12dbeb57a1
We are basically updating the location references everywhere with a location id and adding the location relationship to the models to be able to fetch location names when needed. This also makes the LocationNameConverter model obsolete, so we are removing it. Use model id as value for Sequel::Model in resource creation form Use id of the location as preselected value in Postgres update form
90 lines
2.8 KiB
Ruby
90 lines
2.8 KiB
Ruby
# frozen_string_literal: true
|
|
|
|
require_relative "../spec_helper"
|
|
|
|
RSpec.describe InferenceEndpoint do
|
|
subject(:inference_endpoint) do
|
|
described_class.new(
|
|
name: "ie-name",
|
|
location_id: Location::HETZNER_FSN1_ID,
|
|
model_name: "model-name"
|
|
) { _1.id = "c76fcd0c-3fb0-40cc-8732-d71869ee1341" }
|
|
end
|
|
|
|
describe "#display_states" do
|
|
let(:strand) { instance_double(Strand) }
|
|
|
|
before do
|
|
allow(inference_endpoint).to receive(:strand).and_return(strand)
|
|
end
|
|
|
|
context "when state is running" do
|
|
before { allow(strand).to receive(:label).and_return("wait") }
|
|
|
|
it "returns 'running'" do
|
|
expect(inference_endpoint.display_state).to eq("running")
|
|
end
|
|
end
|
|
|
|
context "when state is deleting" do
|
|
before { allow(strand).to receive(:label).and_return("destroy") }
|
|
|
|
it "returns 'deleting'" do
|
|
expect(inference_endpoint.display_state).to eq("deleting")
|
|
end
|
|
end
|
|
|
|
context "when state is creating" do
|
|
before { allow(strand).to receive(:label).and_return("wait_replicas") }
|
|
|
|
it "returns 'creating'" do
|
|
expect(inference_endpoint.display_state).to eq("creating")
|
|
end
|
|
end
|
|
end
|
|
|
|
describe "#path" do
|
|
it "returns the correct path" do
|
|
expect(inference_endpoint.path).to eq("/location/eu-central-h1/inference-endpoint/ie-name")
|
|
end
|
|
end
|
|
|
|
shared_examples "chat completion request" do |development|
|
|
let(:http) { instance_double(Net::HTTP, read_timeout: 30) }
|
|
let(:load_balancer) { instance_double(LoadBalancer, health_check_protocol: "https") }
|
|
|
|
before do
|
|
allow(Net::HTTP).to receive(:new).and_return(http)
|
|
allow(http).to receive(:use_ssl=).with(true)
|
|
allow(http).to receive(:read_timeout=).with(30)
|
|
allow(inference_endpoint).to receive(:load_balancer).and_return(load_balancer)
|
|
end
|
|
|
|
it "sends the request correctly" do
|
|
if development
|
|
allow(Config).to receive(:development?).and_return(true)
|
|
allow(http).to receive(:verify_mode=).with(OpenSSL::SSL::VERIFY_NONE)
|
|
end
|
|
|
|
expect(http).to receive(:request) do |req|
|
|
expect(req).to be_an_instance_of(Net::HTTP::Post)
|
|
expect(req["Content-Type"]).to eq("application/json")
|
|
expect(req["Authorization"]).to eq("Bearer api_key")
|
|
expect(req.body).to eq({model: "model-name", messages: [{role: "user", content: "what's a common greeting?"}]}.to_json)
|
|
end.and_return("hello")
|
|
|
|
expect(inference_endpoint.chat_completion_request("what's a common greeting?", "hostname", "api_key")).to eq("hello")
|
|
end
|
|
end
|
|
|
|
describe "#chat_completion_request" do
|
|
context "when production" do
|
|
it_behaves_like "chat completion request", false
|
|
end
|
|
|
|
context "when development" do
|
|
it_behaves_like "chat completion request", true
|
|
end
|
|
end
|
|
end
|