Visible changes in the SynEdu documentation and release archive.
Version 0.5.0¶
Highlights¶
Subgraph matching now distinguishes monomorphisms from induced isomorphisms, symmetry reduction uses complete automorphisms, and charge, aromaticity, bond order, valence, and hydrogen handling are preserved consistently.
The canonicalization lesson now connects molecular parsing, Weisfeiler--Lehman refinement, individualization--refinement search, canonical atom ranking, and serialization in one notation-consistent workflow.
Forward and backward predictions now use the same graph-native renderer. Responsive SVG cards keep complete reactions visible and clearly distinguish the reference, matching predictions, and other candidates.
All nine lessons have deterministic notebook exports with working Colab, Binder, download, and local-run links. Each export is tested both in the repository and as a standalone downloaded notebook.
Mathematical and chemical correctness¶
Molecular graph matching now uses explicit node and edge semantics for elements, formal charges, aromaticity, and bond order.
Orbit-based match deduplication was replaced by full automorphism handling, avoiding incorrect equivalence classes for symmetric graphs.
Wildcard completion selects the smallest chemically allowed open valence, including charged and aromatic atoms, and rejects unsupported bond orders.
Explicit-to-implicit hydrogen conversion preserves hydrogens that cannot be absorbed safely.
MCS helpers validate their inputs, retain historical positional arguments, and maximize atom count by default to match the lesson metrics.
Reaction balance checks include both atom counts and net formal charge; auxiliary-species imputation respects the direction of the atom imbalance.
Definitions, symbols, equations, and citations were aligned across the lessons, with journal references and DOI metadata updated where available.
Learning experience¶
The homepage, lesson cards, learning-path progress, navigation, footer, dark mode, quizzes, solutions, and discussion sections share one visual system.
The sticky header shows the current page title after scrolling and truncates long lesson names without overlapping the navigation controls.
The atom-map canonicalization figure is maintained as LaTeX/TikZ source and published as an accessible SVG.
The external-learning collection now includes TeachOpenCADD and the introductory pharmaceutical-AI course.
Build and release reliability¶
Fast tests cover graph semantics, MCS behavior, reaction balancing, visualization output, portable links, and site-asset publication.
CI propagates failures through logged pipelines, executes the documentation, verifies asset-injection idempotency, and publishes portable notebook artifacts.
Read the Docs builds use deployment-prefix-safe links and inject the same static enhancements as local and CI builds.
Release builds install Python 3.11 explicitly, validate the Git tag against the project version, check wheel and source archives with Twine, generate checksums, and attest pushed release artifacts.
Version 0.1.0¶
The nine-notebook route is organized into three stages: fundamentals, rule library construction, and rule application.
The active docs build moved from Sphinx / Jupyter Book 1 to Jupyter Book 2 and MyST.
Each lesson’s source of truth is a Jupytext MyST Markdown file
(synedu/S0X/notebook.md), rendered and executed directly by the site.
Portable .ipynb exports are versioned under docs/downloads/.
The docs cite the published Zenodo archive, with a concept DOI for the latest
version and a version DOI for v0.1.0.
Full commit history and release notes live on GitHub.