discourse-ai/spec/models/ai_tool_spec.rb

679 lines
20 KiB
Ruby

# frozen_string_literal: true
RSpec.describe AiTool do
fab!(:llm_model) { Fabricate(:llm_model, name: "claude-2") }
let(:llm) { DiscourseAi::Completions::Llm.proxy("custom:#{llm_model.id}") }
fab!(:topic)
fab!(:post) { Fabricate(:post, topic: topic, raw: "bananas are a tasty fruit") }
fab!(:bot_user) { Discourse.system_user }
def create_tool(
parameters: nil,
script: nil,
rag_chunk_tokens: nil,
rag_chunk_overlap_tokens: nil
)
AiTool.create!(
name: "test #{SecureRandom.uuid}",
tool_name: "test_#{SecureRandom.uuid.underscore}",
description: "test",
parameters:
parameters || [{ name: "query", type: "string", description: "perform a search" }],
script: script || "function invoke(params) { return params; }",
created_by_id: 1,
summary: "Test tool summary",
rag_chunk_tokens: rag_chunk_tokens || 374,
rag_chunk_overlap_tokens: rag_chunk_overlap_tokens || 10,
)
end
it "it can run a basic tool" do
tool = create_tool
expect(tool.signature).to eq(
{
name: tool.tool_name,
description: "test",
parameters: [{ name: "query", type: "string", description: "perform a search" }],
},
)
runner = tool.runner({ "query" => "test" }, llm: nil, bot_user: nil)
expect(runner.invoke).to eq("query" => "test")
end
it "can perform HTTP requests with various verbs" do
%i[post put delete patch].each do |verb|
script = <<~JS
function invoke(params) {
result = http.#{verb}("https://example.com/api",
{
headers: { TestHeader: "TestValue" },
body: JSON.stringify({ data: params.data })
}
);
return result.body;
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "data" => "test data" }, llm: nil, bot_user: nil)
stub_request(verb, "https://example.com/api").with(
body: "{\"data\":\"test data\"}",
headers: {
"Accept" => "*/*",
"Testheader" => "TestValue",
"User-Agent" => "Discourse AI Bot 1.0 (https://www.discourse.org)",
},
).to_return(status: 200, body: "Success", headers: {})
result = runner.invoke
expect(result).to eq("Success")
end
end
it "can perform GET HTTP requests, with 1 param" do
script = <<~JS
function invoke(params) {
result = http.get("https://example.com/" + params.query);
return result.body;
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "query" => "test" }, llm: nil, bot_user: nil)
stub_request(:get, "https://example.com/test").with(
headers: {
"Accept" => "*/*",
"User-Agent" => "Discourse AI Bot 1.0 (https://www.discourse.org)",
},
).to_return(status: 200, body: "Hello World", headers: {})
result = runner.invoke
expect(result).to eq("Hello World")
end
it "is limited to MAX http requests" do
script = <<~JS
function invoke(params) {
let i = 0;
while (i < 21) {
http.get("https://example.com/");
i += 1;
}
return "will not happen";
}
JS
tool = create_tool(script: script)
runner = tool.runner({}, llm: nil, bot_user: nil)
stub_request(:get, "https://example.com/").to_return(
status: 200,
body: "Hello World",
headers: {
},
)
expect { runner.invoke }.to raise_error(DiscourseAi::Personas::ToolRunner::TooManyRequestsError)
end
it "can perform GET HTTP requests" do
script = <<~JS
function invoke(params) {
result = http.get("https://example.com/" + params.query,
{ headers: { TestHeader: "TestValue" } }
);
return result.body;
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "query" => "test" }, llm: nil, bot_user: nil)
stub_request(:get, "https://example.com/test").with(
headers: {
"Accept" => "*/*",
"Testheader" => "TestValue",
"User-Agent" => "Discourse AI Bot 1.0 (https://www.discourse.org)",
},
).to_return(status: 200, body: "Hello World", headers: {})
result = runner.invoke
expect(result).to eq("Hello World")
end
it "will not timeout on slow HTTP reqs" do
script = <<~JS
function invoke(params) {
result = http.get("https://example.com/" + params.query,
{ headers: { TestHeader: "TestValue" } }
);
return result.body;
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "query" => "test" }, llm: nil, bot_user: nil)
stub_request(:get, "https://example.com/test").to_return do
sleep 0.01
{ status: 200, body: "Hello World", headers: {} }
end
tool = create_tool(script: script)
runner = tool.runner({ "query" => "test" }, llm: nil, bot_user: nil)
runner.timeout = 10
result = runner.invoke
expect(result).to eq("Hello World")
end
it "has access to llm truncation tools" do
script = <<~JS
function invoke(params) {
return llm.truncate("Hello World", 1);
}
JS
tool = create_tool(script: script)
runner = tool.runner({}, llm: llm, bot_user: nil)
result = runner.invoke
expect(result).to eq("Hello")
end
it "is able to run llm completions" do
script = <<~JS
function invoke(params) {
return llm.generate("question two") + llm.generate(
{ messages: [
{ type: "system", content: "system message" },
{ type: "user", content: "user message" }
]}
);
}
JS
tool = create_tool(script: script)
result = nil
prompts = nil
responses = ["Hello ", "World"]
DiscourseAi::Completions::Llm.with_prepared_responses(responses) do |_, _, _prompts|
runner = tool.runner({}, llm: llm, bot_user: nil)
result = runner.invoke
prompts = _prompts
end
prompt =
DiscourseAi::Completions::Prompt.new(
"system message",
messages: [{ type: :user, content: "user message" }],
)
expect(result).to eq("Hello World")
expect(prompts[0]).to eq("question two")
expect(prompts[1]).to eq(prompt)
end
it "can timeout slow JS" do
script = <<~JS
function invoke(params) {
while (true) {}
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "query" => "test" }, llm: nil, bot_user: nil)
runner.timeout = 5
result = runner.invoke
expect(result[:error]).to eq("Script terminated due to timeout")
end
context "when defining RAG fragments" do
fab!(:cloudflare_embedding_def)
before do
SiteSetting.authorized_extensions = "txt"
SiteSetting.ai_embeddings_selected_model = cloudflare_embedding_def.id
SiteSetting.ai_embeddings_enabled = true
Jobs.run_immediately!
end
def create_upload(content, filename)
upload = nil
Tempfile.create(filename) do |file|
file.write(content)
file.rewind
upload = UploadCreator.new(file, filename).create_for(Discourse.system_user.id)
end
upload
end
def stub_embeddings
# this is a trick, we get ever increasing embeddings, this gives us in turn
# 100% consistent search results
@counter = 0
stub_request(:post, cloudflare_embedding_def.url).to_return(
status: 200,
body: lambda { |req| { result: { data: [([@counter += 2] * 1024)] } }.to_json },
headers: {
},
)
end
it "allows search within uploads" do
stub_embeddings
upload1 = create_upload(<<~TXT, "test.txt")
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
TXT
upload2 = create_upload(<<~TXT, "test.txt")
30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
TXT
tool = create_tool(rag_chunk_tokens: 10, rag_chunk_overlap_tokens: 4, script: <<~JS)
function invoke(params) {
let result1 = index.search("testing a search", { limit: 1 });
let result2 = index.search("testing another search", { limit: 3, filenames: ["test.txt"] });
return [result1, result2];
}
JS
RagDocumentFragment.link_target_and_uploads(tool, [upload1.id, upload2.id])
result = tool.runner({}, llm: nil, bot_user: nil).invoke
expected = [
[{ "fragment" => "44 45 46 47 48 49 50", "metadata" => nil }],
[
{ "fragment" => "44 45 46 47 48 49 50", "metadata" => nil },
{ "fragment" => "36 37 38 39 40 41 42 43 44 45", "metadata" => nil },
{ "fragment" => "30 31 32 33 34 35 36 37", "metadata" => nil },
],
]
expect(result).to eq(expected)
# will force a reindex
tool.rag_chunk_tokens = 5
tool.rag_chunk_overlap_tokens = 2
tool.save!
# this part of the API is a bit awkward, maybe we should do it
# automatically
RagDocumentFragment.update_target_uploads(tool, [upload1.id, upload2.id])
result = tool.runner({}, llm: nil, bot_user: nil).invoke
# this is flaking, it is not critical cause it relies on vector search
# that may not be 100% deterministic
# expected = [
# [{ "fragment" => "48 49 50", "metadata" => nil }],
# [
# { "fragment" => "48 49 50", "metadata" => nil },
# { "fragment" => "45 46 47", "metadata" => nil },
# { "fragment" => "42 43 44", "metadata" => nil },
# ],
# ]
expect(result.length).to eq(2)
expect(result[0][0]["fragment"].length).to eq(8)
expect(result[1].length).to eq(3)
end
end
context "when using the topic API" do
it "can fetch topic details" do
script = <<~JS
function invoke(params) {
return discourse.getTopic(params.topic_id);
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "topic_id" => topic.id }, llm: nil, bot_user: nil)
result = runner.invoke
expect(result["id"]).to eq(topic.id)
expect(result["title"]).to eq(topic.title)
expect(result["archetype"]).to eq("regular")
expect(result["posts_count"]).to eq(1)
end
end
context "when using the post API" do
it "can fetch post details" do
script = <<~JS
function invoke(params) {
const post = discourse.getPost(params.post_id);
return {
post: post,
topic: post.topic
}
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "post_id" => post.id }, llm: nil, bot_user: nil)
result = runner.invoke
post_hash = result["post"]
topic_hash = result["topic"]
expect(post_hash["id"]).to eq(post.id)
expect(post_hash["topic_id"]).to eq(topic.id)
expect(post_hash["raw"]).to eq(post.raw)
expect(topic_hash["id"]).to eq(topic.id)
end
end
context "when using the search API" do
before { SearchIndexer.enable }
after { SearchIndexer.disable }
it "can perform a discourse search" do
SearchIndexer.index(topic, force: true)
SearchIndexer.index(post, force: true)
script = <<~JS
function invoke(params) {
return discourse.search({ search_query: params.query });
}
JS
tool = create_tool(script: script)
runner = tool.runner({ "query" => "banana" }, llm: nil, bot_user: nil)
result = runner.invoke
expect(result["rows"].length).to be > 0
expect(result["rows"].first["title"]).to eq(topic.title)
end
end
context "when using the chat API" do
before(:each) do
skip "Chat plugin tests skipped because Chat module is not defined." unless defined?(Chat)
SiteSetting.chat_enabled = true
end
fab!(:chat_user) { Fabricate(:user) }
fab!(:chat_channel) do
Fabricate(:chat_channel).tap do |channel|
Fabricate(
:user_chat_channel_membership,
user: chat_user,
chat_channel: channel,
following: true,
)
end
end
it "can create a chat message" do
script = <<~JS
function invoke(params) {
return discourse.createChatMessage({
channel_name: params.channel_name,
username: params.username,
message: params.message
});
}
JS
tool = create_tool(script: script)
runner =
tool.runner(
{
"channel_name" => chat_channel.name,
"username" => chat_user.username,
"message" => "Hello from the tool!",
},
llm: nil,
bot_user: bot_user, # The user *running* the tool doesn't affect sender
)
initial_message_count = Chat::Message.count
result = runner.invoke
expect(result["success"]).to eq(true), "Tool invocation failed: #{result["error"]}"
expect(result["message"]).to eq("Hello from the tool!")
expect(result["created_at"]).to be_present
expect(result).not_to have_key("error")
# Verify message was actually created in the database
expect(Chat::Message.count).to eq(initial_message_count + 1)
created_message = Chat::Message.find_by(id: result["message_id"])
expect(created_message).not_to be_nil
expect(created_message.message).to eq("Hello from the tool!")
expect(created_message.user_id).to eq(chat_user.id) # Message is sent AS the specified user
expect(created_message.chat_channel_id).to eq(chat_channel.id)
end
it "can create a chat message using channel slug" do
chat_channel.update!(name: "My Test Channel", slug: "my-test-channel")
expect(chat_channel.slug).to eq("my-test-channel")
script = <<~JS
function invoke(params) {
return discourse.createChatMessage({
channel_name: params.channel_slug, // Using slug here
username: params.username,
message: params.message
});
}
JS
tool = create_tool(script: script)
runner =
tool.runner(
{
"channel_slug" => chat_channel.slug,
"username" => chat_user.username,
"message" => "Hello via slug!",
},
llm: nil,
bot_user: bot_user,
)
result = runner.invoke
expect(result["success"]).to eq(true), "Tool invocation failed: #{result["error"]}"
# see: https://github.com/rubyjs/mini_racer/issues/348
# expect(result["message_id"]).to be_a(Integer)
created_message = Chat::Message.find_by(id: result["message_id"])
expect(created_message).not_to be_nil
expect(created_message.message).to eq("Hello via slug!")
expect(created_message.chat_channel_id).to eq(chat_channel.id)
end
it "returns an error if the channel is not found" do
script = <<~JS
function invoke(params) {
return discourse.createChatMessage({
channel_name: "non_existent_channel",
username: params.username,
message: params.message
});
}
JS
tool = create_tool(script: script)
runner =
tool.runner(
{ "username" => chat_user.username, "message" => "Test" },
llm: nil,
bot_user: bot_user,
)
initial_message_count = Chat::Message.count
expect { runner.invoke }.to raise_error(
MiniRacer::RuntimeError,
/Channel not found: non_existent_channel/,
)
expect(Chat::Message.count).to eq(initial_message_count) # Verify no message created
end
it "returns an error if the user is not found" do
script = <<~JS
function invoke(params) {
return discourse.createChatMessage({
channel_name: params.channel_name,
username: "non_existent_user",
message: params.message
});
}
JS
tool = create_tool(script: script)
runner =
tool.runner(
{ "channel_name" => chat_channel.name, "message" => "Test" },
llm: nil,
bot_user: bot_user,
)
initial_message_count = Chat::Message.count
expect { runner.invoke }.to raise_error(
MiniRacer::RuntimeError,
/User not found: non_existent_user/,
)
expect(Chat::Message.count).to eq(initial_message_count) # Verify no message created
end
end
context "when updating personas" do
fab!(:ai_persona) do
Fabricate(:ai_persona, name: "TestPersona", system_prompt: "Original prompt")
end
it "can update a persona with proper permissions" do
script = <<~JS
function invoke(params) {
return discourse.updatePersona(params.persona_name, {
system_prompt: params.new_prompt,
temperature: 0.7,
top_p: 0.9
});
}
JS
tool = create_tool(script: script)
runner =
tool.runner(
{ persona_name: "TestPersona", new_prompt: "Updated system prompt" },
llm: nil,
bot_user: bot_user,
)
result = runner.invoke
expect(result["success"]).to eq(true)
expect(result["persona"]["system_prompt"]).to eq("Updated system prompt")
expect(result["persona"]["temperature"]).to eq(0.7)
ai_persona.reload
expect(ai_persona.system_prompt).to eq("Updated system prompt")
expect(ai_persona.temperature).to eq(0.7)
expect(ai_persona.top_p).to eq(0.9)
end
end
context "when fetching persona information" do
fab!(:ai_persona) do
Fabricate(
:ai_persona,
name: "TestPersona",
description: "Test description",
system_prompt: "Test system prompt",
temperature: 0.8,
top_p: 0.9,
vision_enabled: true,
tools: ["Search", ["WebSearch", { param: "value" }, true]],
)
end
it "can fetch a persona by name" do
script = <<~JS
function invoke(params) {
const persona = discourse.getPersona(params.persona_name);
return persona;
}
JS
tool = create_tool(script: script)
runner = tool.runner({ persona_name: "TestPersona" }, llm: nil, bot_user: bot_user)
result = runner.invoke
expect(result["id"]).to eq(ai_persona.id)
expect(result["name"]).to eq("TestPersona")
expect(result["description"]).to eq("Test description")
expect(result["system_prompt"]).to eq("Test system prompt")
expect(result["temperature"]).to eq(0.8)
expect(result["top_p"]).to eq(0.9)
expect(result["vision_enabled"]).to eq(true)
expect(result["tools"]).to include("Search")
expect(result["tools"][1]).to be_a(Array)
end
it "raises an error when the persona doesn't exist" do
script = <<~JS
function invoke(params) {
return discourse.getPersona("NonExistentPersona");
}
JS
tool = create_tool(script: script)
runner = tool.runner({}, llm: nil, bot_user: bot_user)
expect { runner.invoke }.to raise_error(MiniRacer::RuntimeError, /Persona not found/)
end
it "can update a persona after fetching it" do
script = <<~JS
function invoke(params) {
const persona = discourse.getPersona("TestPersona");
return persona.update({
system_prompt: "Updated through getPersona().update()",
temperature: 0.5
});
}
JS
tool = create_tool(script: script)
runner = tool.runner({}, llm: nil, bot_user: bot_user)
result = runner.invoke
expect(result["success"]).to eq(true)
ai_persona.reload
expect(ai_persona.system_prompt).to eq("Updated through getPersona().update()")
expect(ai_persona.temperature).to eq(0.5)
end
end
end