Class: ActiveHarness::Providers::Azure

Inherits:
Base
  • Object
show all
Defined in:
lib/active_harness/providers/azure.rb

Overview

Azure OpenAI Service — deployment-based API. learn.microsoft.com/en-us/azure/ai-services/openai/reference

The ‘model:` parameter is treated as the **deployment name** you created in the Azure portal (not the underlying model name).

Required config (or ENV fallback):

config.azure_api_base    — "https://my-resource.openai.azure.com"
config.azure_api_key     — your resource API key
                           (alternatively: config.azure_ai_auth_token for OAuth bearer)

Optional config:

config.azure_api_version — defaults to "2024-05-01-preview"

Resulting endpoint:

POST {azure_api_base}/openai/deployments/{deployment}/chat/completions
     ?api-version={azure_api_version}

Example agent config:

model do
  use provider: :azure, model: "my-gpt4o-deployment", temperature: 0.7
end

Constant Summary

Constants inherited from Base

Base::HTTP, Base::STREAMING_HTTP

Instance Method Summary collapse

Instance Method Details

#call(model:, messages:, temperature: 0.7) ⇒ Object 



28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
 
# File 'lib/active_harness/providers/azure.rb', line 28

def call(model:, messages:, temperature: 0.7)
  url = build_url(model)

  raw  = post_json(url,
    headers: {
      "Content-Type" => "application/json"
    }.merge(auth_header),
    body: { messages: messages, temperature: temperature }
  )
  data = parse!(raw)
  handle_error!(data)

  {
    content:  data.dig("choices", 0, "message", "content").to_s.strip,
    provider: :azure,
    model:    data["model"] || model,
    usage:    extract_usage_openai(data)
  }
end