david/library-manager
1
0
Fork 0
Code Issues 2 Pull requests Projects Releases 10 Packages Wiki Activity Actions
Page: Development Roadmap
Pages
Acquisition and Import Configuration Deployment Development Roadmap Home Metadata Organization Profiles Usage v0.1.0 Lab Release v0.1.1 Lab Refresh v0.2.0 Lab Refresh v0.2.3 Readarr Navigation Refresh v0.2.4 Release Display Fix v0.2.5 Completed Download Import Bridge v0.2.6 Output Profile Controls v0.2.7 Acquisition Milestone v0.2.8 Docker Healthcheck
25 Development Roadmap
david edited this page 2026-06-19 09:07:30 -07:00
Table of Contents

Development Roadmap

Forgejo milestones track the main project phases. Repo docs remain the detailed source of design decisions.

Milestones

  • M0 Planning and Safety Foundation
  • M1 Scanner and Dry-Run Inventory: closed
  • M2 Metadata Model and Resolver: closed
  • M3 Organization and Downstream Profiles: closed
  • M4 Acquisition Integrations: closed for the lab milestone
  • M5 Web App Scaffold and Operator UI: authenticated Readarr-style lab console present
  • M6 Deployment and Operations: Docker healthcheck and tag policy present
  • M7 Documentation and Wiki Completeness: #13 open

Current Issues

  • #1 Implement read-only fixture scanner foundation: closed
  • #2 Define versioned Docker release packaging: closed by v0.2.8
  • #3 Explore final project name and logo direction
  • #4 Design scale-aware scanner persistence and database strategy: closed
  • #5 Design metadata source lookup and enrichment workflow: closed
  • #6 Design organization dry-run plans and downstream profiles: closed
  • #7 Add local prototype operator UI: closed
  • #8 Add acquisition foundation: closed
  • #9 Implement first real download client adapter: closed
  • #10 Add Proxmox LXC prototype installer: closed
  • #11 Back Add New and Activity with durable APIs: closed by v0.2.x
  • #12 Add real indexer and search configuration: closed by v0.2.7
  • #13 Publish lab docs and wiki refresh: ongoing with each release

Naming

  • Product name: Pagemaster
  • Tagline: "Look to the books"
  • Name/logo issue: #3

v0.1.0 Lab Release Slice

The v0.1.0 lab slice packages the first operable Pagemaster shell:

  • React/Vite operator console embedded in the Go server.
  • First-visitor setup for the initial operator account.
  • Role foundation for operator/requester/viewer accounts.
  • Server-side password hashing and setup/status APIs.
  • App configuration summary endpoint.
  • SQLite schema additions for users, sessions, settings, secrets, output profiles, automation profiles, and per-root apply behavior.
  • Docker packaging as pagemaster:0.1.0.
  • Calibre included in the image for self-contained ebook conversion tooling.
  • Docker smoke verification for /api/health, /api/setup/status, and the frontend shell.

Follow-up before a broader real-library lab run:

  • Enforce sessions beyond initial setup.
  • Build in-app configuration pages for roots, output profiles, secrets, and automation policy.
  • Add real acquisition configuration for Prowlarr/Newznab/Torznab and download clients.
  • Research and document MyAnonamouse setup accurately before implementing source-specific behavior.

v0.2.8 Docker Healthcheck

The current LXC/mainline prototype reports v0.2.8. It keeps the v0.2.7 acquisition milestone and adds Docker package hardening: the image has a built-in pagemaster healthcheck command, the Dockerfile declares HEALTHCHECK, and the Docker docs define immutable version tags plus reserved moving-tag policy for lab and latest.

Next slices should focus on provider identifiers in matching/ranking, controlled auto-grab and auto-import policy gates, MyAnonamouse/provider-specific behavior after verified setup research, downstream Audiobookshelf/Grimmory apply triggers, and upgrade/operator documentation.

M2 Result

The backend/scanner direction has moved from JavaScript to Go. The current repo keeps the scanner as a Go package under internal/scanner, shared synthetic stress helpers under internal/scannertest, SQLite persistence under internal/store, metadata resolver contracts under internal/metadata, and CLIs under cmd/.

M2 now includes:

  • db/schema.sql: scale-aware schema with source roots, scan runs, files, candidates, evidence, duplicate signals, import plans, and acquisition/domain tables.
  • internal/store: SQLite schema application, scanner-report persistence, source-root-scoped candidates, re-scan state, import-plan history, and indexed review counters.
  • internal/metadata: backend provider contract for wanted-author lookup and missing-item enrichment.
  • make check-large-scan: synthetic 30,000-file scanner stress check outside the repo.
  • make check-persistence: synthetic 30,000-file SQLite write validation outside the repo.
  • Go tests for scanner behavior, schema contract checks, scanner-report persistence, missing/moved file re-scans, repeated import-history growth, indexed review counters, and metadata resolver behavior.

Latest synthetic scanner check processed 30,000 files in 404 ms and produced 30,000 candidates, 30,000 evidence records, and 30,000 dry-run import plan actions.

Latest SQLite persistence check scanned 30,000 files in 409 ms and persisted the report in 3209 ms, producing 30,000 scan files, 30,000 candidates, 30,000 candidate-file links, 30,000 metadata evidence rows, one import plan, and 30,000 import plan actions.

SQLite is accepted for the first app scaffold, with the option to revisit after real-library testing.

M3 Progress

The first organization checkpoint is present:

  • internal/organization: Audiobookshelf dry-run planner.
  • organization_plans and organization_plan_actions: persisted dry-run organization plans.
  • make check-organization: fixture organization plan smoke command.
  • Tests for non-destructive copy-shaped operations, narrator folders, disc folders, duplicate review, weak metadata review, and persistence.

Current dry-run behavior:

  • Organize ready candidates.
  • Review duplicates and weak metadata.
  • Hold partial downloads, unsupported files, and sidecar-only candidates.
  • Preserve audiobook disc folders as Disc N.
  • Keep file operations non-destructive until review/apply behavior is explicitly designed.

The M3 foundation now includes downstream dry-run profiles plus copy-only approve/apply/audit behavior. Follow-up work should refine individual action review UI, retry UI, collision resolution choices, audit retention controls, and optional Audiobookshelf scan handoff.

Prototype Result

The first testable operator prototype is present:

  • cmd/library-manager serve: local web server.
  • internal/app: Go-hosted dashboard and JSON API.
  • /api/health: health check.
  • /api/dashboard: persisted scanner/review/organization summary.
  • Configured-root scan APIs persist reports, build organization plans, and return updated dashboard state.
  • make prototype: starts the prototype on http://127.0.0.1:8090.

The prototype uses SQLite under DATA_DIR and applies the schema idempotently on startup, so it can be restarted against the same development database.

Documentation Practice

Each implementation slice should update:

  • Repo planning docs under docs/.
  • The Forgejo milestone or issue state.
  • Wiki pages that explain setup, configuration, usage, or operator-visible behavior.