Update models and add check for assistants model on startup (#998)

* Remove old models from .env.template

* Make sure we always have a visible model in assistant settings

* Always show correct info in chat window

* Add migration that runs on every server start to update assistant model

* change key

* fix tests

* Update src/lib/migrations/migrations.ts

Co-authored-by: Mishig <[email protected]>

* review

* Update src/lib/components/chat/ChatWindow.svelte

Co-authored-by: Mishig <[email protected]>

* revert duplicate Model:

* simplify deprecated check

---------

Co-authored-by: Mishig <[email protected]>

.env.template

@@ -60,35 +60,6 @@ MODELS=`[
       }
   },
   {
-      "name": "meta-llama/Llama-2-70b-chat-hf",
-      "description": "The latest and biggest model from Meta, fine-tuned for chat.",
-      "logoUrl": "https://huggingface.co/datasets/huggingchat/models-logo/resolve/main/meta-logo.png",
-      "websiteUrl": "https://ai.meta.com/llama/",
-      "preprompt": "",
-      "chatPromptTemplate" : "<s>[INST] <<SYS>>\n{{preprompt}}\n<</SYS>>\n\n{{#each messages}}{{#ifUser}}{{content}} [/INST] {{/ifUser}}{{#ifAssistant}}{{content}} </s><s>[INST] {{/ifAssistant}}{{/each}}",
-      "promptExamples": [
-        {
-          "title": "Write an email from bullet list",
-          "prompt": "As a restaurant owner, write a professional email to the supplier to get these products every week: \n\n- Wine (x10)\n- Eggs (x24)\n- Bread (x12)"
-        }, {
-          "title": "Code a snake game",
-          "prompt": "Code a basic snake game in python, give explanations for each step."
-        }, {
-          "title": "Assist in a task",
-          "prompt": "How do I make a delicious lemon cheesecake?"
-        }
-      ],
-      "parameters": {
-        "temperature": 0.1,
-        "top_p": 0.95,
-        "repetition_penalty": 1.2,
-        "top_k": 50,
-        "truncate": 3072,
-        "max_new_tokens": 1024,
-        "stop" : ["</s>", "</s><s>[INST]"]
-      }
-    },
-    {
       "name" : "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
       "description" : "Nous Hermes 2 Mixtral 8x7B DPO is the new flagship Nous Research model trained over the Mixtral 8x7B MoE LLM.",
       "logoUrl": "https://huggingface.co/datasets/huggingchat/models-logo/resolve/main/nous-logo.png",
@@ -118,37 +89,6 @@ MODELS=`[
         "stop": ["<|im_end|>"]
       }
     },
-    {
-      "name": "codellama/CodeLlama-70b-Instruct-hf",
-      "displayName": "codellama/CodeLlama-70b-Instruct-hf",
-      "description": "Code Llama, a state of the art code model from Meta. Now in 70B!",
-      "logoUrl": "https://huggingface.co/datasets/huggingchat/models-logo/resolve/main/meta-logo.png",
-      "websiteUrl": "https://ai.meta.com/blog/code-llama-large-language-model-coding/",
-      "modelUrl": "https://huggingface.co/codellama/CodeLlama-70b-Instruct-hf",
-      "preprompt": "",
-      "chatPromptTemplate" : "<s>{{#if @root.preprompt}}Source: system\n\n {{@root.preprompt}} <step> {{/if}}{{#each messages}}{{#ifUser}}Source: user\n\n {{content}} <step> {{/ifUser}}{{#ifAssistant}}Source: assistant\n\n {{content}} <step> {{/ifAssistant}}{{/each}}Source: assistant\nDestination: user\n\n ",
-      "promptExamples": [
-        {
-          "title": "Fibonacci in Python",
-          "prompt": "Write a python function to calculate the nth fibonacci number."
-        }, {
-          "title": "JavaScript promises",
-          "prompt": "How can I wait for multiple JavaScript promises to fulfill before doing something with their values?"
-        }, {
-          "title": "Rust filesystem",
-          "prompt": "How can I load a file from disk in Rust?"
-        }
-      ],
-      "parameters": {
-        "temperature": 0.1,
-        "top_p": 0.95,
-        "repetition_penalty": 1.2,
-        "top_k": 50,
-        "truncate": 4096,
-        "max_new_tokens": 4096,
-        "stop": ["<step>", " <step>", " <step> "],
-      }
-      },
     {
       "name": "mistralai/Mistral-7B-Instruct-v0.1",
       "displayName": "mistralai/Mistral-7B-Instruct-v0.1",
@@ -214,38 +154,6 @@ MODELS=`[
         }
       ]
     },
-      {
-    "name": "openchat/openchat-3.5-0106",
-    "displayName": "openchat/openchat-3.5-0106",
-    "description": "OpenChat 3.5 is the #1 model on MT-Bench, with only 7B parameters.",
-    "logoUrl": "https://huggingface.co/datasets/huggingchat/models-logo/resolve/main/openchat-logo.png",
-    "websiteUrl": "https://huggingface.co/openchat/openchat-3.5-0106",
-    "modelUrl": "https://huggingface.co/openchat/openchat-3.5-0106",
-    "tokenizer": "openchat/openchat-3.5-0106",
-    "preprompt": "",
-    "chatPromptTemplate" : "<s>{{#each messages}}{{#ifUser}}GPT4 Correct User: {{#if @first}}{{#if @root.preprompt}}{{@root.preprompt}}\n{{/if}}{{/if}}{{content}}<|end_of_turn|>GPT4 Correct Assistant:{{/ifUser}}{{#ifAssistant}}{{content}}<|end_of_turn|>{{/ifAssistant}}{{/each}}",
-    "parameters": {
-      "temperature": 0.6,
-      "top_p": 0.95,
-      "repetition_penalty": 1.2,
-      "top_k": 50,
-      "truncate": 6016,
-      "max_new_tokens": 2048,
-      "stop": ["<|end_of_turn|>"]
-    },
-    "promptExamples": [
-      {
-        "title": "Write an email from bullet list",
-        "prompt": "As a restaurant owner, write a professional email to the supplier to get these products every week: \n\n- Wine (x10)\n- Eggs (x24)\n- Bread (x12)"
-      }, {
-        "title": "Code a snake game",
-        "prompt": "Code a basic snake game in python, give explanations for each step."
-      }, {
-        "title": "Assist in a task",
-        "prompt": "How do I make a delicious lemon cheesecake?"
-      }
-    ]
-  }
 ]`
 
 OLD_MODELS=`[
@@ -256,7 +164,10 @@ OLD_MODELS=`[
   {"name":"openchat/openchat-3.5-1210"},
   {"name": "tiiuae/falcon-180B-chat"},
   {"name": "codellama/CodeLlama-34b-Instruct-hf"},
-  {"name": "google/gemma-7b-it"}
+  {"name": "google/gemma-7b-it"},
+  {"name": "meta-llama/Llama-2-70b-chat-hf"},
+  {"name": "codellama/CodeLlama-70b-Instruct-hf"},
+  {"name": "openchat/openchat-3.5-0106"}
 ]`
 
 TASK_MODEL='mistralai/Mistral-7B-Instruct-v0.1'

src/lib/components/AssistantSettings.svelte

@@ -17,6 +17,7 @@
 	import IconInternet from "./icons/IconInternet.svelte";
 	import TokensCounter from "./TokensCounter.svelte";
 	import HoverTooltip from "./HoverTooltip.svelte";
+	import { findCurrentModel } from "$lib/utils/models";
 
 	type ActionData = {
 		error: boolean;
@@ -45,11 +46,7 @@
 		const module = await import("browser-image-resizer");
 		compress = module.readAndCompressImage;
 
-		if (assistant) {
-			modelId = assistant.modelId;
-		} else {
-			modelId = models.find((model) => model.id === $settings.activeModel)?.id ?? models[0].id;
-		}
+		modelId = findCurrentModel(models, assistant ? assistant.modelId : $settings.activeModel).id;
 	});
 
 	let inputMessage1 = assistant?.exampleInputs[0] ?? "";

src/lib/components/chat/ChatWindow.svelte

@@ -309,7 +309,9 @@
 							<ChatInput value="Sorry, something went wrong. Please try again." disabled={true} />
 						{:else}
 							<ChatInput
-								placeholder="Ask anything"
+								placeholder={isReadOnly
+									? "This conversation is read-only. Start a new one to continue!"
+									: "Ask anything"}
 								bind:value={message}
 								on:submit={handleSubmit}
 								on:beforeinput={(ev) => {
@@ -353,16 +355,33 @@
 				<p>
 					Model:
 					{#if !assistant}
-						<a href="{base}/settings/{currentModel.id}" class="hover:underline"
-							>{currentModel.displayName}</a
-						>{:else}
-						{@const model = models.find((m) => m.id === assistant?.modelId)}
-						<a
-							href="{base}/settings/assistants/{assistant._id}"
-							class="inline-flex items-center border-b hover:text-gray-600 dark:border-gray-700 dark:hover:text-gray-300"
-							>{model?.displayName}<CarbonCaretDown class="text-xxs" /></a
-						>{/if} <span class="max-sm:hidden">·</span><br class="sm:hidden" /> Generated content may
-					be inaccurate or false.
+						{#if models.find((m) => m.id === currentModel.id)}
+							<a
+								href="{base}/settings/{currentModel.id}"
+								class="inline-flex items-center hover:underline"
+								>{currentModel.displayName}<CarbonCaretDown class="text-xxs" /></a
+							>
+						{:else}
+							<span class="inline-flex items-center line-through dark:border-gray-700">
+								{currentModel.id}
+							</span>
+						{/if}
+					{:else}
+						{@const model = models.find((m) => m.id === currentModel.id)}
+						{#if model}
+							<a
+								href="{base}/settings/assistants/{assistant._id}"
+								class="inline-flex items-center border-b hover:text-gray-600 dark:border-gray-700 dark:hover:text-gray-300"
+								>{model?.displayName}<CarbonCaretDown class="text-xxs" /></a
+							>
+						{:else}
+							<span class="inline-flex items-center line-through dark:border-gray-700">
+								{currentModel.id}
+							</span>
+						{/if}
+					{/if}
+					<span class="max-sm:hidden">·</span><br class="sm:hidden" /> Generated content may be inaccurate
+					or false.
 				</p>
 				{#if messages.length}
 					<button

src/lib/migrations/migrations.spec.ts

@@ -3,7 +3,7 @@ import { migrations } from "./routines";
 import { acquireLock, isDBLocked, refreshLock, releaseLock } from "./lock";
 import { collections } from "$lib/server/database";
 
-const LOCK_KEY = "migrations";
+const LOCK_KEY = "migrations.test";
 
 describe("migrations", () => {
 	it("should not have duplicates guid", async () => {
@@ -21,7 +21,7 @@ describe("migrations", () => {
 		expect(locks.length).toBe(1);
 		expect(semaphores).toBeDefined();
 		expect(semaphores.length).toBe(1);
-		expect(semaphores?.[0].key).toBe("migrations");
+		expect(semaphores?.[0].key).toBe(LOCK_KEY);
 	});
 
 	it("should read the lock correctly", async () => {

src/lib/migrations/migrations.ts

@@ -14,14 +14,6 @@ export async function checkAndRunMigrations() {
 	// check if all migrations have already been run
 	const migrationResults = await collections.migrationResults.find().toArray();
 
-	// if all the migrations._id are in the migrationResults, we can exit early
-	if (
-		migrations.every((m) => migrationResults.some((m2) => m2._id.toString() === m._id.toString()))
-	) {
-		console.log("[MIGRATIONS] All migrations already applied.");
-		return;
-	}
-
 	console.log("[MIGRATIONS] Begin check...");
 
 	// connect to the database
@@ -52,12 +44,12 @@ export async function checkAndRunMigrations() {
 	// iterate over all migrations
 	for (const migration of migrations) {
 		// check if the migration has already been applied
-		const existingMigrationResult = migrationResults.find(
-			(m) => m._id.toString() === migration._id.toString()
-		);
+		const shouldRun =
+			migration.runEveryTime ||
+			!migrationResults.find((m) => m._id.toString() === migration._id.toString());
 
 		// check if the migration has already been applied
-		if (existingMigrationResult) {
+		if (!shouldRun) {
 			console.log(`[MIGRATIONS] "${migration.name}" already applied. Skipping...`);
 		} else {
 			// check the modifiers to see if some cases match
@@ -71,8 +63,12 @@ export async function checkAndRunMigrations() {
 				continue;
 			}
 
-			// otherwise all is good and we cna run the migration
-			console.log(`[MIGRATIONS] "${migration.name}" not applied yet. Applying...`);
+			// otherwise all is good and we can run the migration
+			console.log(
+				`[MIGRATIONS] "${migration.name}" ${
+					migration.runEveryTime ? "should run every time" : "not applied yet"
+				}. Applying...`
+			);
 
 			await collections.migrationResults.updateOne(
 				{ _id: migration._id },

src/lib/migrations/routines/02-update-assistants-models.ts

@@ -0,0 +1,28 @@
+import type { Migration } from ".";
+import { getCollections } from "$lib/server/database";
+import { ObjectId } from "mongodb";
+
+const updateAssistantsModels: Migration = {
+	_id: new ObjectId("5f9f3f3f3f3f3f3f3f3f3f3f"),
+	name: "Update deprecated models in assistants with the default model",
+	up: async (client) => {
+		const models = (await import("$lib/server/models")).models;
+
+		const { assistants } = getCollections(client);
+
+		const modelIds = models.map((el) => el.id); // string[]
+		const defaultModelId = models[0].id;
+
+		// Find all assistants whose modelId is not in modelIds, and update it to use defaultModelId
+		await assistants.updateMany(
+			{ modelId: { $nin: modelIds } },
+			{ $set: { modelId: defaultModelId } }
+		);
+
+		return true;
+	},
+	runEveryTime: true,
+	runForHuggingChat: "only",
+};
+
+export default updateAssistantsModels;

src/lib/migrations/routines/index.ts

@@ -1,6 +1,7 @@
 import type { MongoClient, ObjectId } from "mongodb";
 
 import updateSearchAssistant from "./01-update-search-assistants";
+import updateAssistantsModels from "./02-update-assistants-models";
 
 export interface Migration {
 	_id: ObjectId;
@@ -9,6 +10,7 @@ export interface Migration {
 	down?: (client: MongoClient) => Promise<boolean>;
 	runForFreshInstall?: "only" | "never"; // leave unspecified to run for both
 	runForHuggingChat?: "only" | "never"; // leave unspecified to run for both
+	runEveryTime?: boolean;
 }
 
-export const migrations: Migration[] = [updateSearchAssistant];
+export const migrations: Migration[] = [updateSearchAssistant, updateAssistantsModels];