Expose routes over OpenAPI

.github/workflows/lint-md.yml

@@ -30,7 +30,8 @@ jobs:
   lint-md:
     name: Linting Markdown
     runs-on: ubuntu-latest
-    needs: setup-node-modulessteps:
+    needs: setup-node-modules
+    steps:
       - name: Git Checkout
         uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 

.npmrc

@@ -0,0 +1 @@
+package-lock=true

CONTRIBUTING.md

@@ -38,12 +38,31 @@ Steps for downloading and setting up AI Warp for local development.
  is located at `ai-warp-app/platformatic.json`. **Note: this will be overwrited
  every time you generate the test app.**
 
- 8. Start the test app.
+ 8. Start the test app. From the `app-warp-ai` folder, run:
 
     ```bash
-    npm start
+    node ../dist/cli/start.js
     ```
 
+### Testing a local model with llama2
+
+To test a local model with with llama2, you can use the following to
+download the model we used for testing:
+
+```bash
+curl -L -O https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF/resolve/main/mistral-7b-instruct-v0.2.Q8_0.gguf
+```
+
+Then, in your `platformatic.json` file, add:
+
+```json
+  "aiProvider": {
+    "llama2": {
+      "modelPath": "./mistral-7b-instruct-v0.2.Q8_0.gguf"
+    }
+  },
+```
+
 ## Important Notes
 
 * AI Warp needs to be rebuilt for any code change to take affect in your test

index.ts

@@ -16,18 +16,18 @@ const stackable: Stackable<AiWarpConfig> = async function (fastify, opts) {
   await fastify.register(fastifyUser as any, config.auth)
   await fastify.register(authPlugin, opts)
 
-  await fastify.register(warpPlugin, opts) // needs to be registered here for fastify.ai to be decorated
-
-  await fastify.register(rateLimitPlugin, opts)
-  await fastify.register(apiPlugin, opts)
-
   if (config.showAiWarpHomepage !== undefined && config.showAiWarpHomepage) {
     await fastify.register(fastifyStatic, {
       root: join(import.meta.dirname, 'static')
     })
   }
 
   await fastify.register(platformaticService, opts)
+
+  await fastify.register(warpPlugin, opts) // needs to be registered here for fastify.ai to be decorated
+
+  await fastify.register(rateLimitPlugin, opts)
+  await fastify.register(apiPlugin, opts)
 }
 
 stackable.configType = 'ai-warp-app'

package.json

@@ -18,8 +18,8 @@
     "lint-md": "markdownlint-cli2 .",
     "lint-md:fix": "markdownlint-cli2 --fix .",
     "test": "npm run test:unit && npm run test:e2e && npm run test:types",
-    "test:unit": "node --test --test-reporter=@reporters/github --test-reporter-destination=stdout --test-reporter=spec --test-reporter-destination=stdout --import=tsx ./tests/unit/index.ts",
-    "test:e2e": "node --test --test-reporter=@reporters/github --test-reporter-destination=stdout --test-reporter=spec --test-reporter-destination=stdout --import=tsx ./tests/e2e/index.ts",
+    "test:unit": "node --test --test-reporter=@reporters/github --test-reporter-destination=stdout --test-reporter=spec --test-reporter-destination=stdout --import=tsx --test-concurrency=1 ./tests/unit/*",
+    "test:e2e": "node --test --test-reporter=@reporters/github --test-reporter-destination=stdout --test-reporter=spec --test-reporter-destination=stdout --import=tsx --test-concurrency=1 ./tests/e2e/*",
     "test:types": "tsd"
   },
   "engines": {

plugins/api.ts

@@ -23,7 +23,7 @@ const plugin: FastifyPluginAsyncTypebox = async (fastify) => {
           response: Type.String()
         }),
         default: Type.Object({
-          code: Type.String(),
+          code: Type.Optional(Type.String()),
           message: Type.String()
         })
       }

tests/e2e/api.test.ts

@@ -8,6 +8,8 @@ import { buildAiWarpApp } from '../utils/stackable.js'
 import { AZURE_DEPLOYMENT_NAME, AZURE_MOCK_HOST } from '../utils/mocks/azure.js'
 import { MOCK_CONTENT_RESPONSE, buildExpectedStreamBodyString } from '../utils/mocks/base.js'
 import { OLLAMA_MOCK_HOST } from '../utils/mocks/ollama.js'
+import { mockAllProviders } from '../utils/mocks/index.js'
+mockAllProviders()
 
 const expectedStreamBody = buildExpectedStreamBodyString()
 
@@ -108,6 +110,7 @@ for (const { name, config } of providers) {
           prompt: 'asd'
         })
       })
+
       assert.strictEqual(res.headers.get('content-type'), 'text/event-stream')
 
       assert.strictEqual(chunkCallbackCalled, true)
@@ -170,3 +173,59 @@ it('calls the preResponseCallback', async () => {
 
   await app.close()
 })
+
+it('provides all paths in OpenAPI', async () => {
+  const [app, port] = await buildAiWarpApp({
+    aiProvider: {
+      openai: {
+        model: 'gpt-3.5-turbo',
+        apiKey: ''
+      }
+    }
+  })
+
+  await app.start()
+
+  const res = await fetch(`http://localhost:${port}/documentation/json`)
+  const body = await res.json()
+
+  assert.deepStrictEqual(Object.keys(body.paths), [
+    '/api/v1/prompt',
+    '/api/v1/stream'
+  ])
+
+  await app.close()
+})
+
+it('prompt with wrong JSON', async () => {
+  const [app, port] = await buildAiWarpApp({
+    aiProvider: {
+      openai: {
+        model: 'gpt-3.5-turbo',
+        apiKey: ''
+      }
+    }
+  })
+
+  await app.start()
+
+  const res = await fetch(`http://localhost:${port}/api/v1/prompt`, {
+    method: 'POST',
+    headers: {
+      'content-type': 'application/json'
+    },
+    body: JSON.stringify({
+      prompt: 'asd'
+    }).slice(0, 10)
+  })
+
+  assert.strictEqual(res.status, 400)
+
+  const body = await res.json()
+
+  assert.deepStrictEqual(body, {
+    message: 'Unexpected end of JSON input'
+  })
+
+  await app.close()
+})

tests/e2e/auth.test.ts

@@ -4,6 +4,8 @@ import assert from 'node:assert'
 import { buildAiWarpApp } from '../utils/stackable.js'
 import { AiWarpConfig } from '../../config.js'
 import { authConfig, createToken } from '../utils/auth.js'
+import { mockAllProviders } from '../utils/mocks/index.js'
+mockAllProviders()
 
 const aiProvider: AiWarpConfig['aiProvider'] = {
   openai: {

tests/e2e/rate-limiting.test.ts

@@ -5,6 +5,8 @@ import fastifyPlugin from 'fastify-plugin'
 import { AiWarpConfig } from '../../config.js'
 import { buildAiWarpApp } from '../utils/stackable.js'
 import { authConfig, createToken } from '../utils/auth.js'
+import { mockAllProviders } from '../utils/mocks/index.js'
+mockAllProviders()
 
 const aiProvider: AiWarpConfig['aiProvider'] = {
   openai: {
@@ -28,7 +30,15 @@ it('calls ai.rateLimiting.max callback', async () => {
 
     await app.start()
 
-    const res = await fetch(`http://localhost:${port}`)
+    const res = await fetch(`http://localhost:${port}/api/v1/prompt`, {
+      method: 'POST',
+      headers: {
+        'content-type': 'application/json'
+      },
+      body: JSON.stringify({
+        prompt: 'asd'
+      })
+    })
     assert.strictEqual(callbackCalled, true)
     assert.strictEqual(res.headers.get('x-ratelimit-limit'), `${expectedMax}`)
   } finally {
@@ -50,7 +60,15 @@ it('calls ai.rateLimiting.allowList callback', async () => {
 
     await app.start()
 
-    await fetch(`http://localhost:${port}`)
+    await fetch(`http://localhost:${port}/api/v1/prompt`, {
+      method: 'POST',
+      headers: {
+        'content-type': 'application/json'
+      },
+      body: JSON.stringify({
+        prompt: 'asd'
+      })
+    })
     assert.strictEqual(callbackCalled, true)
   } finally {
     await app.close()
@@ -76,13 +94,21 @@ it('calls ai.rateLimiting.onBanReach callback', async () => {
 
       app.ai.rateLimiting.errorResponseBuilder = () => {
         errorResponseBuilderCalled = true
-        return { error: 'rate limited' }
+        return { message: 'rate limited' }
       }
     }))
 
     await app.start()
 
-    await fetch(`http://localhost:${port}`)
+    await fetch(`http://localhost:${port}/api/v1/prompt`, {
+      method: 'POST',
+      headers: {
+        'content-type': 'application/json'
+      },
+      body: JSON.stringify({
+        prompt: 'asd'
+      })
+    })
     assert.strictEqual(onBanReachCalled, true)
     assert.strictEqual(errorResponseBuilderCalled, true)
   } finally {
@@ -104,7 +130,15 @@ it('calls ai.rateLimiting.keyGenerator callback', async () => {
 
     await app.start()
 
-    await fetch(`http://localhost:${port}`)
+    await fetch(`http://localhost:${port}/api/v1/prompt`, {
+      method: 'POST',
+      headers: {
+        'content-type': 'application/json'
+      },
+      body: JSON.stringify({
+        prompt: 'asd'
+      })
+    })
     assert.strictEqual(callbackCalled, true)
   } finally {
     await app.close()
@@ -120,13 +154,21 @@ it('calls ai.rateLimiting.errorResponseBuilder callback', async () => {
       app.ai.rateLimiting.max = () => 0
       app.ai.rateLimiting.errorResponseBuilder = () => {
         callbackCalled = true
-        return { error: 'rate limited' }
+        return { message: 'rate limited' }
       }
     }))
 
     await app.start()
 
-    await fetch(`http://localhost:${port}`)
+    await fetch(`http://localhost:${port}/api/v1/prompt`, {
+      method: 'POST',
+      headers: {
+        'content-type': 'application/json'
+      },
+      body: JSON.stringify({
+        prompt: 'asd'
+      })
+    })
     assert.strictEqual(callbackCalled, true)
   } finally {
     await app.close()
@@ -156,17 +198,27 @@ it('uses the max for a specific claim', async () => {
   try {
     await app.start()
 
-    let res = await fetch(`http://localhost:${port}`, {
+    let res = await fetch(`http://localhost:${port}/api/v1/prompt`, {
+      method: 'POST',
       headers: {
-        Authorization: `Bearer ${createToken({ rateLimitMax: '10' })}`
-      }
+        Authorization: `Bearer ${createToken({ rateLimitMax: '10' })}`,
+        'content-type': 'application/json'
+      },
+      body: JSON.stringify({
+        prompt: 'asd'
+      })
     })
     assert.strictEqual(res.headers.get('x-ratelimit-limit'), '10')
 
-    res = await fetch(`http://localhost:${port}`, {
+    res = await fetch(`http://localhost:${port}/api/v1/prompt`, {
+      method: 'POST',
       headers: {
-        Authorization: `Bearer ${createToken({ rateLimitMax: '100' })}`
-      }
+        Authorization: `Bearer ${createToken({ rateLimitMax: '100' })}`,
+        'content-type': 'application/json'
+      },
+      body: JSON.stringify({
+        prompt: 'asd'
+      })
     })
     assert.strictEqual(res.headers.get('x-ratelimit-limit'), '100')
   } finally {

tests/unit/ai-providers.test.ts

@@ -10,6 +10,8 @@ import { MOCK_CONTENT_RESPONSE, buildExpectedStreamBodyString } from '../utils/m
 import { OLLAMA_MOCK_HOST } from '../utils/mocks/ollama.js'
 import { AZURE_DEPLOYMENT_NAME, AZURE_MOCK_HOST } from '../utils/mocks/azure.js'
 import { mockLlama2 } from '../utils/mocks/llama2.js'
+import { mockAllProviders } from '../utils/mocks/index.js'
+mockAllProviders()
 
 const expectedStreamBody = buildExpectedStreamBodyString()
 

tests/unit/generator.test.ts

@@ -7,6 +7,8 @@ import { join } from 'node:path'
 import AiWarpGenerator from '../../lib/generator.js'
 import { generateGlobalTypesFile } from '../../lib/templates/types.js'
 import { generatePluginWithTypesSupport } from '@platformatic/generators/lib/create-plugin.js'
+import { mockAllProviders } from '../utils/mocks/index.js'
+mockAllProviders()
 
 const tempDirBase = join(import.meta.dirname, 'tmp')
 

tests/utils/stackable.ts

@@ -21,13 +21,14 @@ export async function buildAiWarpApp (config: AiWarpConfig): Promise<[FastifyIns
     server: {
       port,
       forceCloseConnections: true,
-      healthCheck: {
-        enabled: false
-      },
+      healthCheck: false,
       logger: {
         level: 'silent'
       }
     },
+    service: {
+      openapi: true
+    },
     ...config
   }, stackable)