add an OpenAI-compatible provider as a generic Enterprise LLM adapter (…

…#3218)

Signed-off-by: Stephen Gutekanst <stephen@sourcegraph.com>

lib/shared/src/configuration.ts

@@ -37,6 +37,7 @@ export interface Configuration {
         | 'anthropic'
         | 'fireworks'
         | 'unstable-openai'
+        | 'experimental-openaicompatible'
         | 'experimental-ollama'
         | null
     autocompleteAdvancedModel: string | null

lib/shared/src/sourcegraph-api/completions/client.ts

@@ -94,8 +94,12 @@ export abstract class SourcegraphCompletionsClient {
         apiVersion: number,
         signal?: AbortSignal
     ): AsyncGenerator<CompletionGeneratorValue> {
-        // This is a technique to convert a function that takes callbacks to an async generator.
+        // Provide default stop sequence for starchat models.
+        if (!params.stopSequences && params?.model?.startsWith('openaicompatible/starchat')) {
+            params.stopSequences = ['<|end|>']
+        }
 
+        // This is a technique to convert a function that takes callbacks to an async generator.
         const values: Promise<CompletionGeneratorValue>[] = []
         let resolve: ((value: CompletionGeneratorValue) => void) | undefined
         values.push(

vscode/CHANGELOG.md

@@ -6,6 +6,7 @@ This is a log of all notable changes to Cody for VS Code. [Unreleased] changes a
 
 ### Added
 
+- Cody Enterprise users now have access to an `experimental-openaicompatible` which allows bringing your own LLM via any OpenAI-compatible API. For now, this is only supported with Starchat and specific configurations - but we continue to generalize this work to support more models and OpenAI-compatible endpoints. [pull/3218](https://github.com/sourcegraph/cody/pull/3218)
 - Edit/Chat: Cody now expands the selection to the nearest enclosing function, if available, before attempting to expand to the nearest enclosing block. [pull/3507](https://github.com/sourcegraph/cody/pull/3507)
 - Edit: New `cody.edit.preInstruction` configuration option for adding custom instruction at the end of all your requests. [pull/3542](https://github.com/sourcegraph/cody/pull/3542)
 - Edit: Add support for the new `cody.edit.preInstruction` setting. [pull/3542](https://github.com/sourcegraph/cody/pull/3542)

vscode/package.json

@@ -970,7 +970,7 @@
         "cody.autocomplete.advanced.provider": {
           "type": "string",
           "default": null,
-          "enum": [null, "anthropic", "fireworks", "unstable-openai", "experimental-ollama"],
+          "enum": [null, "anthropic", "fireworks", "unstable-openai", "experimental-ollama", "experimental-openaicompatible"],
           "markdownDescription": "The provider used for code autocomplete. Most providers other than `anthropic` require the `cody.autocomplete.advanced.serverEndpoint` and `cody.autocomplete.advanced.accessToken` settings to also be set. Check the Cody output channel for error messages if autocomplete is not working as expected."
         },
         "cody.autocomplete.advanced.serverEndpoint": {

vscode/src/completions/providers/create-provider.test.ts

@@ -87,6 +87,33 @@ describe('createProviderConfig', () => {
             expect(provider?.model).toBe('starcoder-hybrid')
         })
 
+        it('returns "experimental-openaicompatible" provider config and corresponding model if specified', async () => {
+            const provider = await createProviderConfig(
+                getVSCodeConfigurationWithAccessToken({
+                    autocompleteAdvancedProvider: 'experimental-openaicompatible',
+                    autocompleteAdvancedModel: 'starchat-16b-beta',
+                }),
+                dummyCodeCompletionsClient,
+                dummyAuthStatus
+            )
+            expect(provider?.identifier).toBe('experimental-openaicompatible')
+            expect(provider?.model).toBe('starchat-16b-beta')
+        })
+
+        it('returns "experimental-openaicompatible" provider config if specified in settings and default model', async () => {
+            const provider = await createProviderConfig(
+                getVSCodeConfigurationWithAccessToken({
+                    autocompleteAdvancedProvider: 'experimental-openaicompatible',
+                }),
+                dummyCodeCompletionsClient,
+                dummyAuthStatus
+            )
+            expect(provider?.identifier).toBe('experimental-openaicompatible')
+            // TODO(slimsag): make this default to starchat2 once added
+            // specifically just when using `experimental-openaicompatible`
+            expect(provider?.model).toBe('starcoder-hybrid')
+        })
+
         it('returns "openai" provider config if specified in VSCode settings; model is ignored', async () => {
             const provider = await createProviderConfig(
                 getVSCodeConfigurationWithAccessToken({

vscode/src/completions/providers/create-provider.ts

@@ -17,6 +17,7 @@ import {
     type FireworksOptions,
     createProviderConfig as createFireworksProviderConfig,
 } from './fireworks'
+import { createProviderConfig as createOpenAICompatibleProviderConfig } from './openaicompatible'
 import type { ProviderConfig } from './provider'
 import { createProviderConfig as createUnstableOpenAIProviderConfig } from './unstable-openai'
 
@@ -52,6 +53,15 @@ export async function createProviderConfig(
             case 'anthropic': {
                 return createAnthropicProviderConfig({ client, model })
             }
+            case 'experimental-openaicompatible': {
+                return createOpenAICompatibleProviderConfig({
+                    client,
+                    model: config.autocompleteAdvancedModel ?? model ?? null,
+                    timeouts: config.autocompleteTimeouts,
+                    authStatus,
+                    config,
+                })
+            }
             case 'experimental-ollama':
             case 'unstable-ollama': {
                 return createExperimentalOllamaProviderConfig(
@@ -102,6 +112,14 @@ export async function createProviderConfig(
                     authStatus,
                     config,
                 })
+            case 'experimental-openaicompatible':
+                return createOpenAICompatibleProviderConfig({
+                    client,
+                    timeouts: config.autocompleteTimeouts,
+                    model: model ?? null,
+                    authStatus,
+                    config,
+                })
             case 'aws-bedrock':
             case 'anthropic':
                 return createAnthropicProviderConfig({