Skip to content

v8.3.0

Choose a tag to compare

@I-HiMo-I I-HiMo-I released this 09 Aug 11:45

v8.3.0 — 2026-07-02

Small release — hybrid name-cleaning: precise description-match + capitalisation fallback

What changed

Combined both approaches into one robust two-step method:

  1. Precise match (primary): since the description field is already scraped correctly and separately, look for that exact text inside the title and cut the title off right there — no guessing when we already know the answer.
  2. Capitalisation heuristic (fallback): used only when there's no description to compare against, or it doesn't appear in the title. Detects the boundary between an ALL-CAPS name and following sentence-case text, handling both glued-with-no-space and normal space-separated cases.

Verified against 7 real-world scenarios including the exact reported case:

IN:   MANCHESTER FURNITURE SHOWJanuary Furniture Show is devoted to
      contract buyers, interior designers, specifierr or procurement
      officers. From mainstream volume to high-end premium design
NAME: MANCHESTER FURNITURE SHOW

All 7 test cases pass, including titles with no contamination (left untouched), glued acronym boundaries, and year-glued-to-text boundaries.

  • split_title_and_explanation() now accepts the description as a second parameter — the call site in render_card() was updated to actually pass it through (this was missing in v8.2.0, meaning the precise-match step could never run — likely contributing to the earlier issue)
  • Verified brace/parenthesis balance across all 9 PHP files before packaging this time, plus a full method review, given the previous version had a reported problem