Add enterprise-grade documentation, packaging, CI, and runnable examples

[Copilot]

The repo had a single-line README and no packaging, CI, or usage examples — making it hard to onboard users or contribute.

Documentation

• README.md — full rewrite: architecture diagram, feature table, installation, quick-start snippets, API overview, training guide, edge export, VictorOS integration, tri-model section, project structure
• docs/installation.md — system requirements, all install paths (source/editable/pip), GPU setup, verification script, troubleshooting
• docs/user_guide.md — 13-section guide covering every component end-to-end
• docs/api.md — parameter tables and return-value docs for all 20+ public classes/functions
• docs/architecture.md — design rationale, attention maths, containment system tiers, VictorOS event taxonomy, memory footprints, trade-off table

Packaging & Infrastructure

• pyproject.toml — package metadata, dependencies, dev/benchmarks extras, lgt-export CLI entry point, pytest + coverage config
• __init__.py — __version__ = "0.1.0", clean __all__ re-exporting the full public API
• LICENSE — MIT
• CHANGELOG.md / CONTRIBUTING.md — standard project hygiene
• .github/workflows/ci.yml — matrix test on Python 3.9 + 3.11, then smoke-tests all example scripts; least-privilege contents: read permissions

Examples (examples/)

Five self-contained runnable scripts:

python examples/basic_inference.py       # continuous embeddings, token IDs, snapshots
python examples/language_model.py        # next-token training with ContainmentProtocol
python examples/victorcos_integration.py # Ledger, MirrorLayer, @victoros_module
python examples/edge_export.py           # quantisation comparison, export pipeline
python examples/tri_model_fusion.py      # world/self/env streams, Mirror Layer callback

No source logic was changed; all 66 existing tests continue to pass.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4a891e4b34

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Include actual Python modules in distribution

Configure packaging to include the repository’s top-level modules, because this packages.find setup discovers no importable package in the current layout (all runtime code is in root-level .py files like lightweight_gravitational_transformer.py, training.py, etc.). As a result, built wheels contain only .dist-info metadata and users installing the project cannot import the library.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Point console script at an existing callable

The console entry point targets export_edge_model:main, but export_edge_model.py does not define a main function (it only has an if __name__ == "__main__" block). Packaging tools will still generate lgt-export, but invoking it will fail at runtime when it cannot resolve the callable.

Useful? React with 👍 / 👎.