Files
kunail0804-ComfyUI-Scene-Co…/CHANGELOG.md
T
f098259b8e build: prepare the Version 1.0.0 release (#32) (#76)
Finalize packaging and release readiness (MASTER_SPEC Phase 8, §7.8, §28).

- pyproject.toml: bump version to 1.0.0; ship the JSON Schemas as package data
  and enable include-package-data.
- MANIFEST.in: include the data directories (schemas, knowledge_base, prompts,
  config, examples, docs) in the source distribution.
- CHANGELOG.md: 1.0.0 release notes.
- docs/definition_of_done.md: the §7.8 Definition-of-Done checklist mapped to
  evidence, with the ComfyUI visual check and the tag/release step flagged as the
  remaining manual steps.
- tests/test_release_readiness.py: assert the version, changelog, DoD doc, and
  that every schema / Knowledge Base / config / prompt data file is present and
  declared for packaging.

Verified: `python -m build --sdist` produces comfyui_scene_compiler-1.0.0 with
all 13 Knowledge Base files, 10 schemas, the system prompt, and the default
config bundled. Tagging v1.0.0 and publishing the release are intentionally left
to a human after the ComfyUI visual check.

Co-authored-by: kunail0804 <kunail0804@gmail.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-09 14:18:33 +02:00

1.8 KiB

Changelog

All notable changes to Scene Compiler are documented here. The project follows Semantic Versioning: MAJOR.MINOR.PATCH.

[1.0.0] — Deterministic Compiler Foundation

The first release: a complete, deterministic natural-language-to-prompt compiler for ComfyUI targeting the Illustrious model family.

Compiler

  • Scene Analyzer — the only LLM stage; turns a natural-language description into Scene JSON via a local Ollama backend, with a bounded retry policy and the official system prompt. Never generates tags or invents information.
  • Scene Validator — validates and normalizes Scene JSON against the schema; recoverable issues warn and continue, hard errors stop.
  • Illustrious Resolver — deterministically resolves concepts to Illustrious tags through the Knowledge Base: normalization, direct alias resolution, recursive expansion, and duplicate removal, with full traceability.
  • Category Splitter — groups Resolved Tags into the 19 canonical categories, preserving order.
  • Prompt Builder — formats the Category Map into one output per category plus reserved Negative and Scene outputs.

Foundation

  • JSON Schemas and immutable typed data models for every inter-stage document.
  • Compiler Result wrapper and a stable message-code registry.
  • Typed, injectable configuration and structured logging.
  • Knowledge Base loader with validation tooling and an initial Illustrious dataset (188 entries).

ComfyUI

  • Eight nodes: Scene Analyzer, Scene Validator, Resolver, Category Splitter, Prompt Builder, Debug Viewer, Knowledge Base Loader, and Configuration.

Quality

  • Full unit, integration, golden, and performance test suites; deterministic golden outputs for ten reference scenes.
  • CI on Python 3.11 and 3.12; documentation and ready-to-use examples.