consider the scope of your request and how it fits with ProxAI’s goals
How to Submit a Feature Request
- Go to our GitHub Issues page
- Click “Feature Request” template
- Fill out the template with:
- Clear description of the feature
- Problem it solves
- Proposed implementation (if you have ideas)
- Alternatives you’ve considered
- Additional context or examples
What Makes a Good Feature Request
A helpful feature request includes:
**Feature**: Add support for streaming responses from Anthropic Claude
**Problem**: Currently, users cannot stream responses when using Claude models, leading to long wait times for complete responses.
**Proposed Solution**: Implement streaming support for Claude models similar to the existing streaming for OpenAI.
**Example Usage**:
```python
response = px.generate_text(
"Write a long story about a journey",
provider="anthropic",
model="claude-3-opus",
stream=True
)
for chunk in response:
print(chunk, end="", flush=True)Alternatives Considered: Using a separate method specifically for streaming.
## Issue Lifecycle
1. **Triage**: New issues are reviewed by maintainers
2. **Labeling**: Issues are labeled by type, priority, and status
3. **Discussion**: Community and maintainers discuss the issue
4. **Implementation**: Issue is assigned and worked on
5. **Review & Merge**: Changes are reviewed and merged
6. **Release**: Fixed/implemented in a release
## Community Voting
We use GitHub reactions to gauge community interest:
- 👍 Support for a feature or confirming a bug
- 👎 Disagreement with a feature direction
The number of reactions helps us prioritize what to work on next.
## Stay Updated
To follow progress on your reported issues:
- Enable notifications on the GitHub issue
- Join our [Discord](https://discord.gg/QhrDkzMHrP) for real-time updates
- Check our release notes for fixed issues
Thank you for helping improve ProxAI!