Get HTML context for LLM-assisted parser development

A text-output alternative to pb_inspect for non-interactive environments such as LLM agents. Prints a compact structured summary of page HTML — meta tags, JSON-LD blocks, <time> elements, most common paragraph-parent selectors, and elements with keyword-bearing class names — to help identify CSS selectors for the four required parser fields without needing a browser.

Usage

pb_html_context(x, n = 1L)

Arguments

x

A data frame from pb_collect.

n

Integer; number of articles to summarise (default 1L).

Value

Invisibly returns a character vector of summaries (one per article). Prints to the console as a side effect.

Examples

if (FALSE) { # \dontrun{
test_data <- pb_collect("https://denverpost.com/feed")
pb_html_context(test_data, n = 3L)
} # }