Sphinx Canonical Integration Phase1
Overview
This document defines the first canonical integration phase after the validated Batch 0 Sphinx foundation.
The user requested continuing the integration of the isolated documentation work rather than stopping at the empty portal shell.
Phase 1 should therefore promote the highest-value isolated outcomes into canonical repository structure without yet attempting the most conflict-heavy migrations.
The intended outcomes are:
integrate canonical Markdown content into the
Sphinxportal without duplicating source-of-truth documents;expose the selected
Sphinxbacklog and the documentation-platform comparison report inside the canonical portal;surface the recovered
NotebookLMarchive state as a canonical documentation status page;add the first real API-reference slice for:
scripts/models/feedforward_network.pyscripts/training/train_feedforward_network.py
Technical Approach
Scope For Phase 1
Phase 1 should extend the Batch 0 foundation with:
MySTsupport for Markdown integration insideSphinx;a canonical
Project Guidepage sourced fromdoc/guide/project_usage_guide.md;a canonical
Technical Notespage sourced from the recoveredSphinxbacklog master note;a canonical
Selected Reportspage sourced from the recovered documentation-platform comparison report;a canonical learning-guide/archive status page that records the recovered
NotebookLMarchive inventory;canonical API pages for the first two repository modules selected in the backlog.
Why MyST Is Needed Now
Batch 0 deliberately deferred Markdown integration.
Phase 1 now needs it because the repository sources to expose canonically already exist in Markdown:
doc/guide/project_usage_guide.mddoc/technical/...sphinx_documentation_architecture_backlog_and_github_pages_plan.mddoc/reports/analysis/...code_documentation_platform_comparison_report.md
Using MyST allows the Sphinx portal to include those canonical Markdown files through wrappers instead of duplicating or rewriting them into .rst immediately.
API Integration Strategy
The first API slice should remain conservative.
It should not try to expose every helper in the training module yet. Instead:
add richer Google-style docstrings to the module-level entry points and primary public objects;
create explicit API pages with
autoclassandautofunctiondirectives for the selected callables;avoid expanding into broader training infrastructure modules until the first slice renders cleanly.
Explicit Non-Goals
Phase 1 still does not include:
migration of learning-guide folders from
doc/reports/analysis/learning_guides/todoc/guide/movement of imported
NotebookLMmedia into canonical guide foldersGitHub Pagesworkflow implementationREADME.mdlink to a live documentation sitebroad repository-wide docstring rewrites beyond the two selected modules
Involved Components
README.mdMain project document that must reference this technical note.doc/README.mdInternal documentation index that must also reference this technical note.requirements.txtDependency list that must start trackingmyst-parser.docs/conf.pyCanonical Sphinx configuration to extend with Markdown support.docs/index.rstPortal home page that must expose the new integrated sections.doc/guide/project_usage_guide.mdCanonical usage guide to expose inside the Sphinx portal through a wrapper.doc/technical/2026-03/2026-03-22/2026-03-22-12-40-00_sphinx_documentation_architecture_backlog_and_github_pages_plan.mdCanonical backlog master note to expose through the portal.doc/reports/analysis/2026-03-22-10-28-00_code_documentation_platform_comparison_report.mdCanonical analysis report to expose through the portal.reference/isolated_handoff/notebooklm_exports_provenance_manifest.mdRecovered archive tree to document canonically without yet migrating files into guide folders.scripts/models/feedforward_network.pyFirst canonical model API module to document.scripts/training/train_feedforward_network.pyFirst canonical training API module to document.
Implementation Steps
Create this technical note and register it in
README.mdanddoc/README.md.Add
myst-parserto the canonical dependency list and enable it indocs/conf.py.Add canonical wrapper pages for:
the project usage guide
the Sphinx backlog master note
the documentation-platform comparison report
Add a canonical archive-status page for the recovered
NotebookLMimports.Improve the docstrings and module descriptions for the first API slice.
Add canonical API pages for the selected model and training modules.
Validate the integrated portal with a real
Sphinx -WHTML build.