Skip to content

feat(base): RenderStability declarative API + Qwen3 / Qwen3-VL / DefaultRenderer values#44

Open
mvanhorn wants to merge 1 commit into
PrimeIntellect-ai:mainfrom
mvanhorn:fix/41-renderstability-declarative-api-qwen3-qwen3-vl-def
Open

feat(base): RenderStability declarative API + Qwen3 / Qwen3-VL / DefaultRenderer values#44
mvanhorn wants to merge 1 commit into
PrimeIntellect-ai:mainfrom
mvanhorn:fix/41-renderstability-declarative-api-qwen3-qwen3-vl-def

Conversation

@mvanhorn
Copy link
Copy Markdown

Summary

Adds a declarative RenderStability enum (stable / beta / experimental / deprecated) on the base renderer and populates the stability values for every shipped renderer module. #41 specifically called out that consumers want to filter by stability without source-diving.

Why this matters

Today there's no programmatic way to know whether a renderer is production-tier or experimental. The PR maps each renderer module's status to the enum so downstream tools can branch on it.

Changes

  • renderers/stability.py (new) - the RenderStability enum and helper.
  • renderers/base.py - stability attribute on the base renderer with a default.
  • renderers/__init__.py - re-export the enum from the package root.
  • Per-renderer modules (Qwen3, Qwen3-VL, Qwen3.5, Qwen3.6, DeepSeek-V3, GLM 4.5/5, GPT-OSS, Kimi K2/K2.5, MiniMax-M2, Nemotron-3, default) - declarative stability = RenderStability.X annotation.
  • tests/test_stability.py - tests that the enum round-trips and that each renderer module declares a value.

Testing

New stability tests + existing renderer tests pass locally.

Fixes #41

…ultRenderer values

PrimeIntellect-ai#41 asked for a declarative way to mark renderer stability and to
populate the values for the Qwen3 family and the default renderer.
Adds the RenderStability enum, a stability attribute on the base
renderer, and the values for Qwen3, Qwen3-VL, Qwen3.5, Qwen3.6,
DeepSeek-V3, GLM 4.5 / 5, GPT-OSS, Kimi K2 / K2.5, MiniMax-M2,
Nemotron-3, and DefaultRenderer.

Fixes PrimeIntellect-ai#41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

feat: renderer self-describes prefix-stability for SFT / RL / inference consumers

1 participant