README

AI-powered glossary content engine for Contao 5 — keyword pool management, multi-step generation, SEO-focused entries, internal linking, and full backend control.

Status: Phase 2 (Data layer). Schema + Models + live KPIs in the dashboard. Generation pipeline comes in Phase 3+.

Requirements

PHP 8.2+
Contao 5.3+
Symfony 6.4 / 7.0+

Install

composer require tahericreate/glossary
vendor/bin/contao-console contao:migrate          # creates the 4 tables
vendor/bin/contao-console cache:clear --env=prod

After install, the Glossary entry appears under Content in the Contao backend.

Data model (Phase 2)

Four DCA tables, all defined with full sql specs so contao:migrate creates them automatically:

Table	Purpose	Key indexes
`tl_glossary_keyword`	Keyword pool with lifecycle (pending → approved → generated → published / skipped / archived)	`alias` (unique), `status`, `batchId`, `semanticHash`
`tl_glossary_entry`	The actual glossary articles (title, content, SEO meta, FAQ data, schema JSON-LD, internal links, CTA)	`alias` (unique), `status`, `letter`, `(language, status)`
`tl_glossary_job`	Generation jobs (queue, progress, tokens, cost, step tracking)	`uuid` (unique), `status`, `(status, tstamp)`
`tl_glossary_log`	Verbose per-job logs (closed = append-only in backend)	`(jobId, tstamp)`

Status fields are backed by typed PHP enums:

Tahericreate\Glossary\Keyword\KeywordStatus
Tahericreate\Glossary\Entry\EntryStatus
Tahericreate\Glossary\Job\JobStatus
Tahericreate\Glossary\Job\JobType

Phase 2 file additions

File	Purpose
`contao/dca/tl_glossary_keyword.php`	Keyword DCA + schema
`contao/dca/tl_glossary_entry.php`	Entry DCA + schema (versioning enabled)
`contao/dca/tl_glossary_job.php`	Job DCA + schema
`contao/dca/tl_glossary_log.php`	Log DCA (append-only)
`contao/languages/en/tl_glossary_*.php`	English labels for backend UI
`contao/config/config.php`	Registers `$GLOBALS['TL_MODELS']`
`src/Keyword/KeywordStatus.php`	Lifecycle enum
`src/Entry/EntryStatus.php`	Lifecycle enum
`src/Job/JobStatus.php`	Lifecycle enum (with `isTerminal()`, `isActive()`)
`src/Job/JobType.php`	Job type enum
`src/Model/GlossaryKeywordModel.php`	Contao Model
`src/Model/GlossaryEntryModel.php`	Contao Model
`src/Model/GlossaryJobModel.php`	Contao Model
`src/Model/GlossaryLogModel.php`	Contao Model
`src/Service/KpiAggregator.php`	Aggregates COUNT-by-status for dashboard
`src/Service/SlugGenerator.php`	URL-safe, collision-free alias generation

The dashboard now displays live counts from the database. On a fresh install before contao:migrate runs, the aggregator tolerates missing tables and shows zeros instead of crashing.

Roadmap

Phase 3 — AI Client abstraction (AiClientInterface, OpenAiClient, ClaudeClient) + Prompt Builder with Global / Dynamic layer separation
Phase 4 — Keyword Management UI: Pool generation, status workflow, Duplicate Detection (using semanticHash from Phase 2 schema)
Phase 5 — Job/Step Pipeline: Research → Prompt Assembly → AI Generation → Post-Process → Internal Linking → Publish
Phase 6 — Scheduler + Cron
Phase 7 — Frontend Modules (Alphabet nav, search, entry layout, related)
Phase 8 — SEO + Schema generation + Internal Linking engine

Architecture notes

Settings storage — JSON file in var/glossary/config.json. API keys never touch the database.
Data storage — DCA tables (Phase 2 ✓). Status fields use string columns backed by enum classes in PHP.
Aliases — SlugGenerator produces ASCII-folded, collision-free slugs cross-table. Falls back to a random suffix in the (very rare) case of 1000 collisions.
KPIs — Read directly from DB via Doctrine DBAL (no model hydration). Tolerates missing tables on fresh installs.
Admin-only — Every backend controller calls assertAdmin() at the top.

tahericreate/glossary

包简介

README 文档

README

Requirements

Install

Data model (Phase 2)

Phase 2 file additions

Roadmap

Architecture notes

统计信息

GitHub 信息

其他信息

承接程序开发