WordPress Taxonomy Pipeline Pt.2: Wikidata + Breadcrumb JSON & WP Plugin Demo
Part 2 of 2 | AI-Powered WordPress Taxonomy Migration
Continuing from Part 1 — Step 3 (Wikidata enrichment) has finished
processing 100 WordPress tags. Now I complete the pipeline and wire the
output directly into WordPress.
What you will see:
Step 3 results — Each tag now carries a Wikidata canonical Q-number,
multilingual labels, and a human-readable description. “Airbnb” is no
longer just a string — it is a knowledge-graph entity with provenance,
description, and links to related concepts.
Step 4 — Write enriched proposals to two custom MySQL tables
(wp_breadcrumb_terms, wp_breadcrumb_proposals) that sit alongside
WordPress core tables without touching them. Each term gets a structured
JSON breadcrumb array, e.g.:
[
{“label”: “Home”, “url”: “/”},
{“label”: “Tech”, “url”: “/category/tech”},
{“label”: “Airbnb”, “url”: “/tag/airbnb”}
]
Command used:
python source/pipeline/004_step_4_breadcrumb_proposal.py \
–limit 100 –taxonomy post_tag –no-dry-run
WordPress plugin demo — The breadcrumb-migration plugin imports the
validated JSON proposals back into WordPress. I show how the admin UI
displays the enriched proposals and how the structured breadcrumb data
appears on tag archive pages in the Docker environment.
Stack used:
– Python 3.9 + Wikidata REST API
– MySQL custom tables alongside WordPress core
– breadcrumb-migration WordPress plugin
– Docker local WordPress environment (localhost:8080 / localhost:8081)
Why this matters:
The JSON breadcrumb proposals are the bridge between raw taxonomy strings
and structured, machine-readable content. Schema-markup breadcrumbs signal
to search engines AND AI agents where each page sits in your site
hierarchy. That is critical for both classical SEO link equity and GEO —
making your content usable by LLMs that reason over structured knowledge,
not free-form HTML.
Full article: https://wp.me/p3Vuhl-3rb
GitHub: https://github.com/bflaven/ia_usages/tree/main/ia_seo_ia_semantic_breadcrumb_webmcp
Missed Part 1? Watch the setup + Steps 1 and 2 first.
Related Content
Tag(s) : Agile, AI, AI-generated, CMS, Development, Docker, IA, JSON, MySQL, NLP, P.O, Plugin, PO, POC, Python, WordPress, Workflow
Categorie(s) : Agile, Anaconda, Development, Experiences, Python, Tutorials, Videos