Perplexed.Search.Hit (Perplexed v0.2.0)

One search result, normalized across providers.

description is the provider's text about the page for this query: a snippet, highlights, or a long summary, depending on the provider. content is page text when the provider returns it (DeepSeek with need_content: true), otherwise nil. document is the scraped page when the search ran with scrape: true and the page was readable; otherwise scrape_error says why it was not (for example {:unreadable, %{attempts: [direct: :blocked, impersonated: :blocked], cost: 0.0}}). engines names the upstream engines that returned the page, for metasearch providers such as Perplexed.Search.SearXNG; other providers leave it empty. metadata holds what else the provider said about the result, such as DeepSeek's rerank_score, authority, snippet and site_name; Perplexed.answer/2 shows the ranking step the ones it knows (see Perplexed.Pipeline).

Summary

Types

t()

Types

t()

@type t() :: %Perplexed.Search.Hit{
  content: String.t() | nil,
  description: String.t() | nil,
  document: Perplexed.Document.t() | nil,
  engines: [String.t()],
  metadata: map(),
  provider: atom(),
  published_at: String.t() | nil,
  scrape_error: term(),
  title: String.t() | nil,
  url: String.t()
}