Analyses the rendered article HTML to extract insightful titles,
descriptions, and keywords. This runs after buildContent
produces the article body so that metadata truly reflects what the
reader will see — not mechanical counts from the raw data payload.
The analysis extracts:
Headings (h2/h3) as topic indicators
The lede paragraph for a content-based description
Key statistics (numbers, percentages) for title highlights
Entity names (committees, legislation titles) for keywords
Description
Content-based metadata analysis for articles.
Analyses the rendered article HTML to extract insightful titles, descriptions, and keywords. This runs after buildContent produces the article body so that metadata truly reflects what the reader will see — not mechanical counts from the raw data payload.
The analysis extracts: