Module: Legion::LLM::API::Namespaces::OpenAI::Embeddings

Extended by:
Legion::Logging::Helper
Defined in:
lib/legion/llm/api/namespaces/openai/embeddings.rb

Class Method Summary collapse

Class Method Details

.registered(app) ⇒ Object 



15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
 
# File 'lib/legion/llm/api/namespaces/openai/embeddings.rb', line 15

def self.registered(app)
  log.debug('[llm][api][namespaces][openai][embeddings] registering routes')

  app.post '/v1/embeddings' do
    require_llm!
    body  = parse_request_body
    input = body[:input]
    model = body[:model] || Legion::Settings[:llm][:default_model]

    if input.nil? || (input.respond_to?(:empty?) && input.empty?)
      return openai_error('input is required', type: 'invalid_request_error',
                                              param: 'input', code: nil, status_code: 400)
    end

    encoding_format = body[:encoding_format].to_s

    # One entry per input item, input order preserved (N -> N). A
    # single string goes through the same result-hash shape as a
    # batch so the response shaping below is one code path.
    if input.is_a?(Array)
      embed_results = Legion::LLM.embed_batch(input, model: model)
      log.info("[llm][api][namespaces][openai][embeddings] action=accepted model=#{model} input_count=#{input.size}")
    else
      embed_results = [Legion::LLM.embed(input, model: model)]
      log.info("[llm][api][namespaces][openai][embeddings] action=accepted model=#{model} input_length=#{input.to_s.length}")
    end

    entries = embed_results.map do |result|
      vector = if result.is_a?(Hash)
                 result[:vector] || result['vector'] ||
                   result[:embedding] || result['embedding']
               else
                 result
               end
      {
        vector: vector.is_a?(Array) ? vector : [],
        tokens: result.is_a?(Hash) ? (result[:tokens] || result['tokens']) : nil
      }
    end

    response_body = Legion::LLM::API::Translators::OpenAIResponse.format_embeddings(
      entries, model: model, input_texts: input, encoding_format: encoding_format
    )

    # The embed result hash carries the selected lane's attribution
    # (provider/instance/model) at the top level — flat-hash path.
    # Every batch entry shares the one selected lane, so entry 0
    # attributes the whole response.
    first_result = embed_results.first
    set_routing_response_headers(routing: first_result) if first_result.is_a?(Hash)

    log.info("[llm][api][namespaces][openai][embeddings] action=complete model=#{model} entries=#{entries.size} dims=#{entries.first[:vector].size}")

    audit_provider = first_result.is_a?(Hash) ? first_result[:provider].to_s : ''
    Legion::LLM::Audit.emit_prompt(
      request_id:   SecureRandom.uuid,
      caller:       build_server_caller(source: 'openai_embeddings', path: request.path, env: env),
      routing:      { model: model, provider: audit_provider.empty? ? 'embed' : audit_provider },
      tokens:       { input_tokens: response_body.dig(:usage, :prompt_tokens), output_tokens: 0 },
      request_type: 'embedding',
      timestamp:    Time.now
    )

    content_type :json
    Legion::JSON.dump(response_body)
  rescue Legion::LLM::AuthError => e
    handle_exception(e, level: :error, handled: true, operation: 'llm.api.namespaces.openai.embeddings.auth')
    openai_error(e.message, type: 'authentication_error', status_code: 401)
  rescue Legion::LLM::ProviderDown, Legion::LLM::ProviderError => e
    handle_exception(e, level: :error, handled: true, operation: 'llm.api.namespaces.openai.embeddings.provider')
    openai_error(e.message, type: 'server_error', status_code: 502)
  rescue Legion::LLM::Errors::RoutingRejected => e
    translate_routing_rejected(e, dialect: :openai, operation: 'llm.api.namespaces.openai.embeddings.routing_rejected')
  rescue StandardError => e
    handle_exception(e, level: :error, handled: false, operation: 'llm.api.namespaces.openai.embeddings')
    openai_error(e.message, type: 'server_error', status_code: 500)
  end

  log.debug('[llm][api][namespaces][openai][embeddings] routes registered')
rescue StandardError => e
  handle_exception(e, level: :error, handled: false, operation: 'llm.api.namespaces.openai.embeddings.register')
end