shopdb-flask/docs/ROADMAP.md

Roadmap

shopdb-flask is pre-1.0 on both series. Current contract and product versions are in PROJECT-MAP.md, which is generated from the code rather than typed here. This document captures what stands between today and a stable 1.0.0 release. Maintained as scope evolves; supersedes nothing in the ADRs.

Phase status

Phase	Status	Commit
0 - Lock platform contract, naming, style enforcement	DONE	d6725c0
1 - pytest baseline, production hardening, pinned requirements	DONE	2d1bb83
2 - Plugin contract surface and compliance tests	DONE	5fefb53
3 - Manifest-first loader, shopdb.api namespace, auto-register blueprints	DONE	6f085a1
4 - Plugin scaffolding (flask plugin new)	DONE	8eb9362
5 - Alembic baseline, per-site deploy, ADRs to docs/adr	DONE	d4e3ac9
6 - Multi-site distribution readiness (settings-driven branding/ServiceNow/floor plan, security closeout, docs + Docker frontend build, release engineering)	DONE	v0.5.0-v0.7.0

The last big milestone before 1.0 is the legacy-ASP data import plus a production pilot deployment; the framework work below is what remains after that.

What's left before tagging 1.0.0

Must-have

• Asset model fully wired. Asset.mapx, Asset.mapy columns, AssetRelationship.inheritsposition column, and the relationshiptypepropagations M:N table (RelationshipTypePropagation model; propagation lives on RelationshipType, not AssetRelationship). Models match the locked ADR-001 surface that resolve_asset_position already targets.
• Equipment data migration script for facilities migrating from legacy ASP shopdb. One-shot script under scripts/migration/. Per ADR-001, only category='Equipment' AND machinenumber IS NOT NULL migrates.
• Printers retirement. The printers plugin already runs on the asset architecture (blueprint printers_asset_bp); any remaining legacy printer-table cleanup is coordinated with the equipment migration.
• External plugin UI packaging. The Vue-side hook contract ships (ADR-010: get_settings_cards / get_asset_panels / get_map_overlays / get_asset_presentation) and route gating is backend-driven (ADR-009), but plugin routes/views still live in core frontend/src. Let an external plugin ship its own Vue bundle so adopters can add UI without editing core.

Nice-to-have

• Full palette theming. brand_primary_color and a few brand colors are settings-driven, but the rest of the CSS palette (surfaces, borders, accents) is still hardcoded in style.css. A complete theming pass would expose the palette as CSS variables a site can override, not just the brand colors.
• Frontend scaffolding skill (the backend has flask plugin new; the frontend stub is currently manual copy-paste).
• Marketplace listing site (PLUGINS.md is a one-pager; a proper listing with links to sister-site plugins becomes useful when there are more than three external plugins).
• Plugin contract surface diff tooling. Today version bumps are manual judgment; a CI check that diffs the contract surface against the previous tag would catch missed bumps. See ADR-002.
• Calibration cycles, maintenance windows, downtime tracking (domain extensions; would likely live in the machines and measuringtools plugins).

Deferred (out of scope for 1.0)

• Multi-tenancy (rejected by ADR-004; revisit only if more than five sites adopt and operational overhead becomes painful).
• Pip-installable plugins (deferred per ADR-003 v2). Filesystem distribution stays the v1 model.
• Event bus on BasePlugin (removed per ADR-001; add via new ADR if a real use case appears).
• Frontend rebuild beyond Vue 3 + Pinia + Vite (the existing stack is fine).
• Secret and file sharing as a Tech Tool (tabled 2026-08-21). A one-view, expiring link for handing someone a credential, in place of a chat message or a sticky note. NOT a Tools-plugin candidate as it stands: that plugin is browser-only by design (get_blueprint and get_models are both empty) and this needs storage, expiry and an audit trail, so it would be its own plugin. Decisions already reached, to save rediscovering them: the sender is always authenticated; a FILE recipient should be logged in too, which costs nothing on an intranet-only server where everyone with the link already has an account; separate permissions for sharing a secret and sharing a file, because the risk classes differ; and zero-knowledge encryption is right for a credential but questionable for a controlled document, where "the server cannot say what was in it" is a liability rather than a feature. Export control should confirm internal link sharing in writing before it carries drawings or NC programs.

What 1.0.0 means

Tagging 1.0.0 is a commitment that the contract surface is stable for at least the next minor version cycle. Plugin authors at sister sites can pin core_version: ">=1.0.0,<2.0.0" and trust their plugin will not break on framework patches.

Concretely, 1.0.0 ships when:

1. ADR-001's full contract surface is implemented in code, not just documented (Asset.mapx, RelationshipType.propagatesthroughid, etc.)
2. The contract test suite covers every documented hook with both happy-path and a "broken plugin" failure-isolation test
3. At least one external plugin (likely measuringtools from the scaffold canary) has been built end-to-end with no contract changes required mid-build
4. The deploy runbook (DEPLOY.md) has been validated by an actual fresh-host deploy

Decision log pointers

When a roadmap item gets prioritized, document the why in a new ADR and link from this file. The ADRs are the canonical source for design decisions; this file is the prioritized backlog.

• ADR-001 - Asset as platform contract
• ADR-002 - Plugin contract versioning
• ADR-003 - Plugin distribution model
• ADR-004 - Deployment topology (per-site)
• ADR-005 - Equipment vs measuringtools
• ADR-006 - Collector contract pattern
• ADR-007 - Product versioning and releases
• ADR-008 - Plugin migration ownership (per-plugin chains)
• ADR-009 - Frontend plugin route gating
• ADR-010 - Frontend plugin hook contract
• ADR-011 - Machines rename + modeltypes retyping
• ADR-012 - GE-Enforce manifest ownership