Huge News!Announcing our $40M Series B led by Abstract Ventures.Learn More →
Socket
Sign inDemoInstall
Socket
Sign inDemoInstall

@google-cloud/vertexai

Package Overview
Dependencies
Maintainers
2
Versions
23
Alerts
File Explorer

Advanced tools

License
Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

@google-cloud/vertexai - npm Package Compare versions

Comparing version 1.1.0 to 1.2.0

build/src/functions/test/pre_fetch_processing_test.d.ts
build/src/functions/test/pre_fetch_processing_test.js
build/src/functions/test/pre_fetch_processing_test.js.map
build/src/models/util.d.ts
build/src/models/util.js
build/src/models/util.js.map
src/functions/test/pre_fetch_processing_test.ts
src/models/util.ts

2

.release-please-manifest.json
{
".": "1.1.0"
".": "1.2.0"
}

2

build/src/functions/count_tokens.d.ts

@@ -23,2 +23,2 @@ /**

*/
export declare function countTokens(location: string, project: string, publisherModelEndpoint: string, token: Promise<string | null | undefined>, request: CountTokensRequest, apiEndpoint?: string, requestOptions?: RequestOptions): Promise<CountTokensResponse>;
export declare function countTokens(location: string, resourcePath: string, token: Promise<string | null | undefined>, request: CountTokensRequest, apiEndpoint?: string, requestOptions?: RequestOptions): Promise<CountTokensResponse>;

8

build/src/functions/count_tokens.js

@@ -21,2 +21,3 @@ "use strict";

const errors_1 = require("../types/errors");
const constants = require("../util/constants");
const post_fetch_processing_1 = require("./post_fetch_processing");

@@ -29,8 +30,7 @@ const post_request_1 = require("./post_request");

*/
async function countTokens(location, project, publisherModelEndpoint, token, request, apiEndpoint, requestOptions) {
async function countTokens(location, resourcePath, token, request, apiEndpoint, requestOptions) {
const response = await (0, post_request_1.postRequest)({
region: location,
project: project,
resourcePath: publisherModelEndpoint,
resourceMethod: 'countTokens',
resourcePath: resourcePath,
resourceMethod: constants.COUNT_TOKENS_METHOD,
token: await token,

@@ -37,0 +37,0 @@ data: request,

4

build/src/functions/generate_content.d.ts

@@ -23,3 +23,3 @@ /**

import { GenerateContentRequest, GenerateContentResult, GenerationConfig, RequestOptions, SafetySetting, StreamGenerateContentResult, Tool } from '../types/content';
export declare function generateContent(location: string, project: string, publisherModelEndpoint: string, token: Promise<string | null | undefined>, request: GenerateContentRequest | string, apiEndpoint?: string, generationConfig?: GenerationConfig, safetySettings?: SafetySetting[], tools?: Tool[], requestOptions?: RequestOptions): Promise<GenerateContentResult>;
export declare function generateContent(location: string, resourcePath: string, token: Promise<string | null | undefined>, request: GenerateContentRequest | string, apiEndpoint?: string, generationConfig?: GenerationConfig, safetySettings?: SafetySetting[], tools?: Tool[], requestOptions?: RequestOptions): Promise<GenerateContentResult>;
/**

@@ -32,2 +32,2 @@ * Make an async stream request to generate content. The response will be

*/
export declare function generateContentStream(location: string, project: string, publisherModelEndpoint: string, token: Promise<string | null | undefined>, request: GenerateContentRequest | string, apiEndpoint?: string, generationConfig?: GenerationConfig, safetySettings?: SafetySetting[], tools?: Tool[], requestOptions?: RequestOptions): Promise<StreamGenerateContentResult>;
export declare function generateContentStream(location: string, resourcePath: string, token: Promise<string | null | undefined>, request: GenerateContentRequest | string, apiEndpoint?: string, generationConfig?: GenerationConfig, safetySettings?: SafetySetting[], tools?: Tool[], requestOptions?: RequestOptions): Promise<StreamGenerateContentResult>;

20

build/src/functions/generate_content.js

@@ -25,3 +25,3 @@ "use strict";

const pre_fetch_processing_1 = require("./pre_fetch_processing");
async function generateContent(location, project, publisherModelEndpoint, token, request, apiEndpoint, generationConfig, safetySettings, tools, requestOptions) {
async function generateContent(location, resourcePath, token, request, apiEndpoint, generationConfig, safetySettings, tools, requestOptions) {
var _a, _b, _c;

@@ -42,9 +42,9 @@ request = (0, pre_fetch_processing_1.formatContentRequest)(request, generationConfig, safetySettings);

region: location,
project: project,
resourcePath: publisherModelEndpoint,
resourcePath,
resourceMethod: constants.GENERATE_CONTENT_METHOD,
token: await token,
data: generateContentRequest,
apiEndpoint: apiEndpoint,
requestOptions: requestOptions,
apiEndpoint,
requestOptions,
apiVersion: (0, pre_fetch_processing_1.getApiVersion)(request),
}).catch(e => {

@@ -66,3 +66,3 @@ throw new errors_1.GoogleGenerativeAIError('exception posting request to model', e);

*/
async function generateContentStream(location, project, publisherModelEndpoint, token, request, apiEndpoint, generationConfig, safetySettings, tools, requestOptions) {
async function generateContentStream(location, resourcePath, token, request, apiEndpoint, generationConfig, safetySettings, tools, requestOptions) {
var _a, _b, _c;

@@ -83,9 +83,9 @@ request = (0, pre_fetch_processing_1.formatContentRequest)(request, generationConfig, safetySettings);

region: location,
project: project,
resourcePath: publisherModelEndpoint,
resourcePath,
resourceMethod: constants.STREAMING_GENERATE_CONTENT_METHOD,
token: await token,
data: generateContentRequest,
apiEndpoint: apiEndpoint,
requestOptions: requestOptions,
apiEndpoint,
requestOptions,
apiVersion: (0, pre_fetch_processing_1.getApiVersion)(request),
}).catch(e => {

@@ -92,0 +92,0 @@ throw new errors_1.GoogleGenerativeAIError('exception posting request', e);

9

build/src/functions/post_fetch_processing.js

@@ -234,2 +234,3 @@ "use strict";

groundingAttributions: [],
retrievalQueries: [],
};

@@ -246,2 +247,10 @@ const groundingMetadataAggregated = (_a = aggregatedCandidate.groundingMetadata) !== null && _a !== void 0 ? _a : emptyGroundingMetadata;

}
if (groundingMetadataChunk.retrievalQueries) {
groundingMetadataAggregated.retrievalQueries =
groundingMetadataAggregated.retrievalQueries.concat(groundingMetadataChunk.retrievalQueries);
}
if (groundingMetadataChunk.searchEntryPoint) {
groundingMetadataAggregated.searchEntryPoint =
groundingMetadataChunk.searchEntryPoint;
}
return groundingMetadataAggregated;

@@ -248,0 +257,0 @@ }

3

build/src/functions/post_request.d.ts

@@ -22,5 +22,4 @@ /**

*/
export declare function postRequest({ region, project, resourcePath, resourceMethod, token, data, apiEndpoint, requestOptions, apiVersion, }: {
export declare function postRequest({ region, resourcePath, resourceMethod, token, data, apiEndpoint, requestOptions, apiVersion, }: {
region: string;
project: string;
resourcePath: string;

@@ -27,0 +26,0 @@ resourceMethod: string;

4

build/src/functions/post_request.js

@@ -33,5 +33,5 @@ "use strict";

*/
async function postRequest({ region, project, resourcePath, resourceMethod, token, data, apiEndpoint, requestOptions, apiVersion = 'v1', }) {
async function postRequest({ region, resourcePath, resourceMethod, token, data, apiEndpoint, requestOptions, apiVersion = 'v1', }) {
const vertexBaseEndpoint = apiEndpoint !== null && apiEndpoint !== void 0 ? apiEndpoint : `${region}-${API_BASE_PATH}`;
let vertexEndpoint = `https://${vertexBaseEndpoint}/${apiVersion}/projects/${project}/locations/${region}/${resourcePath}:${resourceMethod}`;
let vertexEndpoint = `https://${vertexBaseEndpoint}/${apiVersion}/${resourcePath}:${resourceMethod}`;
// Use server sent events for streamGenerateContent

@@ -38,0 +38,0 @@ if (resourceMethod === constants.STREAMING_GENERATE_CONTENT_METHOD) {

3

build/src/functions/pre_fetch_processing.d.ts

@@ -21,1 +21,4 @@ /**

export declare function validateGenerationConfig(generationConfig: GenerationConfig): GenerationConfig;
export declare function getApiVersion(request: GenerateContentRequest): 'v1' | 'v1beta1';
export declare function hasVertexRagStore(request: GenerateContentRequest): boolean;
export declare function hasVertexAISearch(request: GenerateContentRequest): boolean;

36

build/src/functions/pre_fetch_processing.js

@@ -19,3 +19,4 @@ "use strict";

Object.defineProperty(exports, "__esModule", { value: true });
exports.validateGenerationConfig = exports.validateGenerateContentRequest = exports.formatContentRequest = void 0;
exports.hasVertexAISearch = exports.hasVertexRagStore = exports.getApiVersion = exports.validateGenerationConfig = exports.validateGenerateContentRequest = exports.formatContentRequest = void 0;
const errors_1 = require("../types/errors");
const constants = require("../util/constants");

@@ -48,2 +49,5 @@ function formatContentRequest(request, generationConfig, safetySettings) {

}
if (hasVertexAISearch(request) && hasVertexRagStore(request)) {
throw new errors_1.ClientError('Found both vertexAiSearch and vertexRagStore field are set in tool. Either set vertexAiSearch or vertexRagStore.');
}
}

@@ -60,2 +64,32 @@ exports.validateGenerateContentRequest = validateGenerateContentRequest;

exports.validateGenerationConfig = validateGenerationConfig;
function getApiVersion(request) {
return hasVertexRagStore(request) ? 'v1beta1' : 'v1';
}
exports.getApiVersion = getApiVersion;
function hasVertexRagStore(request) {
var _a;
for (const tool of (_a = request === null || request === void 0 ? void 0 : request.tools) !== null && _a !== void 0 ? _a : []) {
const retrieval = tool.retrieval;
if (!retrieval)
continue;
if (retrieval.vertexRagStore) {
return true;
}
}
return false;
}
exports.hasVertexRagStore = hasVertexRagStore;
function hasVertexAISearch(request) {
var _a;
for (const tool of (_a = request === null || request === void 0 ? void 0 : request.tools) !== null && _a !== void 0 ? _a : []) {
const retrieval = tool.retrieval;
if (!retrieval)
continue;
if (retrieval.vertexAiSearch) {
return true;
}
}
return false;
}
exports.hasVertexAISearch = hasVertexAISearch;
//# sourceMappingURL=pre_fetch_processing.js.map

99

build/src/functions/test/functions_test.js

@@ -25,5 +25,4 @@ "use strict";

const StreamFunctions = require("../post_fetch_processing");
const TEST_PROJECT = 'test-project';
const TEST_LOCATION = 'test-location';
const TEST_PUBLISHER_MODEL_ENDPOINT = 'test-publisher-model-endpoint';
const TEST_RESOURCE_PATH = 'test-resource-path';
const TEST_TOKEN = 'testtoken';

@@ -36,2 +35,8 @@ const TEST_TOKEN_PROMISE = Promise.resolve(TEST_TOKEN);

];
const CONTENTS = [
{
role: 'user',
parts: [{ text: 'What is the weater like in Boston?' }],
},
];
const TEST_USER_CHAT_MESSAGE_WITH_GCS_FILE = [

@@ -186,2 +191,7 @@ {

];
const TEST_TOOLS_WITH_RAG = [
{
retrieval: { vertexRagStore: { ragResources: [{ ragCorpus: 'ragCorpus' }] } },
},
];
const fetchResponseObj = {

@@ -218,3 +228,3 @@ status: 200,

fetchSpy = spyOn(global, 'fetch').and.resolveTo(response);
const resp = await (0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResponseBody);

@@ -228,3 +238,3 @@ });

});
await expectAsync((0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT, TEST_REQUEST_OPTIONS)).toBeRejected();
await expectAsync((0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT, TEST_REQUEST_OPTIONS)).toBeRejected();
expect(fetchSpy.calls.allArgs()[0][1].signal).toBeInstanceOf(AbortSignal);

@@ -245,3 +255,3 @@ });

spyOn(global, 'fetch').and.resolveTo(response);
await expectAsync((0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT)).toBeRejected();
await expectAsync((0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT)).toBeRejected();
});

@@ -261,3 +271,3 @@ it('throw ClientError when not OK and 4XX', async () => {

spyOn(global, 'fetch').and.resolveTo(response);
await expectAsync((0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT)).toBeRejected();
await expectAsync((0, count_tokens_1.countTokens)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT)).toBeRejected();
});

@@ -279,3 +289,3 @@ });

});
await expectAsync((0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT, TEST_GENERATION_CONFIG, TEST_SAFETY_SETTINGS, TEST_EMPTY_TOOLS, TEST_REQUEST_OPTIONS)).toBeRejected();
await expectAsync((0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT, TEST_GENERATION_CONFIG, TEST_SAFETY_SETTINGS, TEST_EMPTY_TOOLS, TEST_REQUEST_OPTIONS)).toBeRejected();
expect(fetchSpy.calls.allArgs()[0][1].signal).toBeInstanceOf(AbortSignal);

@@ -291,3 +301,3 @@ });

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResult);

@@ -300,3 +310,3 @@ });

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, TEST_CHAT_MESSAGE_TEXT, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, TEST_CHAT_MESSAGE_TEXT, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResult);

@@ -312,3 +322,3 @@ });

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResult);

@@ -320,3 +330,3 @@ });

};
await expectAsync((0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT)).toBeRejectedWithError(URIError);
await expectAsync((0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT)).toBeRejectedWithError(URIError);
});

@@ -333,3 +343,3 @@ it('returns a GenerateContentResponse when passed safetySettings and generationConfig', async () => {

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResult);

@@ -342,3 +352,3 @@ });

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_ENDPOINT_BASE_PATH);
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_ENDPOINT_BASE_PATH);
expect(fetchSpy.calls.allArgs()[0][0].toString()).toContain(TEST_ENDPOINT_BASE_PATH);

@@ -353,3 +363,3 @@ });

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, reqWithEmptyConfigs, TEST_API_ENDPOINT);
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, reqWithEmptyConfigs, TEST_API_ENDPOINT);
const requestArgs = fetchSpy.calls.allArgs()[0][1];

@@ -367,3 +377,3 @@ if (typeof requestArgs === 'object' && requestArgs) {

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, reqWithEmptyConfigs, TEST_API_ENDPOINT);
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, reqWithEmptyConfigs, TEST_API_ENDPOINT);
const requestArgs = fetchSpy.calls.allArgs()[0][1];

@@ -380,3 +390,3 @@ if (typeof requestArgs === 'object' && requestArgs) {

fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect((_a = resp.response.candidates[0].citationMetadata) === null || _a === void 0 ? void 0 : _a.citations.length).toEqual(TEST_MODEL_RESPONSE.candidates[0].citationMetadata.citations.length);

@@ -386,5 +396,3 @@ });

const req = {
contents: [
{ role: 'user', parts: [{ text: 'What is the weater like in Boston?' }] },
],
contents: CONTENTS,
tools: TEST_TOOLS_WITH_FUNCTION_DECLARATION,

@@ -396,3 +404,3 @@ };

fetchSpy.and.resolveTo(new Response(JSON.stringify(TEST_MODEL_RESPONSE_WITH_FUNCTION_CALL), fetchResponseObj));
const actualResult = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const actualResult = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(actualResult).toEqual(expectedResult);

@@ -407,8 +415,3 @@ expect(types_1.GenerateContentResponseHandler.getFunctionCallsFromCandidate(actualResult.response.candidates[0])).toHaveSize(1);

const req = {
contents: [
{
role: 'user',
parts: [{ text: 'What is the weater like in Boston?' }],
},
],
contents: CONTENTS,
tools: TEST_TOOLS_WITH_FUNCTION_DECLARATION,

@@ -420,3 +423,3 @@ };

fetchSpy.and.resolveTo(new Response(JSON.stringify(TEST_MODEL_RESPONSE_WITH_INVALID_DATA), fetchResponseObj));
const actualResult = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const actualResult = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(actualResult).toEqual(expectedResult);

@@ -427,14 +430,28 @@ expect(types_1.GenerateContentResponseHandler.getFunctionCallsFromCandidate((_a = actualResult.response.candidates) === null || _a === void 0 ? void 0 : _a[0])).toHaveSize(0);

const req = {
contents: [
{
role: 'user',
parts: [{ text: 'What is the weater like in Boston?' }],
},
],
contents: CONTENTS,
tools: TEST_TOOLS_WITH_FUNCTION_DECLARATION,
};
fetchSpy.and.resolveTo(new Response(JSON.stringify({}), fetchResponseObj));
const actualResult = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const actualResult = await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(actualResult.response.candidates).not.toBeDefined();
});
it('should use v1 apiVersion', async () => {
const request = {
contents: CONTENTS,
};
fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, request, TEST_API_ENDPOINT);
const vertexEndpoint = fetchSpy.calls.allArgs()[0][0];
expect(vertexEndpoint).toContain('/v1/');
});
it('should use v1beta1 apiVersion when set RAG in tools', async () => {
const request = {
contents: CONTENTS,
tools: TEST_TOOLS_WITH_RAG,
};
fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
await (0, generate_content_1.generateContent)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, request, TEST_API_ENDPOINT);
const vertexEndpoint = fetchSpy.calls.allArgs()[0][0];
expect(vertexEndpoint).toContain('/v1beta1/');
});
});

@@ -461,3 +478,3 @@ describe('generateContentStream', () => {

});
await expectAsync((0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT, TEST_GENERATION_CONFIG, TEST_SAFETY_SETTINGS, TEST_EMPTY_TOOLS, TEST_REQUEST_OPTIONS)).toBeRejected();
await expectAsync((0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT, TEST_GENERATION_CONFIG, TEST_SAFETY_SETTINGS, TEST_EMPTY_TOOLS, TEST_REQUEST_OPTIONS)).toBeRejected();
expect(fetchSpy.calls.allArgs()[0][1].signal).toBeInstanceOf(AbortSignal);

@@ -475,3 +492,3 @@ });

spyOn(StreamFunctions, 'processStream').and.resolveTo(expectedResult);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResult);

@@ -486,3 +503,3 @@ });

spyOn(StreamFunctions, 'processStream').and.resolveTo(expectedResult);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, TEST_API_ENDPOINT, TEST_CHAT_MESSAGE_TEXT);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, TEST_API_ENDPOINT, TEST_CHAT_MESSAGE_TEXT);
expect(resp).toEqual(expectedResult);

@@ -500,3 +517,3 @@ });

spyOn(StreamFunctions, 'processStream').and.resolveTo(expectedResult);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResult);

@@ -514,3 +531,3 @@ });

spyOn(StreamFunctions, 'processStream').and.resolveTo(expectedResult);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const resp = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(resp).toEqual(expectedResult);

@@ -531,3 +548,3 @@ });

spyOn(StreamFunctions, 'processStream').and.resolveTo(expectedStreamResult);
const result = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const result = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(result).toEqual(expectedStreamResult);

@@ -552,3 +569,3 @@ const response = await result.response;

spyOn(StreamFunctions, 'processStream').and.resolveTo(expectedStreamResult);
const actualResult = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_PROJECT, TEST_PUBLISHER_MODEL_ENDPOINT, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
const actualResult = await (0, generate_content_1.generateContentStream)(TEST_LOCATION, TEST_RESOURCE_PATH, TEST_TOKEN_PROMISE, req, TEST_API_ENDPOINT);
expect(actualResult).toEqual(expectedStreamResult);

@@ -555,0 +572,0 @@ const response = await actualResult.response;

13

build/src/functions/test/post_fetch_processing_test.js

@@ -32,4 +32,3 @@ "use strict";

const LOCATION = 'location';
const PROJECT = 'project';
const PUBLISHER_MODEL_ENDPOINT = 'publisher_model_endpoint';
const RESOURCE_PATH = 'RESOURCE_PATH';
const TOKEN = Promise.resolve('token');

@@ -61,3 +60,3 @@ const GENERATE_CONTENT_REQUEST = 'generate_content_request';

spyOn(global, 'fetch').and.resolveTo(fetchResult);
const actualResult = await (0, generate_content_1.generateContent)(LOCATION, PROJECT, PUBLISHER_MODEL_ENDPOINT, TOKEN, GENERATE_CONTENT_REQUEST);
const actualResult = await (0, generate_content_1.generateContent)(LOCATION, RESOURCE_PATH, TOKEN, GENERATE_CONTENT_REQUEST);
const actualResponse = actualResult.response;

@@ -70,3 +69,3 @@ expect(actualResponse).toEqual(test_data_1.UNARY_RESPONSE_1);

spyOn(global, 'fetch').and.resolveTo(fetchResult);
const actualResult = await (0, generate_content_1.generateContent)(LOCATION, PROJECT, PUBLISHER_MODEL_ENDPOINT, TOKEN, GENERATE_CONTENT_REQUEST);
const actualResult = await (0, generate_content_1.generateContent)(LOCATION, RESOURCE_PATH, TOKEN, GENERATE_CONTENT_REQUEST);
const actualResponse = actualResult.response;

@@ -77,3 +76,3 @@ expect(actualResponse).toEqual(expectedResult);

spyOn(global, 'fetch').and.resolveTo(new Response(JSON.stringify({}), fetchResponseObj));
const actualResult = await (0, generate_content_1.generateContent)(LOCATION, PROJECT, PUBLISHER_MODEL_ENDPOINT, TOKEN, GENERATE_CONTENT_REQUEST);
const actualResult = await (0, generate_content_1.generateContent)(LOCATION, RESOURCE_PATH, TOKEN, GENERATE_CONTENT_REQUEST);
const actualResponse = actualResult.response;

@@ -88,3 +87,3 @@ expect(actualResponse).toEqual({});

const processStreamSpy = spyOn(PostFetchFunctions, 'processStream');
await (0, generate_content_1.generateContentStream)(LOCATION, PROJECT, PUBLISHER_MODEL_ENDPOINT, TOKEN, GENERATE_CONTENT_REQUEST);
await (0, generate_content_1.generateContentStream)(LOCATION, RESOURCE_PATH, TOKEN, GENERATE_CONTENT_REQUEST);
const actualArg = processStreamSpy.calls.allArgs()[0][0];

@@ -100,3 +99,3 @@ expect(actualArg).toBeDefined();

spyOn(global, 'fetch').and.resolveTo(fetchResult);
const actualResponse = await (0, count_tokens_1.countTokens)(LOCATION, PROJECT, PUBLISHER_MODEL_ENDPOINT, TOKEN, COUNT_TOKEN_REQUEST);
const actualResponse = await (0, count_tokens_1.countTokens)(LOCATION, RESOURCE_PATH, TOKEN, COUNT_TOKEN_REQUEST);
expect(actualResponse).toEqual(test_data_1.COUNT_TOKENS_RESPONSE_1);

@@ -103,0 +102,0 @@ });

9

build/src/functions/test/post_request_test.js

@@ -6,3 +6,2 @@ "use strict";

const REGION = 'us-central1';
const PROJECT = 'project-id';
const RESOURCE_PATH = 'resource-path';

@@ -26,3 +25,2 @@ const RESOURCE_METHOD = 'resource-method';

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -46,3 +44,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -64,3 +61,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -84,3 +80,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -104,3 +99,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -122,3 +116,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -145,3 +138,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -171,3 +163,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -174,0 +165,0 @@ resourceMethod: RESOURCE_METHOD,

14

build/src/functions/test/test_data.js

@@ -308,2 +308,5 @@ "use strict";

],
searchEntryPoint: {
renderedContent: 'rendered content for later chunk for first candidate',
},
},

@@ -330,2 +333,5 @@ },

],
searchEntryPoint: {
renderedContent: 'rendered content for later chunk for second candidate',
},
},

@@ -415,2 +421,6 @@ },

],
retrievalQueries: [],
searchEntryPoint: {
renderedContent: 'rendered content for later chunk for first candidate',
},
},

@@ -486,2 +496,6 @@ finishReason: 'STOP',

],
retrievalQueries: [],
searchEntryPoint: {
renderedContent: 'rendered content for later chunk for second candidate',
},
},

@@ -488,0 +502,0 @@ finishReason: 'STOP',

4

build/src/models/chat_session.d.ts

@@ -31,3 +31,3 @@ /**

private sendStreamPromise;
private readonly publisherModelEndpoint;
private readonly resourcePath;
private readonly googleAuth;

@@ -111,3 +111,3 @@ protected readonly requestOptions?: RequestOptions;

private sendStreamPromise;
private readonly publisherModelEndpoint;
private readonly resourcePath;
private readonly googleAuth;

@@ -114,0 +114,0 @@ protected readonly requestOptions?: RequestOptions;

39

build/src/models/chat_session.js

@@ -20,5 +20,6 @@ "use strict";

exports.ChatSessionPreview = exports.ChatSession = void 0;
const util_1 = require("./util");
const generate_content_1 = require("../functions/generate_content");
const errors_1 = require("../types/errors");
const util_1 = require("../util");
const util_2 = require("../util");
/**

@@ -46,3 +47,3 @@ * The `ChatSession` class is used to make multiturn send message requests. You

this.googleAuth = request.googleAuth;
this.publisherModelEndpoint = request.publisherModelEndpoint;
this.resourcePath = request.resourcePath;
this.historyInternal = (_a = request.history) !== null && _a !== void 0 ? _a : [];

@@ -55,5 +56,4 @@ this.generationConfig = request.generationConfig;

if (request.systemInstruction) {
request.systemInstruction.role = util_1.constants.SYSTEM_ROLE;
this.systemInstruction = (0, util_1.formulateSystemInstructionIntoContent)(request.systemInstruction);
}
this.systemInstruction = request.systemInstruction;
}

@@ -67,3 +67,3 @@ /**

const tokenPromise = this.googleAuth.getAccessToken().catch(e => {
throw new errors_1.GoogleAuthError(util_1.constants.CREDENTIAL_ERROR_MESSAGE, e);
throw new errors_1.GoogleAuthError(util_2.constants.CREDENTIAL_ERROR_MESSAGE, e);
});

@@ -100,3 +100,3 @@ return tokenPromise;

};
const generateContentResult = await (0, generate_content_1.generateContent)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
const generateContentResult = await (0, generate_content_1.generateContent)(this.location, this.resourcePath, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
throw e;

@@ -157,7 +157,9 @@ });

};
const streamGenerateContentResultPromise = (0, generate_content_1.generateContentStream)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
const streamGenerateContentResultPromise = (0, generate_content_1.generateContentStream)(this.location, this.resourcePath, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
throw e;
});
this.sendStreamPromise = this.appendHistory(streamGenerateContentResultPromise, newContent).catch(e => {
throw new errors_1.GoogleGenerativeAIError('exception appending chat history', e);
// Errors from remote endpoint will be catchable by user from streamGenerateContentResultPromise
// Errors in appendHistory should not throw to cause user's programe exit with code 1
console.error(e);
});

@@ -190,3 +192,3 @@ return streamGenerateContentResultPromise;

this.googleAuth = request.googleAuth;
this.publisherModelEndpoint = request.publisherModelEndpoint;
this.resourcePath = request.resourcePath;
this.historyInternal = (_a = request.history) !== null && _a !== void 0 ? _a : [];

@@ -199,5 +201,4 @@ this.generationConfig = request.generationConfig;

if (request.systemInstruction) {
request.systemInstruction.role = util_1.constants.SYSTEM_ROLE;
this.systemInstruction = (0, util_1.formulateSystemInstructionIntoContent)(request.systemInstruction);
}
this.systemInstruction = request.systemInstruction;
}

@@ -210,3 +211,3 @@ /**

const tokenPromise = this.googleAuth.getAccessToken().catch(e => {
throw new errors_1.GoogleAuthError(util_1.constants.CREDENTIAL_ERROR_MESSAGE, e);
throw new errors_1.GoogleAuthError(util_2.constants.CREDENTIAL_ERROR_MESSAGE, e);
});

@@ -243,3 +244,3 @@ return tokenPromise;

};
const generateContentResult = await (0, generate_content_1.generateContent)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
const generateContentResult = await (0, generate_content_1.generateContent)(this.location, this.resourcePath, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
throw e;

@@ -300,6 +301,10 @@ });

};
const streamGenerateContentResultPromise = (0, generate_content_1.generateContentStream)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
const streamGenerateContentResultPromise = (0, generate_content_1.generateContentStream)(this.location, this.resourcePath, this.fetchToken(), generateContentrequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions).catch(e => {
throw e;
});
this.sendStreamPromise = this.appendHistory(streamGenerateContentResultPromise, newContent);
this.sendStreamPromise = this.appendHistory(streamGenerateContentResultPromise, newContent).catch(e => {
// Errors from remote endpoint will be catchable by user from streamGenerateContentResultPromise
// Errors in appendHistory should not throw to cause user's programe exit with code 1
console.error(e);
});
return streamGenerateContentResultPromise;

@@ -335,4 +340,4 @@ }

function assignRoleToPartsAndValidateSendMessageRequest(parts) {
const userContent = { role: util_1.constants.USER_ROLE, parts: [] };
const functionContent = { role: util_1.constants.USER_ROLE, parts: [] };
const userContent = { role: util_2.constants.USER_ROLE, parts: [] };
const functionContent = { role: util_2.constants.USER_ROLE, parts: [] };
let hasUserContent = false;

@@ -339,0 +344,0 @@ let hasFunctionContent = false;

2

build/src/models/generative_models.d.ts

@@ -36,2 +36,3 @@ /**

private readonly publisherModelEndpoint;
private readonly resourcePath;
private readonly apiEndpoint?;

@@ -155,2 +156,3 @@ /**

private readonly publisherModelEndpoint;
private readonly resourcePath;
private readonly apiEndpoint?;

@@ -157,0 +159,0 @@ /**

73

build/src/models/generative_models.js

@@ -20,6 +20,7 @@ "use strict";

exports.GenerativeModelPreview = exports.GenerativeModel = void 0;
const util_1 = require("./util");
const count_tokens_1 = require("../functions/count_tokens");
const generate_content_1 = require("../functions/generate_content");
const errors_1 = require("../types/errors");
const util_1 = require("../util");
const util_2 = require("../util");
const chat_session_1 = require("./chat_session");

@@ -49,11 +50,7 @@ /**

if (getGenerativeModelParams.systemInstruction) {
getGenerativeModelParams.systemInstruction.role = util_1.constants.SYSTEM_ROLE;
this.systemInstruction = (0, util_1.formulateSystemInstructionIntoContent)(getGenerativeModelParams.systemInstruction);
}
this.systemInstruction = getGenerativeModelParams.systemInstruction;
if (this.model.startsWith('models/')) {
this.publisherModelEndpoint = `publishers/google/${this.model}`;
}
else {
this.publisherModelEndpoint = `publishers/google/models/${this.model}`;
}
this.resourcePath = formulateResourcePathFromModel(this.model, this.project, this.location);
// publisherModelEndpoint is deprecated
this.publisherModelEndpoint = this.resourcePath;
}

@@ -67,3 +64,3 @@ /**

const tokenPromise = this.googleAuth.getAccessToken().catch(e => {
throw new errors_1.GoogleAuthError(util_1.constants.CREDENTIAL_ERROR_MESSAGE, e);
throw new errors_1.GoogleAuthError(util_2.constants.CREDENTIAL_ERROR_MESSAGE, e);
});

@@ -93,3 +90,3 @@ return tokenPromise;

const formulatedRequest = formulateSystemInstructionIntoGenerateContentRequest(request, this.systemInstruction);
return (0, generate_content_1.generateContent)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
return (0, generate_content_1.generateContent)(this.location, this.resourcePath, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
}

@@ -123,3 +120,3 @@ /**

const formulatedRequest = formulateSystemInstructionIntoGenerateContentRequest(request, this.systemInstruction);
return (0, generate_content_1.generateContentStream)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
return (0, generate_content_1.generateContentStream)(this.location, this.resourcePath, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
}

@@ -145,3 +142,3 @@ /**

async countTokens(request) {
return (0, count_tokens_1.countTokens)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), request, this.apiEndpoint, this.requestOptions);
return (0, count_tokens_1.countTokens)(this.location, this.resourcePath, this.fetchToken(), request, this.apiEndpoint, this.requestOptions);
}

@@ -179,3 +176,5 @@ /**

publisherModelEndpoint: this.publisherModelEndpoint,
resourcePath: this.resourcePath,
tools: this.tools,
systemInstruction: this.systemInstruction,
};

@@ -220,11 +219,7 @@ if (request) {

if (getGenerativeModelParams.systemInstruction) {
getGenerativeModelParams.systemInstruction.role = util_1.constants.SYSTEM_ROLE;
this.systemInstruction = (0, util_1.formulateSystemInstructionIntoContent)(getGenerativeModelParams.systemInstruction);
}
this.systemInstruction = getGenerativeModelParams.systemInstruction;
if (this.model.startsWith('models/')) {
this.publisherModelEndpoint = `publishers/google/${this.model}`;
}
else {
this.publisherModelEndpoint = `publishers/google/models/${this.model}`;
}
this.resourcePath = formulateResourcePathFromModel(this.model, this.project, this.location);
// publisherModelEndpoint is deprecated
this.publisherModelEndpoint = this.resourcePath;
}

@@ -238,3 +233,3 @@ /**

const tokenPromise = this.googleAuth.getAccessToken().catch(e => {
throw new errors_1.GoogleAuthError(util_1.constants.CREDENTIAL_ERROR_MESSAGE, e);
throw new errors_1.GoogleAuthError(util_2.constants.CREDENTIAL_ERROR_MESSAGE, e);
});

@@ -263,3 +258,3 @@ return tokenPromise;

const formulatedRequest = formulateSystemInstructionIntoGenerateContentRequest(request, this.systemInstruction);
return (0, generate_content_1.generateContent)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
return (0, generate_content_1.generateContent)(this.location, this.resourcePath, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
}

@@ -293,3 +288,3 @@ /**

const formulatedRequest = formulateSystemInstructionIntoGenerateContentRequest(request, this.systemInstruction);
return (0, generate_content_1.generateContentStream)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
return (0, generate_content_1.generateContentStream)(this.location, this.resourcePath, this.fetchToken(), formulatedRequest, this.apiEndpoint, this.generationConfig, this.safetySettings, this.tools, this.requestOptions);
}

@@ -315,3 +310,3 @@ /**

async countTokens(request) {
return (0, count_tokens_1.countTokens)(this.location, this.project, this.publisherModelEndpoint, this.fetchToken(), request, this.apiEndpoint, this.requestOptions);
return (0, count_tokens_1.countTokens)(this.location, this.resourcePath, this.fetchToken(), request, this.apiEndpoint, this.requestOptions);
}

@@ -349,3 +344,5 @@ /**

publisherModelEndpoint: this.publisherModelEndpoint,
resourcePath: this.resourcePath,
tools: this.tools,
systemInstruction: this.systemInstruction,
};

@@ -366,6 +363,28 @@ if (request) {

exports.GenerativeModelPreview = GenerativeModelPreview;
function formulateResourcePathFromModel(model, project, location) {
let resourcePath;
if (!model) {
throw new errors_1.ClientError('model parameter must not be empty.');
}
if (!model.includes('/')) {
// example 'gemini-1.0-pro'
resourcePath = `projects/${project}/locations/${location}/publishers/google/models/${model}`;
}
else if (model.startsWith('models/')) {
// example 'models/gemini-1.0-pro'
resourcePath = `projects/${project}/locations/${location}/publishers/google/${model}`;
}
else if (model.startsWith('projects/')) {
// example 'projects/my-project/locations/my-location/models/my-tuned-model'
resourcePath = model;
}
else {
throw new errors_1.ClientError('model parameter must be either a Model Garden model ID or a full resource name.');
}
return resourcePath;
}
function formulateRequestToGenerateContentRequest(request) {
if (typeof request === 'string') {
return {
contents: [{ role: util_1.constants.USER_ROLE, parts: [{ text: request }] }],
contents: [{ role: util_2.constants.USER_ROLE, parts: [{ text: request }] }],
};

@@ -377,3 +396,3 @@ }

if (methodRequest.systemInstruction) {
methodRequest.systemInstruction.role = util_1.constants.SYSTEM_ROLE;
methodRequest.systemInstruction = (0, util_1.formulateSystemInstructionIntoContent)(methodRequest.systemInstruction);
return methodRequest;

@@ -380,0 +399,0 @@ }

93

build/src/types/content.d.ts

@@ -24,5 +24,9 @@ /**

project: string;
/** The Google Cloud project location. */
location: string;
/**
* Optional. The Google Cloud project location. If not provided, SDK will
* firtly try to resolve it from run time environment. If no location resolved
* from run time environment, SDK will use default value `us-central1`.
*/
location?: string;
/**
* Optional. The base Vertex AI endpoint to use for the request. If not

@@ -47,6 +51,7 @@ * provided, the default regionalized endpoint (i.e.

contents: Content[];
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -104,6 +109,7 @@ /**

requestOptions?: RequestOptions;
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -131,6 +137,7 @@ /**

tools?: Tool[];
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -525,8 +532,16 @@ /**

export declare interface GoogleDate {
/** Year of the date. Must be from 1 to 9999, or 0 to specify a date without a year. */
/**
* Year of the date. Must be from 1 to 9999, or 0 to specify a date without a
* year.
*/
year?: number;
/** Month of the date. Must be from 1 to 12, or 0 to specify a year without a monthi and day. */
/**
* Month of the date. Must be from 1 to 12, or 0 to specify a year without a
* monthi and day.
*/
month?: number;
/** Day of the date. Must be from 1 to 31 and valid for the year and month.
* or 0 to specify a year by itself or a year and month where the day isn't significant
/**
* Day of the date. Must be from 1 to 31 and valid for the year and month.
* or 0 to specify a year by itself or a year and month where the day isn't
* significant
*/

@@ -553,2 +568,11 @@ day?: number;

/**
* Google search entry point.
*/
export declare interface SearchEntryPoint {
/** Optional. Web content snippet that can be embedded in a web page or an app webview. */
renderedContent?: string;
/** Optional. Base64 encoded JSON representing array of tuple. */
sdkBlob?: string;
}
/**
* A collection of grounding attributions for a piece of content.

@@ -563,2 +587,4 @@ */

groundingAttributions?: GroundingAttribution[];
/** Optional. Google search entry for the following-up web searches. {@link SearchEntryPoint} */
searchEntryPoint?: SearchEntryPoint;
}

@@ -704,3 +730,32 @@ /**

}
export declare interface VertexRagStore {
/**
* Optional. List of corpora for retrieval. Currently only support one corpus
* or multiple files from one corpus. In the future we may open up multiple
* corpora support.
*/
ragResources?: RagResource[];
/** Optional. Number of top k results to return from the selected corpora. */
similarityTopK?: number;
/** Optional. If set this field, results with vector distance smaller than this threshold will be returned. */
vectorDistanceThreshold?: number;
}
/**
* Config of Vertex RagStore grounding checking.
*/
export declare interface RagResource {
/**
* Optional. Vertex RAG Store corpus resource name.
*
* @example
* `projects/{project}/locations/{location}/ragCorpora/{rag_corpus}`
*/
ragCorpus?: string;
/**
* Optional. Set this field to select the files under the ragCorpora for
* retrieval.
*/
ragFileIds?: string[];
}
/**
* Defines a retrieval tool that model can call to access external knowledge.

@@ -723,2 +778,4 @@ */

vertexAiSearch?: VertexAISearch;
/** Optional. Set to use data source powered by Vertex RAG store. */
vertexRagStore?: VertexRagStore;
/**

@@ -832,6 +889,7 @@ * Optional. Disable using the result from this tool in detecting grounding

apiEndpoint?: string;
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -850,6 +908,9 @@ /**

publisherModelEndpoint: string;
/** Optional. The user provided system instructions for the model.
/** The resource path to use for the request. */
resourcePath: string;
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -856,0 +917,0 @@ /**

3

build/src/util/constants.d.ts

@@ -19,6 +19,7 @@ /**

export declare const STREAMING_GENERATE_CONTENT_METHOD = "streamGenerateContent";
export declare const COUNT_TOKENS_METHOD = "countTokens";
export declare const USER_ROLE = "user";
export declare const MODEL_ROLE = "model";
export declare const SYSTEM_ROLE = "system";
export declare const USER_AGENT = "model-builder/1.1.0 grpc-node/1.1.0";
export declare const USER_AGENT = "model-builder/1.2.0 grpc-node/1.2.0";
export declare const CREDENTIAL_ERROR_MESSAGE = "\nUnable to authenticate your request \nDepending on your run time environment, you can get authentication by \n- if in local instance or cloud shell: `!gcloud auth login` \n- if in Colab: \n -`from google.colab import auth` \n -`auth.authenticate_user()` \n- if in service account or other: please follow guidance in https://cloud.google.com/docs/authentication";

5

build/src/util/constants.js
"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.CREDENTIAL_ERROR_MESSAGE = exports.USER_AGENT = exports.SYSTEM_ROLE = exports.MODEL_ROLE = exports.USER_ROLE = exports.STREAMING_GENERATE_CONTENT_METHOD = exports.GENERATE_CONTENT_METHOD = void 0;
exports.CREDENTIAL_ERROR_MESSAGE = exports.USER_AGENT = exports.SYSTEM_ROLE = exports.MODEL_ROLE = exports.USER_ROLE = exports.COUNT_TOKENS_METHOD = exports.STREAMING_GENERATE_CONTENT_METHOD = exports.GENERATE_CONTENT_METHOD = void 0;
/**

@@ -22,2 +22,3 @@ * @license

exports.STREAMING_GENERATE_CONTENT_METHOD = 'streamGenerateContent';
exports.COUNT_TOKENS_METHOD = 'countTokens';
exports.USER_ROLE = 'user';

@@ -27,3 +28,3 @@ exports.MODEL_ROLE = 'model';

const USER_AGENT_PRODUCT = 'model-builder';
const CLIENT_LIBRARY_VERSION = '1.1.0'; // x-release-please-version
const CLIENT_LIBRARY_VERSION = '1.2.0'; // x-release-please-version
const CLIENT_LIBRARY_LANGUAGE = `grpc-node/${CLIENT_LIBRARY_VERSION}`;

@@ -30,0 +31,0 @@ exports.USER_AGENT = `${USER_AGENT_PRODUCT}/${CLIENT_LIBRARY_VERSION} ${CLIENT_LIBRARY_LANGUAGE}`;

12

build/src/vertex_ai.js

@@ -40,3 +40,3 @@ "use strict";

this.project = init.project;
this.location = init.location;
this.location = resolveLocation(init.location);
this.googleAuth = new google_auth_library_1.GoogleAuth(opts);

@@ -178,2 +178,12 @@ this.apiEndpoint = init.apiEndpoint;

}
function resolveLocation(locationFromInput) {
if (locationFromInput) {
return locationFromInput;
}
const inferredLocation = process.env['GOOGLE_CLOUD_REGION'] || process.env['CLOUD_ML_REGION'];
if (inferredLocation) {
return inferredLocation;
}
return 'us-central1';
}
//# sourceMappingURL=vertex_ai.js.map

74

build/system_test/end_to_end_sample_test.js

@@ -22,3 +22,3 @@ "use strict";

const types_1 = require("../src/types");
const PROJECT = process.env.GCLOUD_PROJECT;
const PROJECT = process.env['GCLOUD_PROJECT'];
const LOCATION = 'us-central1';

@@ -79,2 +79,15 @@ const TEXT_REQUEST = {

];
const TOOLS_WITH_RAG = [
{
retrieval: {
vertexRagStore: {
ragResources: [
{
ragCorpus: 'projects/ucaip-sample-tests/locations/us-central1/ragCorpora/6917529027641081856',
},
],
},
},
},
];
const WEATHER_FORECAST = 'super nice';

@@ -259,3 +272,3 @@ const FUNCTION_RESPONSE_PART = [

});
it('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
var _a;

@@ -277,3 +290,3 @@ const request = {

});
it('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
var _a;

@@ -295,3 +308,3 @@ const request = {

});
it('should return a FunctionCall when passed a FunctionDeclaration', async () => {
xit('should return a FunctionCall when passed a FunctionDeclaration', async () => {
var _a;

@@ -314,3 +327,3 @@ const request = {

});
it('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
xit('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
var _a;

@@ -397,2 +410,22 @@ const request = {

});
it('in preview should return grounding metadata when passed a VertexRagStore', async () => {
var _a, _b;
const request = {
contents: [
{
role: 'user',
parts: [
{
text: 'How much gain or loss did Google get in the Motorola Mobile deal in 2014?',
},
],
},
],
tools: TOOLS_WITH_RAG,
};
const result = await generativeTextModelPreview.generateContentStream(request);
const response = await result.response;
expect(response.candidates[0]).toBeTruthy(`sys test failure on generateContent with RAG tool, for resp ${JSON.stringify(response)}`);
expect((_b = (_a = response.candidates[0]) === null || _a === void 0 ? void 0 : _a.groundingMetadata) === null || _b === void 0 ? void 0 : _b.retrievalQueries).toBeTruthy(`sys test failure on generateContent with RAG tool, empty groundingMetadata.retrievalQueries, for resp ${JSON.stringify(response)}`);
});
});

@@ -477,3 +510,3 @@ describe('generateContent', () => {

});
it('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
var _a;

@@ -492,3 +525,3 @@ const request = {

});
it('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
var _a;

@@ -507,5 +540,24 @@ const request = {

});
it('should return a FunctionCall when passed a FunctionDeclaration', async () => {
it('in preview should return grounding metadata when passed a VertexRagStore', async () => {
var _a, _b;
const request = {
contents: [
{
role: 'user',
parts: [
{
text: 'How much gain or loss did Google get in the Motorola Mobile deal in 2014?',
},
],
},
],
tools: TOOLS_WITH_RAG,
};
const resp = await generativeTextModelPreview.generateContent(request);
expect(resp.response.candidates[0]).toBeTruthy(`sys test failure on generateContent with RAG tool, for resp ${JSON.stringify(resp)}`);
expect((_b = (_a = resp.response.candidates[0]) === null || _a === void 0 ? void 0 : _a.groundingMetadata) === null || _b === void 0 ? void 0 : _b.retrievalQueries).toBeTruthy(`sys test failure on generateContent with RAG tool, empty groundingMetadata.retrievalQueries, for resp ${JSON.stringify(resp)}`);
});
xit('should return a FunctionCall when passed a FunctionDeclaration', async () => {
const request = {
contents: [
{ role: 'user', parts: [{ text: 'What is the weather in Boston?' }] },

@@ -523,3 +575,3 @@ ],

});
it('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
xit('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
const request = {

@@ -679,3 +731,3 @@ contents: [

});
it('should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
const chat = generativeTextModel.startChat({

@@ -700,3 +752,3 @@ tools: TOOLS_WITH_FUNCTION_DECLARATION,

});
it('in preview should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('in preview should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
const chat = generativeTextModelPreview.startChat({

@@ -703,0 +755,0 @@ tools: TOOLS_WITH_FUNCTION_DECLARATION,

38

build/test/vertex_ai_test.js

@@ -16,2 +16,40 @@ "use strict";

});
it('no location given, should instantiate VertexAI and VertexAIPreview', () => {
const vertexaiNoLocation = new vertex_ai_1.VertexAI({ project: PROJECT });
const generativeModel = vertexaiNoLocation.getGenerativeModel({
model: 'gemini-pro',
});
const generativeModelPreview = vertexaiNoLocation.preview.getGenerativeModel({
model: 'gemini-pro',
});
expect(vertexaiNoLocation).toBeInstanceOf(vertex_ai_1.VertexAI);
expect(generativeModel).toBeInstanceOf(models_1.GenerativeModel);
expect(generativeModelPreview).toBeInstanceOf(models_1.GenerativeModelPreview);
});
it('location in run time env GOOGLE_CLOUD_REGION, should instantiate VertexAI and VertexAIPreview', () => {
process.env['GOOGLE_CLOUD_REGION'] = 'us-central1';
const vertexaiNoLocation = new vertex_ai_1.VertexAI({ project: PROJECT });
const generativeModel = vertexaiNoLocation.getGenerativeModel({
model: 'gemini-pro',
});
const generativeModelPreview = vertexaiNoLocation.preview.getGenerativeModel({
model: 'gemini-pro',
});
expect(vertexaiNoLocation).toBeInstanceOf(vertex_ai_1.VertexAI);
expect(generativeModel).toBeInstanceOf(models_1.GenerativeModel);
expect(generativeModelPreview).toBeInstanceOf(models_1.GenerativeModelPreview);
});
it('location in run time env CLOUD_ML_REGION, should instantiate VertexAI and VertexAIPreview', () => {
process.env['CLOUD_ML_REGION'] = 'us-central1';
const vertexaiNoLocation = new vertex_ai_1.VertexAI({ project: PROJECT });
const generativeModel = vertexaiNoLocation.getGenerativeModel({
model: 'gemini-pro',
});
const generativeModelPreview = vertexaiNoLocation.preview.getGenerativeModel({
model: 'gemini-pro',
});
expect(vertexaiNoLocation).toBeInstanceOf(vertex_ai_1.VertexAI);
expect(generativeModel).toBeInstanceOf(models_1.GenerativeModel);
expect(generativeModelPreview).toBeInstanceOf(models_1.GenerativeModelPreview);
});
it('given undefined google auth options, should be instantiated', () => {

@@ -18,0 +56,0 @@ expect(vertexai).toBeInstanceOf(vertex_ai_1.VertexAI);

16

CHANGELOG.md
# Changelog
## [1.2.0](https://github.com/googleapis/nodejs-vertexai/compare/v1.1.0...v1.2.0) (2024-05-22)
### Features
* allow users to pass string as system instruction ([a824162](https://github.com/googleapis/nodejs-vertexai/commit/a824162a29b8c6ba3ccae46c492dd28e9c2baf9c))
* enable inference request to tuned model. ([de9c4c2](https://github.com/googleapis/nodejs-vertexai/commit/de9c4c2f8c63a298bd28ab69dae9b6a5d72c20d7))
* infer location if user doesn't specifies it. ([b8d4af1](https://github.com/googleapis/nodejs-vertexai/commit/b8d4af1bb990e95093f446c808194bfc4fe53287))
* support RAG in public preview ([5ade755](https://github.com/googleapis/nodejs-vertexai/commit/5ade7551fe0dbab54bc56f251c32bf3b7802c2c5))
* update grounding metadata ([d3c0a64](https://github.com/googleapis/nodejs-vertexai/commit/d3c0a647248be6be49b6b93a18aad79c10bae6c4))
### Bug Fixes
* log instead of throw appendHistory errors to avoid unhandled rejection ([2ec9e7d](https://github.com/googleapis/nodejs-vertexai/commit/2ec9e7d5519af438eb03b9f21f90b86f2575ac47))
## [1.1.0](https://github.com/googleapis/nodejs-vertexai/compare/v1.0.0...v1.1.0) (2024-04-13)

@@ -4,0 +20,0 @@

2

package.json
{
"name": "@google-cloud/vertexai",
"description": "Vertex Generative AI client for Node.js",
"version": "1.1.0",
"version": "1.2.0",
"license": "Apache-2.0",

@@ -6,0 +6,0 @@ "author": "Google LLC",

1

README.md

@@ -18,2 +18,3 @@ [![NPM Downloads](https://img.shields.io/npm/dm/%40google-cloud%2Fvertexai)](https://www.npmjs.com/package/@google-cloud/vertexai)

1. Make sure your node.js version is 18 or above.
1. [Select](https://console.cloud.google.com/project) or [create](https://cloud.google.com/resource-manager/docs/creating-managing-projects#creating_a_project) a Google Cloud project.

@@ -20,0 +21,0 @@ 1. [Enable billing for your project](https://cloud.google.com/billing/docs/how-to/modify-project).

9

src/functions/count_tokens.ts

@@ -24,2 +24,3 @@ /**

import {GoogleGenerativeAIError} from '../types/errors';
import * as constants from '../util/constants';
import {

@@ -38,4 +39,3 @@ throwErrorIfNotOK,

location: string,
project: string,
publisherModelEndpoint: string,
resourcePath: string,
token: Promise<string | null | undefined>,

@@ -48,5 +48,4 @@ request: CountTokensRequest,

region: location,
project: project,
resourcePath: publisherModelEndpoint,
resourceMethod: 'countTokens',
resourcePath: resourcePath,
resourceMethod: constants.COUNT_TOKENS_METHOD,
token: await token,

@@ -53,0 +52,0 @@ data: request,

24

src/functions/generate_content.ts

@@ -46,2 +46,4 @@ /**

validateGenerationConfig,
hasVertexRagStore,
getApiVersion,
} from './pre_fetch_processing';

@@ -51,4 +53,3 @@

location: string,
project: string,
publisherModelEndpoint: string,
resourcePath: string,
token: Promise<string | null | undefined>,

@@ -81,9 +82,9 @@ request: GenerateContentRequest | string,

region: location,
project: project,
resourcePath: publisherModelEndpoint,
resourcePath,
resourceMethod: constants.GENERATE_CONTENT_METHOD,
token: await token,
data: generateContentRequest,
apiEndpoint: apiEndpoint,
requestOptions: requestOptions,
apiEndpoint,
requestOptions,
apiVersion: getApiVersion(request),
}).catch(e => {

@@ -107,4 +108,3 @@ throw new GoogleGenerativeAIError('exception posting request to model', e);

location: string,
project: string,
publisherModelEndpoint: string,
resourcePath: string,
token: Promise<string | null | undefined>,

@@ -136,9 +136,9 @@ request: GenerateContentRequest | string,

region: location,
project: project,
resourcePath: publisherModelEndpoint,
resourcePath,
resourceMethod: constants.STREAMING_GENERATE_CONTENT_METHOD,
token: await token,
data: generateContentRequest,
apiEndpoint: apiEndpoint,
requestOptions: requestOptions,
apiEndpoint,
requestOptions,
apiVersion: getApiVersion(request),
}).catch(e => {

@@ -145,0 +145,0 @@ throw new GoogleGenerativeAIError('exception posting request', e);

12

src/functions/post_fetch_processing.ts

@@ -26,3 +26,2 @@ /**

GroundingMetadata,
Part,
StreamGenerateContentResult,

@@ -301,2 +300,3 @@ } from '../types/content';

groundingAttributions: [],
retrievalQueries: [],
};

@@ -319,2 +319,12 @@ const groundingMetadataAggregated: GroundingMetadata =

}
if (groundingMetadataChunk.retrievalQueries) {
groundingMetadataAggregated.retrievalQueries =
groundingMetadataAggregated.retrievalQueries!.concat(
groundingMetadataChunk.retrievalQueries
);
}
if (groundingMetadataChunk.searchEntryPoint) {
groundingMetadataAggregated.searchEntryPoint =
groundingMetadataChunk.searchEntryPoint;
}
return groundingMetadataAggregated;

@@ -321,0 +331,0 @@ }

4

src/functions/post_request.ts

@@ -41,3 +41,2 @@ /**

region,
project,
resourcePath,

@@ -52,3 +51,2 @@ resourceMethod,

region: string;
project: string;
resourcePath: string;

@@ -64,3 +62,3 @@ resourceMethod: string;

let vertexEndpoint = `https://${vertexBaseEndpoint}/${apiVersion}/projects/${project}/locations/${region}/${resourcePath}:${resourceMethod}`;
let vertexEndpoint = `https://${vertexBaseEndpoint}/${apiVersion}/${resourcePath}:${resourceMethod}`;

@@ -67,0 +65,0 @@ // Use server sent events for streamGenerateContent

37

src/functions/pre_fetch_processing.ts

@@ -21,4 +21,7 @@ /**

GenerationConfig,
RetrievalTool,
SafetySetting,
Tool,
} from '../types/content';
import {ClientError} from '../types/errors';
import * as constants from '../util/constants';

@@ -59,2 +62,8 @@

}
if (hasVertexAISearch(request) && hasVertexRagStore(request)) {
throw new ClientError(
'Found both vertexAiSearch and vertexRagStore field are set in tool. Either set vertexAiSearch or vertexRagStore.'
);
}
}

@@ -72,1 +81,29 @@

}
export function getApiVersion(
request: GenerateContentRequest
): 'v1' | 'v1beta1' {
return hasVertexRagStore(request) ? 'v1beta1' : 'v1';
}
export function hasVertexRagStore(request: GenerateContentRequest): boolean {
for (const tool of request?.tools ?? []) {
const retrieval = (tool as RetrievalTool).retrieval;
if (!retrieval) continue;
if (retrieval.vertexRagStore) {
return true;
}
}
return false;
}
export function hasVertexAISearch(request: GenerateContentRequest): boolean {
for (const tool of request?.tools ?? []) {
const retrieval = (tool as RetrievalTool).retrieval;
if (!retrieval) continue;
if (retrieval.vertexAiSearch) {
return true;
}
}
return false;
}

139

src/functions/test/functions_test.ts

@@ -40,5 +40,4 @@ /**

const TEST_PROJECT = 'test-project';
const TEST_LOCATION = 'test-location';
const TEST_PUBLISHER_MODEL_ENDPOINT = 'test-publisher-model-endpoint';
const TEST_RESOURCE_PATH = 'test-resource-path';
const TEST_TOKEN = 'testtoken';

@@ -52,2 +51,9 @@ const TEST_TOKEN_PROMISE = Promise.resolve(TEST_TOKEN);

const CONTENTS = [
{
role: 'user',
parts: [{text: 'What is the weater like in Boston?'}],
},
];
const TEST_USER_CHAT_MESSAGE_WITH_GCS_FILE = [

@@ -214,2 +220,8 @@ {

const TEST_TOOLS_WITH_RAG: Tool[] = [
{
retrieval: {vertexRagStore: {ragResources: [{ragCorpus: 'ragCorpus'}]}},
},
];
const fetchResponseObj = {

@@ -255,4 +267,3 @@ status: 200,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -275,4 +286,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -304,4 +314,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -331,4 +340,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -361,4 +369,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -385,4 +392,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -401,4 +407,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -421,4 +426,3 @@ TEST_CHAT_MESSAGE_TEXT,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -438,4 +442,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -460,4 +463,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -476,4 +478,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -497,4 +498,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -519,4 +519,3 @@ reqWithEmptyConfigs,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -539,4 +538,3 @@ reqWithEmptyConfigs,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -555,5 +553,3 @@ req,

const req: GenerateContentRequest = {
contents: [
{role: 'user', parts: [{text: 'What is the weater like in Boston?'}]},
],
contents: CONTENTS,
tools: TEST_TOOLS_WITH_FUNCTION_DECLARATION,

@@ -573,4 +569,3 @@ };

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -597,8 +592,3 @@ req,

const req: GenerateContentRequest = {
contents: [
{
role: 'user',
parts: [{text: 'What is the weater like in Boston?'}],
},
],
contents: CONTENTS,
tools: TEST_TOOLS_WITH_FUNCTION_DECLARATION,

@@ -618,4 +608,3 @@ };

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -635,8 +624,3 @@ req,

const req: GenerateContentRequest = {
contents: [
{
role: 'user',
parts: [{text: 'What is the weater like in Boston?'}],
},
],
contents: CONTENTS,
tools: TEST_TOOLS_WITH_FUNCTION_DECLARATION,

@@ -648,4 +632,3 @@ };

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -657,2 +640,35 @@ req,

});
it('should use v1 apiVersion', async () => {
const request: GenerateContentRequest = {
contents: CONTENTS,
};
fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
await generateContent(
TEST_LOCATION,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,
request,
TEST_API_ENDPOINT
);
const vertexEndpoint = fetchSpy.calls.allArgs()[0][0];
expect(vertexEndpoint).toContain('/v1/');
});
it('should use v1beta1 apiVersion when set RAG in tools', async () => {
const request: GenerateContentRequest = {
contents: CONTENTS,
tools: TEST_TOOLS_WITH_RAG,
};
fetchSpy.and.resolveTo(buildFetchResponse(TEST_MODEL_RESPONSE));
await generateContent(
TEST_LOCATION,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,
request,
TEST_API_ENDPOINT
);
const vertexEndpoint = fetchSpy.calls.allArgs()[0][0];
expect(vertexEndpoint).toContain('/v1beta1/');
});
});

@@ -688,4 +704,3 @@

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -714,4 +729,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -733,4 +747,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -755,4 +768,3 @@ TEST_API_ENDPOINT,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -777,4 +789,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -801,4 +812,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -838,4 +848,3 @@ req,

TEST_LOCATION,
TEST_PROJECT,
TEST_PUBLISHER_MODEL_ENDPOINT,
TEST_RESOURCE_PATH,
TEST_TOKEN_PROMISE,

@@ -842,0 +851,0 @@ req,

18

src/functions/test/post_fetch_processing_test.ts

@@ -43,4 +43,3 @@ /**

const LOCATION = 'location';
const PROJECT = 'project';
const PUBLISHER_MODEL_ENDPOINT = 'publisher_model_endpoint';
const RESOURCE_PATH = 'RESOURCE_PATH';
const TOKEN = Promise.resolve('token');

@@ -90,4 +89,3 @@ const GENERATE_CONTENT_REQUEST = 'generate_content_request';

LOCATION,
PROJECT,
PUBLISHER_MODEL_ENDPOINT,
RESOURCE_PATH,
TOKEN,

@@ -110,4 +108,3 @@ GENERATE_CONTENT_REQUEST

LOCATION,
PROJECT,
PUBLISHER_MODEL_ENDPOINT,
RESOURCE_PATH,
TOKEN,

@@ -127,4 +124,3 @@ GENERATE_CONTENT_REQUEST

LOCATION,
PROJECT,
PUBLISHER_MODEL_ENDPOINT,
RESOURCE_PATH,
TOKEN,

@@ -149,4 +145,3 @@ GENERATE_CONTENT_REQUEST

LOCATION,
PROJECT,
PUBLISHER_MODEL_ENDPOINT,
RESOURCE_PATH,
TOKEN,

@@ -172,4 +167,3 @@ GENERATE_CONTENT_REQUEST

LOCATION,
PROJECT,
PUBLISHER_MODEL_ENDPOINT,
RESOURCE_PATH,
TOKEN,

@@ -176,0 +170,0 @@ COUNT_TOKEN_REQUEST

9

src/functions/test/post_request_test.ts

@@ -6,3 +6,2 @@ import {GenerateContentRequest, RequestOptions} from '../../types';

const REGION = 'us-central1';
const PROJECT = 'project-id';
const RESOURCE_PATH = 'resource-path';

@@ -28,3 +27,2 @@ const RESOURCE_METHOD = 'resource-method';

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -49,3 +47,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -67,3 +64,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -88,3 +84,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -109,3 +104,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -127,3 +121,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -150,3 +143,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -179,3 +171,2 @@ resourceMethod: RESOURCE_METHOD,

region: REGION,
project: PROJECT,
resourcePath: RESOURCE_PATH,

@@ -182,0 +173,0 @@ resourceMethod: RESOURCE_METHOD,

18

src/functions/test/test_data.ts

@@ -326,2 +326,6 @@ /**

],
searchEntryPoint: {
renderedContent:
'rendered content for later chunk for first candidate',
},
},

@@ -348,2 +352,6 @@ },

],
searchEntryPoint: {
renderedContent:
'rendered content for later chunk for second candidate',
},
},

@@ -435,2 +443,7 @@ },

],
retrievalQueries: [],
searchEntryPoint: {
renderedContent:
'rendered content for later chunk for first candidate',
},
},

@@ -506,2 +519,7 @@ finishReason: 'STOP',

],
retrievalQueries: [],
searchEntryPoint: {
renderedContent:
'rendered content for later chunk for second candidate',
},
},

@@ -508,0 +526,0 @@ finishReason: 'STOP',

47

src/models/chat_session.ts

@@ -21,2 +21,3 @@ /**

import {formulateSystemInstructionIntoContent} from './util';
import {

@@ -38,7 +39,3 @@ generateContent,

} from '../types/content';
import {
ClientError,
GoogleAuthError,
GoogleGenerativeAIError,
} from '../types/errors';
import {ClientError, GoogleAuthError} from '../types/errors';
import {constants} from '../util';

@@ -59,3 +56,3 @@

private sendStreamPromise: Promise<void> = Promise.resolve();
private readonly publisherModelEndpoint: string;
private readonly resourcePath: string;
private readonly googleAuth: GoogleAuth;

@@ -84,3 +81,3 @@ protected readonly requestOptions?: RequestOptions;

this.googleAuth = request.googleAuth;
this.publisherModelEndpoint = request.publisherModelEndpoint;
this.resourcePath = request.resourcePath;
this.historyInternal = request.history ?? [];

@@ -93,5 +90,6 @@ this.generationConfig = request.generationConfig;

if (request.systemInstruction) {
request.systemInstruction.role = constants.SYSTEM_ROLE;
this.systemInstruction = formulateSystemInstructionIntoContent(
request.systemInstruction
);
}
this.systemInstruction = request.systemInstruction;
}

@@ -145,4 +143,3 @@

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -230,4 +227,3 @@ generateContentrequest,

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -248,3 +244,5 @@ generateContentrequest,

).catch(e => {
throw new GoogleGenerativeAIError('exception appending chat history', e);
// Errors from remote endpoint will be catchable by user from streamGenerateContentResultPromise
// Errors in appendHistory should not throw to cause user's programe exit with code 1
console.error(e);
});

@@ -268,3 +266,3 @@ return streamGenerateContentResultPromise;

private sendStreamPromise: Promise<void> = Promise.resolve();
private readonly publisherModelEndpoint: string;
private readonly resourcePath: string;
private readonly googleAuth: GoogleAuth;

@@ -293,3 +291,3 @@ protected readonly requestOptions?: RequestOptions;

this.googleAuth = request.googleAuth;
this.publisherModelEndpoint = request.publisherModelEndpoint;
this.resourcePath = request.resourcePath;
this.historyInternal = request.history ?? [];

@@ -302,5 +300,6 @@ this.generationConfig = request.generationConfig;

if (request.systemInstruction) {
request.systemInstruction.role = constants.SYSTEM_ROLE;
this.systemInstruction = formulateSystemInstructionIntoContent(
request.systemInstruction
);
}
this.systemInstruction = request.systemInstruction;
}

@@ -353,4 +352,3 @@

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -439,4 +437,3 @@ generateContentrequest,

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -456,3 +453,7 @@ generateContentrequest,

newContent
);
).catch(e => {
// Errors from remote endpoint will be catchable by user from streamGenerateContentResultPromise
// Errors in appendHistory should not throw to cause user's programe exit with code 1
console.error(e);
});
return streamGenerateContentResultPromise;

@@ -459,0 +460,0 @@ }

92

src/models/generative_models.ts

@@ -21,2 +21,3 @@ /**

import {formulateSystemInstructionIntoContent} from './util';
import {countTokens} from '../functions/count_tokens';

@@ -42,3 +43,3 @@ import {

} from '../types/content';
import {GoogleAuthError} from '../types/errors';
import {ClientError, GoogleAuthError} from '../types/errors';
import {constants} from '../util';

@@ -65,2 +66,3 @@

private readonly publisherModelEndpoint: string;
private readonly resourcePath: string;
private readonly apiEndpoint?: string;

@@ -83,10 +85,13 @@

if (getGenerativeModelParams.systemInstruction) {
getGenerativeModelParams.systemInstruction.role = constants.SYSTEM_ROLE;
this.systemInstruction = formulateSystemInstructionIntoContent(
getGenerativeModelParams.systemInstruction
);
}
this.systemInstruction = getGenerativeModelParams.systemInstruction;
if (this.model.startsWith('models/')) {
this.publisherModelEndpoint = `publishers/google/${this.model}`;
} else {
this.publisherModelEndpoint = `publishers/google/models/${this.model}`;
}
this.resourcePath = formulateResourcePathFromModel(
this.model,
this.project,
this.location
);
// publisherModelEndpoint is deprecated
this.publisherModelEndpoint = this.resourcePath;
}

@@ -135,4 +140,3 @@

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -183,4 +187,3 @@ formulatedRequest,

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -217,4 +220,3 @@ formulatedRequest,

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -257,3 +259,5 @@ request,

publisherModelEndpoint: this.publisherModelEndpoint,
resourcePath: this.resourcePath,
tools: this.tools,
systemInstruction: this.systemInstruction,
};

@@ -293,2 +297,3 @@

private readonly publisherModelEndpoint: string;
private readonly resourcePath: string;
private readonly apiEndpoint?: string;

@@ -311,10 +316,13 @@

if (getGenerativeModelParams.systemInstruction) {
getGenerativeModelParams.systemInstruction.role = constants.SYSTEM_ROLE;
this.systemInstruction = formulateSystemInstructionIntoContent(
getGenerativeModelParams.systemInstruction
);
}
this.systemInstruction = getGenerativeModelParams.systemInstruction;
if (this.model.startsWith('models/')) {
this.publisherModelEndpoint = `publishers/google/${this.model}`;
} else {
this.publisherModelEndpoint = `publishers/google/models/${this.model}`;
}
this.resourcePath = formulateResourcePathFromModel(
this.model,
this.project,
this.location
);
// publisherModelEndpoint is deprecated
this.publisherModelEndpoint = this.resourcePath;
}

@@ -362,4 +370,3 @@

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -410,4 +417,3 @@ formulatedRequest,

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -444,4 +450,3 @@ formulatedRequest,

this.location,
this.project,
this.publisherModelEndpoint,
this.resourcePath,
this.fetchToken(),

@@ -484,3 +489,5 @@ request,

publisherModelEndpoint: this.publisherModelEndpoint,
resourcePath: this.resourcePath,
tools: this.tools,
systemInstruction: this.systemInstruction,
};

@@ -502,2 +509,29 @@

function formulateResourcePathFromModel(
model: string,
project: string,
location: string
): string {
let resourcePath: string;
if (!model) {
throw new ClientError('model parameter must not be empty.');
}
if (!model.includes('/')) {
// example 'gemini-1.0-pro'
resourcePath = `projects/${project}/locations/${location}/publishers/google/models/${model}`;
} else if (model.startsWith('models/')) {
// example 'models/gemini-1.0-pro'
resourcePath = `projects/${project}/locations/${location}/publishers/google/${model}`;
} else if (model.startsWith('projects/')) {
// example 'projects/my-project/locations/my-location/models/my-tuned-model'
resourcePath = model;
} else {
throw new ClientError(
'model parameter must be either a Model Garden model ID or a full resource name.'
);
}
return resourcePath;
}
function formulateRequestToGenerateContentRequest(

@@ -519,3 +553,5 @@ request: GenerateContentRequest | string

if (methodRequest.systemInstruction) {
methodRequest.systemInstruction.role = constants.SYSTEM_ROLE;
methodRequest.systemInstruction = formulateSystemInstructionIntoContent(
methodRequest.systemInstruction
);
return methodRequest;

@@ -522,0 +558,0 @@ }

101

src/types/content.ts

@@ -27,5 +27,9 @@ /**

project: string;
/** The Google Cloud project location. */
location: string;
/**
* Optional. The Google Cloud project location. If not provided, SDK will
* firtly try to resolve it from run time environment. If no location resolved
* from run time environment, SDK will use default value `us-central1`.
*/
location?: string;
/**
* Optional. The base Vertex AI endpoint to use for the request. If not

@@ -51,6 +55,7 @@ * provided, the default regionalized endpoint (i.e.

contents: Content[];
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -111,6 +116,7 @@

requestOptions?: RequestOptions;
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -140,6 +146,7 @@

tools?: Tool[];
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -566,8 +573,16 @@

export declare interface GoogleDate {
/** Year of the date. Must be from 1 to 9999, or 0 to specify a date without a year. */
/**
* Year of the date. Must be from 1 to 9999, or 0 to specify a date without a
* year.
*/
year?: number;
/** Month of the date. Must be from 1 to 12, or 0 to specify a year without a monthi and day. */
/**
* Month of the date. Must be from 1 to 12, or 0 to specify a year without a
* monthi and day.
*/
month?: number;
/** Day of the date. Must be from 1 to 31 and valid for the year and month.
* or 0 to specify a year by itself or a year and month where the day isn't significant
/**
* Day of the date. Must be from 1 to 31 and valid for the year and month.
* or 0 to specify a year by itself or a year and month where the day isn't
* significant
*/

@@ -596,2 +611,12 @@ day?: number;

/**
* Google search entry point.
*/
export declare interface SearchEntryPoint {
/** Optional. Web content snippet that can be embedded in a web page or an app webview. */
renderedContent?: string;
/** Optional. Base64 encoded JSON representing array of tuple. */
sdkBlob?: string;
}
/**
* A collection of grounding attributions for a piece of content.

@@ -606,2 +631,4 @@ */

groundingAttributions?: GroundingAttribution[];
/** Optional. Google search entry for the following-up web searches. {@link SearchEntryPoint} */
searchEntryPoint?: SearchEntryPoint;
}

@@ -757,3 +784,37 @@

export declare interface VertexRagStore {
/**
* Optional. List of corpora for retrieval. Currently only support one corpus
* or multiple files from one corpus. In the future we may open up multiple
* corpora support.
*/
ragResources?: RagResource[];
/** Optional. Number of top k results to return from the selected corpora. */
similarityTopK?: number;
/** Optional. If set this field, results with vector distance smaller than this threshold will be returned. */
vectorDistanceThreshold?: number;
}
/**
* Config of Vertex RagStore grounding checking.
*/
export declare interface RagResource {
/**
* Optional. Vertex RAG Store corpus resource name.
*
* @example
* `projects/{project}/locations/{location}/ragCorpora/{rag_corpus}`
*/
ragCorpus?: string;
/**
* Optional. Set this field to select the files under the ragCorpora for
* retrieval.
*/
ragFileIds?: string[];
}
/**
* Defines a retrieval tool that model can call to access external knowledge.

@@ -781,2 +842,6 @@ */

vertexAiSearch?: VertexAISearch;
/** Optional. Set to use data source powered by Vertex RAG store. */
vertexRagStore?: VertexRagStore;
/**

@@ -892,6 +957,7 @@ * Optional. Disable using the result from this tool in detecting grounding

apiEndpoint?: string;
/** Optional. The user provided system instructions for the model.
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -911,6 +977,9 @@

publisherModelEndpoint: string;
/** Optional. The user provided system instructions for the model.
/** The resource path to use for the request. */
resourcePath: string;
/**
* Optional. The user provided system instructions for the model.
* Note: only text should be used in parts of {@link Content}
*/
systemInstruction?: Content;
systemInstruction?: string | Content;
}

@@ -917,0 +986,0 @@

3

src/util/constants.ts

@@ -19,2 +19,3 @@ /**

export const STREAMING_GENERATE_CONTENT_METHOD = 'streamGenerateContent';
export const COUNT_TOKENS_METHOD = 'countTokens';
export const USER_ROLE = 'user';

@@ -24,3 +25,3 @@ export const MODEL_ROLE = 'model';

const USER_AGENT_PRODUCT = 'model-builder';
const CLIENT_LIBRARY_VERSION = '1.1.0'; // x-release-please-version
const CLIENT_LIBRARY_VERSION = '1.2.0'; // x-release-please-version
const CLIENT_LIBRARY_LANGUAGE = `grpc-node/${CLIENT_LIBRARY_VERSION}`;

@@ -27,0 +28,0 @@ export const USER_AGENT = `${USER_AGENT_PRODUCT}/${CLIENT_LIBRARY_VERSION} ${CLIENT_LIBRARY_LANGUAGE}`;

14

src/vertex_ai.ts

@@ -55,3 +55,3 @@ /**

this.project = init.project;
this.location = init.location;
this.location = resolveLocation(init.location);
this.googleAuth = new GoogleAuth(opts);

@@ -226,1 +226,13 @@ this.apiEndpoint = init.apiEndpoint;

}
function resolveLocation(locationFromInput?: string): string {
if (locationFromInput) {
return locationFromInput;
}
const inferredLocation =
process.env['GOOGLE_CLOUD_REGION'] || process.env['CLOUD_ML_REGION'];
if (inferredLocation) {
return inferredLocation;
}
return 'us-central1';
}

96

system_test/end_to_end_sample_test.ts

@@ -30,3 +30,3 @@ /**

const PROJECT = process.env.GCLOUD_PROJECT;
const PROJECT = process.env['GCLOUD_PROJECT'];
const LOCATION = 'us-central1';

@@ -95,2 +95,17 @@ const TEXT_REQUEST = {

const TOOLS_WITH_RAG = [
{
retrieval: {
vertexRagStore: {
ragResources: [
{
ragCorpus:
'projects/ucaip-sample-tests/locations/us-central1/ragCorpora/6917529027641081856',
},
],
},
},
},
];
const WEATHER_FORECAST = 'super nice';

@@ -403,3 +418,3 @@ const FUNCTION_RESPONSE_PART = [

it('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
const request = {

@@ -432,3 +447,3 @@ contents: [

});
it('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
const request = {

@@ -461,3 +476,3 @@ contents: [

});
it('should return a FunctionCall when passed a FunctionDeclaration', async () => {
xit('should return a FunctionCall when passed a FunctionDeclaration', async () => {
const request = {

@@ -488,3 +503,3 @@ contents: [

});
it('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
xit('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
const request = {

@@ -603,2 +618,33 @@ contents: [

});
it('in preview should return grounding metadata when passed a VertexRagStore', async () => {
const request = {
contents: [
{
role: 'user',
parts: [
{
text: 'How much gain or loss did Google get in the Motorola Mobile deal in 2014?',
},
],
},
],
tools: TOOLS_WITH_RAG,
};
const result =
await generativeTextModelPreview.generateContentStream(request);
const response = await result.response;
expect(response.candidates![0]).toBeTruthy(
`sys test failure on generateContent with RAG tool, for resp ${JSON.stringify(
response
)}`
);
expect(
response.candidates![0]?.groundingMetadata?.retrievalQueries
).toBeTruthy(
`sys test failure on generateContent with RAG tool, empty groundingMetadata.retrievalQueries, for resp ${JSON.stringify(
response
)}`
);
});
});

@@ -719,3 +765,3 @@

});
it('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
const request = {

@@ -745,3 +791,3 @@ contents: [

});
it('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('in preview should return a text when passed a FunctionDeclaration or FunctionResponse', async () => {
const request = {

@@ -770,5 +816,33 @@ contents: [

});
it('should return a FunctionCall when passed a FunctionDeclaration', async () => {
it('in preview should return grounding metadata when passed a VertexRagStore', async () => {
const request = {
contents: [
{
role: 'user',
parts: [
{
text: 'How much gain or loss did Google get in the Motorola Mobile deal in 2014?',
},
],
},
],
tools: TOOLS_WITH_RAG,
};
const resp = await generativeTextModelPreview.generateContent(request);
expect(resp.response.candidates![0]).toBeTruthy(
`sys test failure on generateContent with RAG tool, for resp ${JSON.stringify(
resp
)}`
);
expect(
resp.response.candidates![0]?.groundingMetadata?.retrievalQueries
).toBeTruthy(
`sys test failure on generateContent with RAG tool, empty groundingMetadata.retrievalQueries, for resp ${JSON.stringify(
resp
)}`
);
});
xit('should return a FunctionCall when passed a FunctionDeclaration', async () => {
const request = {
contents: [
{role: 'user', parts: [{text: 'What is the weather in Boston?'}]},

@@ -795,3 +869,3 @@ ],

});
it('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
xit('in preview should return a FunctionCall when passed a FunctionDeclaration', async () => {
const request = {

@@ -1023,3 +1097,3 @@ contents: [

it('should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
const chat = generativeTextModel.startChat({

@@ -1072,3 +1146,3 @@ tools: TOOLS_WITH_FUNCTION_DECLARATION,

});
it('in preview should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
xit('in preview should return a FunctionCall or text when passed a FunctionDeclaration or FunctionResponse', async () => {
const chat = generativeTextModelPreview.startChat({

@@ -1075,0 +1149,0 @@ tools: TOOLS_WITH_FUNCTION_DECLARATION,

44

test/vertex_ai_test.ts

@@ -17,2 +17,46 @@ import {VertexAI} from '../src/vertex_ai';

it('no location given, should instantiate VertexAI and VertexAIPreview', () => {
const vertexaiNoLocation = new VertexAI({project: PROJECT});
const generativeModel = vertexaiNoLocation.getGenerativeModel({
model: 'gemini-pro',
});
const generativeModelPreview =
vertexaiNoLocation.preview.getGenerativeModel({
model: 'gemini-pro',
});
expect(vertexaiNoLocation).toBeInstanceOf(VertexAI);
expect(generativeModel).toBeInstanceOf(GenerativeModel);
expect(generativeModelPreview).toBeInstanceOf(GenerativeModelPreview);
});
it('location in run time env GOOGLE_CLOUD_REGION, should instantiate VertexAI and VertexAIPreview', () => {
process.env['GOOGLE_CLOUD_REGION'] = 'us-central1';
const vertexaiNoLocation = new VertexAI({project: PROJECT});
const generativeModel = vertexaiNoLocation.getGenerativeModel({
model: 'gemini-pro',
});
const generativeModelPreview =
vertexaiNoLocation.preview.getGenerativeModel({
model: 'gemini-pro',
});
expect(vertexaiNoLocation).toBeInstanceOf(VertexAI);
expect(generativeModel).toBeInstanceOf(GenerativeModel);
expect(generativeModelPreview).toBeInstanceOf(GenerativeModelPreview);
});
it('location in run time env CLOUD_ML_REGION, should instantiate VertexAI and VertexAIPreview', () => {
process.env['CLOUD_ML_REGION'] = 'us-central1';
const vertexaiNoLocation = new VertexAI({project: PROJECT});
const generativeModel = vertexaiNoLocation.getGenerativeModel({
model: 'gemini-pro',
});
const generativeModelPreview =
vertexaiNoLocation.preview.getGenerativeModel({
model: 'gemini-pro',
});
expect(vertexaiNoLocation).toBeInstanceOf(VertexAI);
expect(generativeModel).toBeInstanceOf(GenerativeModel);
expect(generativeModelPreview).toBeInstanceOf(GenerativeModelPreview);
});
it('given undefined google auth options, should be instantiated', () => {

@@ -19,0 +63,0 @@ expect(vertexai).toBeInstanceOf(VertexAI);

.kokoro/presubmit/node18/continuous-system-test.cfg

Sorry, the diff of this file is not supported yet

.kokoro/presubmit/node18/docs.sh

Sorry, the diff of this file is not supported yet

.kokoro/presubmit/node18/system-test.cfg

Sorry, the diff of this file is not supported yet

.kokoro/release/docs-devsite.sh

Sorry, the diff of this file is not supported yet

.kokoro/system-test.sh

Sorry, the diff of this file is not supported yet

build/src/functions/count_tokens.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/generate_content.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/post_fetch_processing.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/post_request.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/pre_fetch_processing.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/test/functions_test.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/test/post_fetch_processing_test.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/test/post_request_test.js.map

Sorry, the diff of this file is not supported yet

build/src/functions/test/test_data.js.map

Sorry, the diff of this file is not supported yet

build/src/models/chat_session.js.map

Sorry, the diff of this file is not supported yet

build/src/models/generative_models.js.map

Sorry, the diff of this file is not supported yet

build/src/models/test/models_test.js

Sorry, the diff of this file is too big to display

build/src/models/test/models_test.js.map

Sorry, the diff of this file is not supported yet

build/src/types/content.js.map

Sorry, the diff of this file is not supported yet

build/src/util/constants.js.map

Sorry, the diff of this file is not supported yet

build/src/vertex_ai.js.map

Sorry, the diff of this file is not supported yet

build/system_test/end_to_end_sample_test.js.map

Sorry, the diff of this file is not supported yet

build/test/vertex_ai_test.js.map

Sorry, the diff of this file is not supported yet

src/models/test/models_test.ts

Sorry, the diff of this file is too big to display

SocketSocket SOC 2 Logo

Product

  • Package Alerts
  • Integrations
  • Docs
  • Pricing
  • FAQ
  • Roadmap
  • Changelog

About

Packages

npm

Go

Maven

PyPI

Rubygems

Stay in touch

Get open source security insights delivered straight into your inbox.

  • Terms
  • Privacy
  • Security

Made with ⚡️ by Socket Inc