mirror of
https://github.com/discourse/discourse-ai.git
synced 2025-11-01 23:18:38 +00:00
2c0f535bab
* FEATURE: HyDE-powered semantic search. It relies on the new outlet added on discourse/discourse#23390 to display semantic search results in an unobtrusive way. We'll use a HyDE-backed approach for semantic search, which consists on generating an hypothetical document from a given keywords, which gets transformed into a vector and used in a asymmetric similarity topic search. This PR also reorganizes the internals to have less moving parts, maintaining one hierarchy of DAOish classes for vector-related operations like transformations and querying. Completions and vectors created by HyDE will remain cached on Redis for now, but we could later use Postgres instead. * Missing translation and rate limiting --------- Co-authored-by: Roman Rizzi <rizziromanalejandro@gmail.com>
33 lines
895 B
Ruby
33 lines
895 B
Ruby
# frozen_string_literal: true
|
|
|
|
module DiscourseAi
|
|
module Embeddings
|
|
module HydeGenerators
|
|
class Anthropic < DiscourseAi::Embeddings::HydeGenerators::Base
|
|
def prompt(search_term)
|
|
<<~TEXT
|
|
Given a search term given between <input> tags, generate a forum post about the search term.
|
|
Respond with the generated post between <ai> tags.
|
|
|
|
<input>#{search_term}</input>
|
|
TEXT
|
|
end
|
|
|
|
def models
|
|
%w[claude-instant-1 claude-2]
|
|
end
|
|
|
|
def hypothetical_post_from(query)
|
|
response =
|
|
::DiscourseAi::Inference::AnthropicCompletions.perform!(
|
|
prompt(query),
|
|
SiteSetting.ai_embeddings_semantic_search_hyde_model,
|
|
).dig(:completion)
|
|
|
|
Nokogiri::HTML5.fragment(response).at("ai").text
|
|
end
|
|
end
|
|
end
|
|
end
|
|
end
|