@langchain/google-genai 0.0.3 → 0.0.5

package/README.md

@@ -0,0 +1,163 @@
+# @langchain/google-genai
+
+This package contains the LangChain.js integrations for Gemini through their generative-ai SDK.
+
+## Installation
+
+```bash npm2yarn
+npm install @langchain/google-genai
+```
+
+This package, along with the main LangChain package, depends on [`@langchain/core`](https://npmjs.com/package/@langchain/core/).
+If you are using this package with other LangChain packages, you should make sure that all of the packages depend on the same instance of @langchain/core.
+You can do so by adding appropriate field to your project's `package.json` like this:
+
+```json
+{
+  "name": "your-project",
+  "version": "0.0.0",
+  "dependencies": {
+    "@langchain/google-genai": "^0.0.0",
+    "langchain": "0.0.207"
+  },
+  "resolutions": {
+    "@langchain/core": "0.1.2"
+  },
+  "overrides": {
+    "@langchain/core": "0.1.2"
+  },
+  "pnpm": {
+    "overrides": {
+      "@langchain/core": "0.1.2"
+    }
+  }
+}
+```
+
+The field you need depends on the package manager you're using, but we recommend adding a field for the common `yarn`, `npm`, and `pnpm` to maximize compatibility.
+
+## Chat Models
+
+This package contains the `ChatGoogleGenerativeAI` class, which is the recommended way to interface with the Google Gemini series of models.
+
+To use, install the requirements, and configure your environment.
+
+```bash
+export GOOGLE_API_KEY=your-api-key
+```
+
+Then initialize
+
+```typescript
+import { ChatGoogleGenerativeAI } from "@langchain/google-genai";
+
+const model = new ChatGoogleGenerativeAI({
+  modelName: "gemini-pro",
+  maxOutputTokens: 2048,
+});
+const response = await mode.invoke(new HumanMessage("Hello world!"));
+```
+
+#### Multimodal inputs
+
+Gemini vision model supports image inputs when providing a single chat message. Example:
+
+```bash npm2yarn
+npm install @langchain/core
+```
+
+```typescript
+import fs from "fs";
+import { ChatGoogleGenerativeAI } from "@langchain/google-genai";
+import { HumanMessage } from "@langchain/core/messages";
+
+const vision = new ChatGoogleGenerativeAI({
+  modelName: "gemini-pro-vision",
+  maxOutputTokens: 2048,
+});
+const image = fs.readFileSync("./hotdog.jpg").toString("base64");
+const input = [
+  new HumanMessage({
+    content: [
+      {
+        type: "text",
+        text: "Describe the following image.",
+      },
+      {
+        type: "image_url",
+        image_url: `data:image/png;base64,${image}`,
+      },
+    ],
+  }),
+];
+
+const res = await vision.invoke(input);
+```
+
+The value of `image_url` can be any of the following:
+
+- A public image URL
+- An accessible gcs file (e.g., "gcs://path/to/file.png")
+- A base64 encoded image (e.g., `data:image/png;base64,abcd124`)
+- A PIL image
+
+## Embeddings
+
+This package also adds support for google's embeddings models.
+
+```typescript
+import { GoogleGenerativeAIEmbeddings } from "@langchain/google-genai";
+import { TaskType } from "@google/generative-ai";
+
+const embeddings = new GoogleGenerativeAIEmbeddings({
+  modelName: "embedding-001", // 768 dimensions
+  taskType: TaskType.RETRIEVAL_DOCUMENT,
+  title: "Document title",
+});
+
+const res = await embeddings.embedQuery("OK Google");
+```
+
+## Development
+
+To develop the Google GenAI package, you'll need to follow these instructions:
+
+### Install dependencies
+
+```bash
+yarn install
+```
+
+### Build the package
+
+```bash
+yarn build
+```
+
+Or from the repo root:
+
+```bash
+yarn build --filter=@langchain/google-genai
+```
+
+### Run tests
+
+Test files should live within a `tests/` file in the `src/` folder. Unit tests should end in `.test.ts` and integration tests should
+end in `.int.test.ts`:
+
+```bash
+$ yarn test
+$ yarn test:int
+```
+
+### Lint & Format
+
+Run the linter & formatter to ensure your code is up to standard:
+
+```bash
+yarn lint && yarn format
+```
+
+### Adding new entrypoints
+
+If you add a new file to be exported, either import & re-export from `src/index.ts`, or add it to `scripts/create-entrypoints.js` and run `yarn build` to generate the new entrypoint.

package/dist/chat_models.cjs

@@ -129,6 +129,7 @@ class ChatGoogleGenerativeAI extends chat_models_1.BaseChatModel {
         if (this.topK && this.topK < 0) {
             throw new Error("`topK` must be a positive integer");
         }
+        this.stopSequences = fields?.stopSequences ?? this.stopSequences;
         this.apiKey = fields?.apiKey ?? (0, env_1.getEnvironmentVariable)("GOOGLE_API_KEY");
         if (!this.apiKey) {
             throw new Error("Please set an API key for Google GenerativeAI " +
@@ -162,7 +163,7 @@ class ChatGoogleGenerativeAI extends chat_models_1.BaseChatModel {
     _llmType() {
         return "googlegenerativeai";
     }
-    async _generate(messages, options, _runManager) {
+    async _generate(messages, options, runManager) {
         const prompt = (0, utils_js_1.convertBaseMessagesToContent)(messages, this._isMultimodalModel);
         const res = await this.caller.callWithOptions({ signal: options?.signal }, async () => {
             let output;
@@ -181,9 +182,11 @@ class ChatGoogleGenerativeAI extends chat_models_1.BaseChatModel {
             }
             return output;
         });
-        return (0, utils_js_1.mapGenerateContentResultToChatResult)(res.response);
+        const generationResult = (0, utils_js_1.mapGenerateContentResultToChatResult)(res.response);
+        await runManager?.handleLLMNewToken(generationResult.generations[0].text ?? "");
+        return generationResult;
     }
-    async *_streamResponseChunks(messages, options, _runManager) {
+    async *_streamResponseChunks(messages, options, runManager) {
         const prompt = (0, utils_js_1.convertBaseMessagesToContent)(messages, this._isMultimodalModel);
         const stream = await this.caller.callWithOptions({ signal: options?.signal }, async () => {
             const { stream } = await this.client.generateContentStream({
@@ -197,6 +200,7 @@ class ChatGoogleGenerativeAI extends chat_models_1.BaseChatModel {
                 continue;
             }
             yield chunk;
+            await runManager?.handleLLMNewToken(chunk.text ?? "");
         }
     }
 }

package/dist/chat_models.d.ts

@@ -125,6 +125,6 @@ export declare class ChatGoogleGenerativeAI extends BaseChatModel implements Goo
     constructor(fields?: GoogleGenerativeAIChatInput);
     _combineLLMOutput(): never[];
     _llmType(): string;
-    _generate(messages: BaseMessage[], options: this["ParsedCallOptions"], _runManager?: CallbackManagerForLLMRun): Promise<ChatResult>;
-    _streamResponseChunks(messages: BaseMessage[], options: this["ParsedCallOptions"], _runManager?: CallbackManagerForLLMRun): AsyncGenerator<ChatGenerationChunk>;
+    _generate(messages: BaseMessage[], options: this["ParsedCallOptions"], runManager?: CallbackManagerForLLMRun): Promise<ChatResult>;
+    _streamResponseChunks(messages: BaseMessage[], options: this["ParsedCallOptions"], runManager?: CallbackManagerForLLMRun): AsyncGenerator<ChatGenerationChunk>;
 }

package/dist/chat_models.js

@@ -126,6 +126,7 @@ export class ChatGoogleGenerativeAI extends BaseChatModel {
         if (this.topK && this.topK < 0) {
             throw new Error("`topK` must be a positive integer");
         }
+        this.stopSequences = fields?.stopSequences ?? this.stopSequences;
         this.apiKey = fields?.apiKey ?? getEnvironmentVariable("GOOGLE_API_KEY");
         if (!this.apiKey) {
             throw new Error("Please set an API key for Google GenerativeAI " +
@@ -159,7 +160,7 @@ export class ChatGoogleGenerativeAI extends BaseChatModel {
     _llmType() {
         return "googlegenerativeai";
     }
-    async _generate(messages, options, _runManager) {
+    async _generate(messages, options, runManager) {
         const prompt = convertBaseMessagesToContent(messages, this._isMultimodalModel);
         const res = await this.caller.callWithOptions({ signal: options?.signal }, async () => {
             let output;
@@ -178,9 +179,11 @@ export class ChatGoogleGenerativeAI extends BaseChatModel {
             }
             return output;
         });
-        return mapGenerateContentResultToChatResult(res.response);
+        const generationResult = mapGenerateContentResultToChatResult(res.response);
+        await runManager?.handleLLMNewToken(generationResult.generations[0].text ?? "");
+        return generationResult;
     }
-    async *_streamResponseChunks(messages, options, _runManager) {
+    async *_streamResponseChunks(messages, options, runManager) {
         const prompt = convertBaseMessagesToContent(messages, this._isMultimodalModel);
         const stream = await this.caller.callWithOptions({ signal: options?.signal }, async () => {
             const { stream } = await this.client.generateContentStream({
@@ -194,6 +197,7 @@ export class ChatGoogleGenerativeAI extends BaseChatModel {
                 continue;
             }
             yield chunk;
+            await runManager?.handleLLMNewToken(chunk.text ?? "");
         }
     }
 }

package/dist/utils.cjs

@@ -128,7 +128,7 @@ function mapGenerateContentResultToChatResult(response) {
         message: new messages_1.AIMessage({
             content: text,
             name: content === null ? undefined : content.role,
-            additional_kwargs: {},
+            additional_kwargs: generationInfo,
         }),
         generationInfo,
     };
@@ -149,6 +149,8 @@ function convertResponseContentToChatGenerationChunk(response) {
         message: new messages_1.AIMessageChunk({
             content: text,
             name: content === null ? undefined : content.role,
+            // Each chunk can have unique "generationInfo", and merging strategy is unclear,
+            // so leave blank for now.
             additional_kwargs: {},
         }),
         generationInfo,

package/dist/utils.js

@@ -121,7 +121,7 @@ export function mapGenerateContentResultToChatResult(response) {
         message: new AIMessage({
             content: text,
             name: content === null ? undefined : content.role,
-            additional_kwargs: {},
+            additional_kwargs: generationInfo,
         }),
         generationInfo,
     };
@@ -141,6 +141,8 @@ export function convertResponseContentToChatGenerationChunk(response) {
         message: new AIMessageChunk({
             content: text,
             name: content === null ? undefined : content.role,
+            // Each chunk can have unique "generationInfo", and merging strategy is unclear,
+            // so leave blank for now.
             additional_kwargs: {},
         }),
         generationInfo,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@langchain/google-genai",
-  "version": "0.0.3",
+  "version": "0.0.5",
   "description": "Sample integration for LangChain.js",
   "type": "module",
   "engines": {
@@ -35,7 +35,7 @@
   "license": "MIT",
   "dependencies": {
     "@google/generative-ai": "^0.1.0",
-    "@langchain/core": "~0.1.1"
+    "@langchain/core": "~0.1.3"
   },
   "devDependencies": {
     "@jest/globals": "^29.5.0",