Class: LlmConductor::Clients::BaseClient

Inherits:
Object
  • Object
show all
Includes:
Concerns::Retryable, Prompts
Defined in:
lib/llm_conductor/clients/base_client.rb

Overview

Base client class providing common functionality for all LLM providers including prompt building, token counting, and response formatting.

Every provider request goes through Concerns::Retryable, so transient rate-limit/overload failures are retried according to configuration.max_retries and configuration.retry_delay. See docs/retries.md for the exact policy.

Constant Summary

Constants included from Concerns::Retryable

Concerns::Retryable::JITTER_RATIO, Concerns::Retryable::MAX_BACKOFF_SECONDS, Concerns::Retryable::MAX_RETRY_AFTER_SECONDS, Concerns::Retryable::RETRYABLE_STATUSES, Concerns::Retryable::TRANSPORT_ERRORS

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Concerns::Retryable

sleep_for

Methods included from Prompts

#prompt_analyze_content, #prompt_classify_content, #prompt_custom, #prompt_extract_links, #prompt_summarize_text

Constructor Details

#initialize(model:, type:, params: {}) ⇒ BaseClient

Returns a new instance of BaseClient.



23
24
25
26
27
 
# File 'lib/llm_conductor/clients/base_client.rb', line 23

def initialize(model:, type:, params: {})
  @model = model
  @type = type
  @params = params
end

Instance Attribute Details

#model ⇒ Object (readonly)

Returns the value of attribute model.



21
22
23
 
# File 'lib/llm_conductor/clients/base_client.rb', line 21

def model
  @model
end

#params ⇒ Object (readonly)

Returns the value of attribute params.



21
22
23
 
# File 'lib/llm_conductor/clients/base_client.rb', line 21

def params
  @params
end

#type ⇒ Object (readonly)

Returns the value of attribute type.



21
22
23
 
# File 'lib/llm_conductor/clients/base_client.rb', line 21

def type
  @type
end

Instance Method Details

#generate(data:) ⇒ Object 



29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
 
# File 'lib/llm_conductor/clients/base_client.rb', line 29

def generate(data:)
  prompt = build_prompt(data)
  input_tokens = calculate_tokens(prompt)
  output_text = with_retries { generate_content(prompt) }
  output_tokens = calculate_tokens(output_text || '')

  # Logging AI request metadata if logger is set
  configuration.logger&.debug(
    "Vendor: #{vendor_name}, Model: #{@model} " \
    "Output_tokens: #{output_tokens} Input_tokens: #{input_tokens}"
  )

  build_response(output_text, input_tokens, output_tokens, { prompt: })
rescue StandardError => e
  build_error_response(e)
end

#generate_simple(prompt:) ⇒ Object

Simple generation method that accepts a direct prompt and returns a Response object



47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
 
# File 'lib/llm_conductor/clients/base_client.rb', line 47

def generate_simple(prompt:)
  input_tokens = calculate_tokens(prompt)
  output_text = with_retries { generate_content(prompt) }
  output_tokens = calculate_tokens(output_text || '')

  # Logging AI request metadata if logger is set
  configuration.logger&.debug(
    "Vendor: #{vendor_name}, Model: #{@model} " \
    "Output_tokens: #{output_tokens} Input_tokens: #{input_tokens}"
  )

  build_response(output_text, input_tokens, output_tokens)
rescue StandardError => e
  build_error_response(e)
end