Version 1.8.0
This major platform release establishes the SPM Search Database foundation: one structured discovery system for articles, video, galleries, music, radio, television, news, and SPM pages, with phrase-aware relevance and connected parent-child records.
Release Highlights
One Search Across SPM
A common record model makes very different media types discoverable through one coordinated system.
Results With Relationships
Child matches strengthen their parent result instead of flooding visitors with disconnected tracks, images, or episodes.
Useful Before Clever
Find SPM Content Quickly
The first search version is designed to help someone find a Rush article, gallery, video, or music result quickly and understand why it belongs in the results.
No Unnecessary Prediction Layer
The foundation intentionally excludes AI semantic search, recommendations, predictive queries, elaborate animation, and excessive filters. Structured content and explainable relevance come first.
Unified Content Coverage
A Shared Discovery Layer
Version 1 supports Articles, Video, Galleries, Music, Radio, TV, News, and SPM Pages. Search categories organize results without forcing each media type into an unrelated system.
Universal Search Fields
Every record can carry an ID, content type, parent ID, title, summary, description, URL, thumbnail, tags, category, publication and update dates, status, source type, and source ID.
The Index Is Not the Master Copy
Search points visitors to SPM content and preserves its source identity. The index does not replace the authoritative editorial, media, radio, television, or page record.
MariaDB Data Model
spm_search_items
The central table holds the common identity and searchable presentation fields. Its parent relationship uses controlled deletion behavior so a parent cannot disappear while dependent child records remain attached.
Type-Specific Metadata
spm_search_metadata stores additional key-value details that matter to particular content types without making the core record endlessly wider.
Tags, Locations, and Credits
spm_search_tags, spm_search_item_tags, spm_search_locations, spm_search_item_locations, and spm_search_credits preserve reusable structured meaning for topics, places, creators, artists, photographers, authors, publishers, and their roles.
MariaDB 10.11 Foundation
The architecture was established on MariaDB 10.11 with Unicode-safe utf8mb4 handling and indexes selected for the first measurable search workload.
Parent and Child Relationships
Collections Stay Connected
Gallery images belong to galleries, tracks belong to albums, episodes belong to series, and articles can belong to an article series.
Child Matches Strengthen the Parent
A matching image, track, or episode can raise the relevance of its parent while the results remain centered on the useful destination rather than exposing a long list of repetitive child records.
Balanced Result Types
Per-type result caps prevent one large collection or media category from overwhelming the unified result set.
Explainable Relevance
Title
The title receives the strongest field priority because it most directly identifies what the visitor is trying to find.
Tags, Keywords, Artist, and Creator
Curated subjects and credited people receive strong weight, helping searches connect with meaningful structured identity.
Category, Description, and Location
Category is medium-high, while description and location provide supporting context without overpowering a stronger title or tag match.
Body Text Rescue
Deep body content can rescue a relevant item that would otherwise be missed, but its low weight prevents incidental wording from outranking deliberate metadata.
Phrase and Short-Query Rules
Multi-Word Terms Stay Together First
Structured terms such as Classic Rock, Heavy Metal, Creative AI, and Music Industry are evaluated as phrases before their individual words are allowed to broaden the search.
Bounded Exact and Prefix Fallbacks
Very short or punctuation-heavy searches, including artist names such as U2, call signs, and frequencies, receive controlled exact, beginning-of-title, word, and prefix handling when full-text search alone is insufficient.
Metadata Contributes Deliberately
Tags, credits, locations, and type-specific metadata add measured relevance rather than being flattened into one unexplained block of text.
Foundation Validation
The approved database and relevance foundation progressed through Stage 8C with controlled data. Phrase-aware full-text search, weighted fields, structured metadata scoring, short-query fallbacks, parent-child enrichment, result collapsing, and per-type caps were demonstrated together.
This milestone proves the search model and its core integration. Production ingestion, administrative tooling, the public tabbed search interface, and the complete site inventory remain separate delivery steps.
Known Limitations
The foundation is not yet the complete public search product
Version 1.8.0 validates the database, relationships, and relevance model. Production ingestion and inventory work are documented in later releases, and the complete visitor interface remains additional work.
Search quality depends on source metadata
Strong titles, descriptions, tags, credits, locations, and parent relationships produce better results. Missing or inconsistent source information cannot be fully corrected by ranking rules.
Relevance requires measurement at production scale
The initial indexes and weights are intentionally practical. They should be measured and tuned as the quantity and variety of indexed SPM content grows.
Advanced prediction is intentionally excluded
This release does not provide semantic AI search, recommendations, or predictive queries. The first objective is dependable, understandable retrieval across SPM's own content.