Add tests
This commit is contained in:
Sean Goedecke
31
src/helpers.ts
Normal file
31
src/helpers.ts
Normal file
@@ -0,0 +1,31 @@
|
||||
import * as core from '@actions/core'
|
||||
import * as fs from 'fs'
|
||||
|
||||
/**
|
||||
* Helper function to load content from a file or use fallback input
|
||||
* @param filePathInput - Input name for the file path
|
||||
* @param contentInput - Input name for the direct content
|
||||
* @param defaultValue - Default value to use if neither file nor content is provided
|
||||
* @returns The loaded content
|
||||
*/
|
||||
export function loadContentFromFileOrInput(
|
||||
filePathInput: string,
|
||||
contentInput: string,
|
||||
defaultValue?: string
|
||||
): string {
|
||||
const filePath = core.getInput(filePathInput)
|
||||
const contentString = core.getInput(contentInput)
|
||||
|
||||
if (filePath !== undefined && filePath !== '') {
|
||||
if (!fs.existsSync(filePath)) {
|
||||
throw new Error(`File for ${filePathInput} was not found: ${filePath}`)
|
||||
}
|
||||
return fs.readFileSync(filePath, 'utf-8')
|
||||
} else if (contentString !== undefined && contentString !== '') {
|
||||
return contentString
|
||||
} else if (defaultValue !== undefined) {
|
||||
return defaultValue
|
||||
} else {
|
||||
throw new Error(`Neither ${filePathInput} nor ${contentInput} was set`)
|
||||
}
|
||||
}
|
||||
156
src/inference.ts
Normal file
156
src/inference.ts
Normal file
@@ -0,0 +1,156 @@
|
||||
import * as core from '@actions/core'
|
||||
import ModelClient, { isUnexpected } from '@azure-rest/ai-inference'
|
||||
import { AzureKeyCredential } from '@azure/core-auth'
|
||||
import { MCPClient, executeToolCalls } from './mcp.js'
|
||||
|
||||
export interface InferenceRequest {
|
||||
systemPrompt: string
|
||||
prompt: string
|
||||
modelName: string
|
||||
maxTokens: number
|
||||
endpoint: string
|
||||
token: string
|
||||
}
|
||||
|
||||
export interface InferenceResponse {
|
||||
content: string | null
|
||||
toolCalls?: any[]
|
||||
}
|
||||
|
||||
/**
|
||||
* Simple one-shot inference without tools
|
||||
*/
|
||||
export async function simpleInference(
|
||||
request: InferenceRequest
|
||||
): Promise<string | null> {
|
||||
core.info('Running simple inference without tools')
|
||||
|
||||
const client = ModelClient(
|
||||
request.endpoint,
|
||||
new AzureKeyCredential(request.token),
|
||||
{
|
||||
userAgentOptions: { userAgentPrefix: 'github-actions-ai-inference' }
|
||||
}
|
||||
)
|
||||
|
||||
const requestBody = {
|
||||
messages: [
|
||||
{
|
||||
role: 'system',
|
||||
content: request.systemPrompt
|
||||
},
|
||||
{ role: 'user', content: request.prompt }
|
||||
],
|
||||
max_tokens: request.maxTokens,
|
||||
model: request.modelName
|
||||
}
|
||||
|
||||
const response = await client.path('/chat/completions').post({
|
||||
body: requestBody
|
||||
})
|
||||
|
||||
if (isUnexpected(response)) {
|
||||
throw new Error(
|
||||
'An error occurred while fetching the response (' +
|
||||
response.status +
|
||||
'): ' +
|
||||
response.body
|
||||
)
|
||||
}
|
||||
|
||||
const modelResponse = response.body.choices[0].message.content
|
||||
core.info(`Model response: ${modelResponse || 'No response content'}`)
|
||||
|
||||
return modelResponse
|
||||
}
|
||||
|
||||
/**
|
||||
* MCP-enabled inference with tool execution loop
|
||||
*/
|
||||
export async function mcpInference(
|
||||
request: InferenceRequest,
|
||||
mcpClient: MCPClient
|
||||
): Promise<string | null> {
|
||||
core.info('Running MCP inference with tools')
|
||||
|
||||
const client = ModelClient(
|
||||
request.endpoint,
|
||||
new AzureKeyCredential(request.token),
|
||||
{
|
||||
userAgentOptions: { userAgentPrefix: 'github-actions-ai-inference' }
|
||||
}
|
||||
)
|
||||
|
||||
// Start with the initial conversation
|
||||
let messages: any[] = [
|
||||
{
|
||||
role: 'system',
|
||||
content: request.systemPrompt
|
||||
},
|
||||
{ role: 'user', content: request.prompt }
|
||||
]
|
||||
|
||||
let iterationCount = 0
|
||||
const maxIterations = 5 // Prevent infinite loops
|
||||
|
||||
while (iterationCount < maxIterations) {
|
||||
iterationCount++
|
||||
core.info(`MCP inference iteration ${iterationCount}`)
|
||||
|
||||
const requestBody = {
|
||||
messages: messages,
|
||||
max_tokens: request.maxTokens,
|
||||
model: request.modelName,
|
||||
tools: mcpClient.tools
|
||||
}
|
||||
|
||||
const response = await client.path('/chat/completions').post({
|
||||
body: requestBody
|
||||
})
|
||||
|
||||
if (isUnexpected(response)) {
|
||||
throw new Error(
|
||||
'An error occurred while fetching the response (' +
|
||||
response.status +
|
||||
'): ' +
|
||||
response.body
|
||||
)
|
||||
}
|
||||
|
||||
const assistantMessage = response.body.choices[0].message
|
||||
const modelResponse = assistantMessage.content
|
||||
const toolCalls = assistantMessage.tool_calls
|
||||
|
||||
core.info(`Model response: ${modelResponse || 'No response content'}`)
|
||||
|
||||
messages.push({
|
||||
role: 'assistant',
|
||||
content: modelResponse,
|
||||
...(toolCalls && { tool_calls: toolCalls })
|
||||
})
|
||||
|
||||
if (!toolCalls || toolCalls.length === 0) {
|
||||
core.info('No tool calls requested, ending MCP inference loop')
|
||||
return modelResponse
|
||||
}
|
||||
|
||||
core.info(`Model requested ${toolCalls.length} tool calls`)
|
||||
|
||||
const toolResults = await executeToolCalls(mcpClient.client, toolCalls)
|
||||
messages.push(...toolResults)
|
||||
|
||||
core.info('Tool results added, continuing conversation...')
|
||||
}
|
||||
|
||||
core.warning(
|
||||
`MCP inference loop exceeded maximum iterations (${maxIterations})`
|
||||
)
|
||||
|
||||
// Return the last assistant message content
|
||||
const lastAssistantMessage = messages
|
||||
.slice()
|
||||
.reverse()
|
||||
.find((msg) => msg.role === 'assistant')
|
||||
|
||||
return lastAssistantMessage?.content || null
|
||||
}
|
||||
241
src/main.ts
241
src/main.ts
@@ -1,43 +1,13 @@
|
||||
import * as core from '@actions/core'
|
||||
import ModelClient, { isUnexpected } from '@azure-rest/ai-inference'
|
||||
import { AzureKeyCredential } from '@azure/core-auth'
|
||||
import { Client } from '@modelcontextprotocol/sdk/client/index.js'
|
||||
import { StreamableHTTPClientTransport } from '@modelcontextprotocol/sdk/client/streamableHttp.js'
|
||||
import * as fs from 'fs'
|
||||
import * as os from 'os'
|
||||
import * as path from 'path'
|
||||
import { connectToMCP } from './mcp.js'
|
||||
import { simpleInference, mcpInference, InferenceRequest } from './inference.js'
|
||||
import { loadContentFromFileOrInput } from './helpers.js'
|
||||
|
||||
const RESPONSE_FILE = 'modelResponse.txt'
|
||||
|
||||
/**
|
||||
* Helper function to load content from a file or use fallback input
|
||||
* @param filePathInput - Input name for the file path
|
||||
* @param contentInput - Input name for the direct content
|
||||
* @param defaultValue - Default value to use if neither file nor content is provided
|
||||
* @returns The loaded content
|
||||
*/
|
||||
function loadContentFromFileOrInput(
|
||||
filePathInput: string,
|
||||
contentInput: string,
|
||||
defaultValue?: string
|
||||
): string {
|
||||
const filePath = core.getInput(filePathInput)
|
||||
const contentString = core.getInput(contentInput)
|
||||
|
||||
if (filePath !== undefined && filePath !== '') {
|
||||
if (!fs.existsSync(filePath)) {
|
||||
throw new Error(`File for ${filePathInput} was not found: ${filePath}`)
|
||||
}
|
||||
return fs.readFileSync(filePath, 'utf-8')
|
||||
} else if (contentString !== undefined && contentString !== '') {
|
||||
return contentString
|
||||
} else if (defaultValue !== undefined) {
|
||||
return defaultValue
|
||||
} else {
|
||||
throw new Error(`Neither ${filePathInput} nor ${contentInput} was set`)
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* The main function for the action.
|
||||
*
|
||||
@@ -45,10 +15,8 @@ function loadContentFromFileOrInput(
|
||||
*/
|
||||
export async function run(): Promise<void> {
|
||||
try {
|
||||
// Load prompt content - required
|
||||
const prompt = loadContentFromFileOrInput('prompt-file', 'prompt')
|
||||
|
||||
// Load system prompt with default value
|
||||
const systemPrompt = loadContentFromFileOrInput(
|
||||
'system-prompt-file',
|
||||
'system-prompt',
|
||||
@@ -64,200 +32,34 @@ export async function run(): Promise<void> {
|
||||
}
|
||||
|
||||
const endpoint = core.getInput('endpoint')
|
||||
|
||||
// Get MCP server configuration
|
||||
const mcpServerUrl = 'https://api.githubcopilot.com/mcp/'
|
||||
const enableMcp = core.getBooleanInput('enable-mcp') || false
|
||||
|
||||
let azureTools: any[] = []
|
||||
let mcp: Client | null = null
|
||||
const inferenceRequest: InferenceRequest = {
|
||||
systemPrompt,
|
||||
prompt,
|
||||
modelName,
|
||||
maxTokens,
|
||||
endpoint,
|
||||
token
|
||||
}
|
||||
|
||||
// Connect to MCP server if enabled
|
||||
if (enableMcp || true) {
|
||||
core.info('Connecting to GitHub MCP server...' + token)
|
||||
let modelResponse: string | null = null
|
||||
|
||||
const transport = new StreamableHTTPClientTransport(
|
||||
new URL(mcpServerUrl),
|
||||
{
|
||||
requestInit: {
|
||||
headers: {
|
||||
Authorization: `Bearer ${token}`
|
||||
}
|
||||
}
|
||||
}
|
||||
)
|
||||
if (enableMcp) {
|
||||
const mcpClient = await connectToMCP(token)
|
||||
|
||||
mcp = new Client({
|
||||
name: 'ai-inference-action',
|
||||
version: '1.0.0',
|
||||
transport
|
||||
})
|
||||
|
||||
try {
|
||||
await mcp.connect(transport)
|
||||
} catch (mcpError) {
|
||||
core.warning(`Failed to connect to MCP server: ${mcpError}`)
|
||||
// Continue without tools if MCP connection fails
|
||||
return
|
||||
if (mcpClient) {
|
||||
modelResponse = await mcpInference(inferenceRequest, mcpClient)
|
||||
} else {
|
||||
core.warning('MCP connection failed, falling back to simple inference')
|
||||
modelResponse = await simpleInference(inferenceRequest)
|
||||
}
|
||||
|
||||
core.info('Successfully connected to MCP server')
|
||||
|
||||
// Pull tool metadata
|
||||
const tools = await mcp.listTools()
|
||||
core.info(`Retrieved ${tools.tools?.length || 0} tools from MCP server`)
|
||||
|
||||
// Map MCP → Azure tool definitions
|
||||
azureTools = (tools.tools || []).map((t) => ({
|
||||
type: 'function',
|
||||
function: {
|
||||
name: t.name,
|
||||
description: t.description,
|
||||
parameters: t.inputSchema
|
||||
}
|
||||
}))
|
||||
|
||||
core.info(`Mapped ${azureTools.length} tools for Azure AI Inference`)
|
||||
} else {
|
||||
modelResponse = await simpleInference(inferenceRequest)
|
||||
}
|
||||
|
||||
const client = ModelClient(endpoint, new AzureKeyCredential(token), {
|
||||
userAgentOptions: { userAgentPrefix: 'github-actions-ai-inference' }
|
||||
})
|
||||
|
||||
const requestBody: any = {
|
||||
messages: [
|
||||
{
|
||||
role: 'system',
|
||||
content: systemPrompt
|
||||
},
|
||||
{ role: 'user', content: prompt }
|
||||
],
|
||||
max_tokens: maxTokens,
|
||||
model: modelName
|
||||
}
|
||||
|
||||
// Add tools if available
|
||||
if (azureTools.length > 0) {
|
||||
requestBody.tools = azureTools
|
||||
}
|
||||
|
||||
const response = await client.path('/chat/completions').post({
|
||||
body: requestBody
|
||||
})
|
||||
|
||||
if (isUnexpected(response)) {
|
||||
throw new Error(
|
||||
'An error occurred while fetching the response (' +
|
||||
response.status +
|
||||
'): ' +
|
||||
response.body
|
||||
)
|
||||
}
|
||||
|
||||
let modelResponse: string | null =
|
||||
response.body.choices[0].message.content
|
||||
|
||||
core.info(`Model response: ${response || 'No response content'}`)
|
||||
|
||||
// Handle tool calls if present
|
||||
const toolCalls = response.body.choices[0].message.tool_calls
|
||||
if (toolCalls && toolCalls.length > 0 && mcp) {
|
||||
core.info(`Model requested ${toolCalls.length} tool calls`)
|
||||
|
||||
// Execute tool calls via MCP and continue the conversation
|
||||
const toolResults: any[] = []
|
||||
|
||||
for (const toolCall of toolCalls) {
|
||||
core.info(
|
||||
`Executing tool: ${toolCall.function.name} with args: ${toolCall.function.arguments}`
|
||||
)
|
||||
|
||||
try {
|
||||
// Parse the arguments from JSON string
|
||||
const args = JSON.parse(toolCall.function.arguments)
|
||||
|
||||
// Call the tool via MCP
|
||||
const result = await mcp.callTool({
|
||||
name: toolCall.function.name,
|
||||
arguments: args
|
||||
})
|
||||
|
||||
core.info(`Tool ${toolCall.function.name} executed successfully`)
|
||||
|
||||
// Store the result for the follow-up conversation
|
||||
toolResults.push({
|
||||
tool_call_id: toolCall.id,
|
||||
role: 'tool',
|
||||
name: toolCall.function.name,
|
||||
content: JSON.stringify(result.content)
|
||||
})
|
||||
|
||||
} catch (toolError) {
|
||||
core.warning(`Failed to execute tool ${toolCall.function.name}: ${toolError}`)
|
||||
|
||||
// Add error result to continue conversation
|
||||
toolResults.push({
|
||||
tool_call_id: toolCall.id,
|
||||
role: 'tool',
|
||||
name: toolCall.function.name,
|
||||
content: `Error: ${toolError}`
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// If we have tool results, continue the conversation
|
||||
if (toolResults.length > 0) {
|
||||
core.info('Continuing conversation with tool results...')
|
||||
|
||||
// Build the follow-up request with the original conversation + tool results
|
||||
const followUpMessages = [
|
||||
{
|
||||
role: 'system',
|
||||
content: systemPrompt
|
||||
},
|
||||
{ role: 'user', content: prompt },
|
||||
{
|
||||
role: 'assistant',
|
||||
content: modelResponse,
|
||||
tool_calls: toolCalls
|
||||
},
|
||||
...toolResults
|
||||
]
|
||||
|
||||
const followUpRequest: any = {
|
||||
messages: followUpMessages,
|
||||
max_tokens: maxTokens,
|
||||
model: modelName
|
||||
}
|
||||
|
||||
// Add tools again for potential follow-up tool calls
|
||||
if (azureTools.length > 0) {
|
||||
followUpRequest.tools = azureTools
|
||||
}
|
||||
|
||||
const followUpResponse = await client.path('/chat/completions').post({
|
||||
body: followUpRequest
|
||||
})
|
||||
|
||||
if (isUnexpected(followUpResponse)) {
|
||||
core.warning(
|
||||
'Failed to get follow-up response after tool execution: ' +
|
||||
followUpResponse.status + ': ' + followUpResponse.body
|
||||
)
|
||||
} else {
|
||||
const finalResponse = followUpResponse.body.choices[0].message.content
|
||||
core.info(`Final response after tool execution: ${finalResponse}`)
|
||||
|
||||
// Update the model response to the final one
|
||||
modelResponse = finalResponse || modelResponse
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Set outputs for other workflow steps to use
|
||||
core.setOutput('response', modelResponse || '')
|
||||
|
||||
// Save the response to a file in case the response overflow the output limit
|
||||
const responseFilePath = path.join(tempDir(), RESPONSE_FILE)
|
||||
core.setOutput('response-file', responseFilePath)
|
||||
|
||||
@@ -265,7 +67,6 @@ export async function run(): Promise<void> {
|
||||
fs.writeFileSync(responseFilePath, modelResponse, 'utf-8')
|
||||
}
|
||||
} catch (error) {
|
||||
// Fail the workflow run if an error occurs
|
||||
if (error instanceof Error) {
|
||||
core.setFailed(error.message)
|
||||
} else {
|
||||
|
||||
129
src/mcp.ts
Normal file
129
src/mcp.ts
Normal file
@@ -0,0 +1,129 @@
|
||||
import * as core from '@actions/core'
|
||||
import { Client } from '@modelcontextprotocol/sdk/client/index.js'
|
||||
import { StreamableHTTPClientTransport } from '@modelcontextprotocol/sdk/client/streamableHttp.js'
|
||||
|
||||
export interface ToolResult {
|
||||
tool_call_id: string
|
||||
role: 'tool'
|
||||
name: string
|
||||
content: string
|
||||
}
|
||||
|
||||
export interface MCPClient {
|
||||
client: Client
|
||||
tools: any[]
|
||||
}
|
||||
|
||||
/**
|
||||
* Connect to the MCP server and retrieve available tools
|
||||
*/
|
||||
export async function connectToMCP(token: string): Promise<MCPClient | null> {
|
||||
const mcpServerUrl = 'https://api.githubcopilot.com/mcp/'
|
||||
|
||||
core.info('Connecting to GitHub MCP server...')
|
||||
|
||||
const transport = new StreamableHTTPClientTransport(new URL(mcpServerUrl), {
|
||||
requestInit: {
|
||||
headers: {
|
||||
Authorization: `Bearer ${token}`
|
||||
}
|
||||
}
|
||||
})
|
||||
|
||||
const client = new Client({
|
||||
name: 'ai-inference-action',
|
||||
version: '1.0.0',
|
||||
transport
|
||||
})
|
||||
|
||||
try {
|
||||
await client.connect(transport)
|
||||
} catch (mcpError) {
|
||||
core.warning(`Failed to connect to MCP server: ${mcpError}`)
|
||||
return null
|
||||
}
|
||||
|
||||
core.info('Successfully connected to MCP server')
|
||||
|
||||
// Pull tool metadata
|
||||
const toolsResponse = await client.listTools()
|
||||
core.info(
|
||||
`Retrieved ${toolsResponse.tools?.length || 0} tools from MCP server`
|
||||
)
|
||||
|
||||
// Map MCP → Azure tool definitions
|
||||
const tools = (toolsResponse.tools || []).map((t) => ({
|
||||
type: 'function',
|
||||
function: {
|
||||
name: t.name,
|
||||
description: t.description,
|
||||
parameters: t.inputSchema
|
||||
}
|
||||
}))
|
||||
|
||||
core.info(`Mapped ${tools.length} tools for Azure AI Inference`)
|
||||
|
||||
return { client, tools }
|
||||
}
|
||||
|
||||
/**
|
||||
* Execute a single tool call via MCP
|
||||
*/
|
||||
export async function executeToolCall(
|
||||
mcpClient: Client,
|
||||
toolCall: any
|
||||
): Promise<ToolResult> {
|
||||
core.info(
|
||||
`Executing tool: ${toolCall.function.name} with args: ${toolCall.function.arguments}`
|
||||
)
|
||||
|
||||
try {
|
||||
// Parse the arguments from JSON string
|
||||
const args = JSON.parse(toolCall.function.arguments)
|
||||
|
||||
// Call the tool via MCP
|
||||
const result = await mcpClient.callTool({
|
||||
name: toolCall.function.name,
|
||||
arguments: args
|
||||
})
|
||||
|
||||
core.info(`Tool ${toolCall.function.name} executed successfully`)
|
||||
|
||||
// Return the result formatted for the conversation
|
||||
return {
|
||||
tool_call_id: toolCall.id,
|
||||
role: 'tool',
|
||||
name: toolCall.function.name,
|
||||
content: JSON.stringify(result.content)
|
||||
}
|
||||
} catch (toolError) {
|
||||
core.warning(
|
||||
`Failed to execute tool ${toolCall.function.name}: ${toolError}`
|
||||
)
|
||||
|
||||
// Return error result to continue conversation
|
||||
return {
|
||||
tool_call_id: toolCall.id,
|
||||
role: 'tool',
|
||||
name: toolCall.function.name,
|
||||
content: `Error: ${toolError}`
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Execute all tool calls from a response
|
||||
*/
|
||||
export async function executeToolCalls(
|
||||
mcpClient: Client,
|
||||
toolCalls: any[]
|
||||
): Promise<ToolResult[]> {
|
||||
const toolResults: ToolResult[] = []
|
||||
|
||||
for (const toolCall of toolCalls) {
|
||||
const result = await executeToolCall(mcpClient, toolCall)
|
||||
toolResults.push(result)
|
||||
}
|
||||
|
||||
return toolResults
|
||||
}
|
||||
Reference in New Issue
Block a user