modelfusion 0.132.0 → 0.134.0

package/model-provider/llamacpp/LlamaCppCompletionModel.js

@@ -1,322 +0,0 @@
-import { z } from "zod";
-import { callWithRetryAndThrottle } from "../../core/api/callWithRetryAndThrottle.js";
-import { createJsonResponseHandler, postJsonToApi, } from "../../core/api/postToApi.js";
-import { zodSchema } from "../../core/schema/ZodSchema.js";
-import { parseJSON } from "../../core/schema/parseJSON.js";
-import { validateTypes } from "../../core/schema/validateTypes.js";
-import { AbstractModel } from "../../model-function/AbstractModel.js";
-import { ObjectFromTextStreamingModel } from "../../model-function/generate-object/ObjectFromTextStreamingModel.js";
-import { PromptTemplateTextStreamingModel } from "../../model-function/generate-text/PromptTemplateTextStreamingModel.js";
-import { textGenerationModelProperties, } from "../../model-function/generate-text/TextGenerationModel.js";
-import { AsyncQueue } from "../../util/AsyncQueue.js";
-import { parseEventSourceStream } from "../../util/streaming/parseEventSourceStream.js";
-import { LlamaCppApiConfiguration } from "./LlamaCppApiConfiguration.js";
-import { failedLlamaCppCallResponseHandler } from "./LlamaCppError.js";
-import { Text } from "./LlamaCppPrompt.js";
-import { LlamaCppTokenizer } from "./LlamaCppTokenizer.js";
-import { convertJsonSchemaToGBNF } from "./convertJsonSchemaToGBNF.js";
-export class LlamaCppCompletionModel extends AbstractModel {
-    constructor(settings = {}) {
-        super({ settings });
-        Object.defineProperty(this, "provider", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: "llamacpp"
-        });
-        Object.defineProperty(this, "tokenizer", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        this.tokenizer = new LlamaCppTokenizer(this.settings.api);
-    }
-    get modelName() {
-        return null;
-    }
-    get contextWindowSize() {
-        return this.settings.contextWindowSize;
-    }
-    async callAPI(prompt, callOptions, options) {
-        const api = this.settings.api ?? new LlamaCppApiConfiguration();
-        const responseFormat = options.responseFormat;
-        const abortSignal = callOptions.run?.abortSignal;
-        return callWithRetryAndThrottle({
-            retry: api.retry,
-            throttle: api.throttle,
-            call: async () => postJsonToApi({
-                url: api.assembleUrl(`/completion`),
-                headers: api.headers({
-                    functionType: callOptions.functionType,
-                    functionId: callOptions.functionId,
-                    run: callOptions.run,
-                    callId: callOptions.callId,
-                }),
-                body: {
-                    stream: responseFormat.stream,
-                    prompt: prompt.text,
-                    image_data: prompt.images != null
-                        ? Object.entries(prompt.images).map(([id, data]) => ({
-                            id: +id,
-                            data,
-                        }))
-                        : undefined,
-                    temperature: this.settings.temperature,
-                    top_k: this.settings.topK,
-                    top_p: this.settings.topP,
-                    min_p: this.settings.minP,
-                    n_predict: this.settings.maxGenerationTokens,
-                    n_keep: this.settings.nKeep,
-                    stop: this.settings.stopSequences,
-                    tfs_z: this.settings.tfsZ,
-                    typical_p: this.settings.typicalP,
-                    repeat_penalty: this.settings.repeatPenalty,
-                    repeat_last_n: this.settings.repeatLastN,
-                    penalize_nl: this.settings.penalizeNl,
-                    presence_penalty: this.settings.presencePenalty,
-                    frequency_penalty: this.settings.frequencyPenalty,
-                    penalty_prompt: this.settings.penaltyPrompt,
-                    mirostat: this.settings.mirostat,
-                    mirostat_tau: this.settings.mirostatTau,
-                    mirostat_eta: this.settings.mirostatEta,
-                    grammar: this.settings.grammar,
-                    seed: this.settings.seed,
-                    ignore_eos: this.settings.ignoreEos,
-                    logit_bias: this.settings.logitBias,
-                    n_probs: this.settings.nProbs,
-                    cache_prompt: this.settings.cachePrompt,
-                    slot_id: this.settings.slotId,
-                },
-                failedResponseHandler: failedLlamaCppCallResponseHandler,
-                successfulResponseHandler: responseFormat.handler,
-                abortSignal,
-            }),
-        });
-    }
-    get settingsForEvent() {
-        const eventSettingProperties = [
-            ...textGenerationModelProperties,
-            "contextWindowSize",
-            "temperature",
-            "topK",
-            "topP",
-            "minP",
-            "nKeep",
-            "tfsZ",
-            "typicalP",
-            "repeatPenalty",
-            "repeatLastN",
-            "penalizeNl",
-            "presencePenalty",
-            "frequencyPenalty",
-            "penaltyPrompt",
-            "mirostat",
-            "mirostatTau",
-            "mirostatEta",
-            "grammar",
-            "seed",
-            "ignoreEos",
-            "logitBias",
-            "nProbs",
-            "cachePrompt",
-            "slotId",
-        ];
-        return Object.fromEntries(Object.entries(this.settings).filter(([key]) => eventSettingProperties.includes(key)));
-    }
-    async countPromptTokens(prompt) {
-        const tokens = await this.tokenizer.tokenize(prompt.text);
-        return tokens.length;
-    }
-    async doGenerateTexts(prompt, options) {
-        return this.processTextGenerationResponse(await this.callAPI(prompt, options, {
-            responseFormat: LlamaCppCompletionResponseFormat.json,
-        }));
-    }
-    restoreGeneratedTexts(rawResponse) {
-        return this.processTextGenerationResponse(validateTypes({
-            value: rawResponse,
-            schema: zodSchema(llamaCppTextGenerationResponseSchema),
-        }));
-    }
-    processTextGenerationResponse(rawResponse) {
-        return {
-            rawResponse,
-            textGenerationResults: [
-                {
-                    text: rawResponse.content,
-                    finishReason: rawResponse.stopped_eos || rawResponse.stopped_word
-                        ? "stop"
-                        : rawResponse.stopped_limit
-                            ? "length"
-                            : "unknown",
-                },
-            ],
-            usage: {
-                promptTokens: rawResponse.tokens_evaluated,
-                completionTokens: rawResponse.tokens_predicted,
-                totalTokens: rawResponse.tokens_evaluated + rawResponse.tokens_predicted,
-            },
-        };
-    }
-    doStreamText(prompt, options) {
-        return this.callAPI(prompt, options, {
-            responseFormat: LlamaCppCompletionResponseFormat.deltaIterable,
-        });
-    }
-    extractTextDelta(delta) {
-        return delta.content;
-    }
-    asObjectGenerationModel(promptTemplate) {
-        return "adaptModel" in promptTemplate
-            ? new ObjectFromTextStreamingModel({
-                model: promptTemplate.adaptModel(this),
-                template: promptTemplate,
-            })
-            : new ObjectFromTextStreamingModel({
-                model: this,
-                template: promptTemplate,
-            });
-    }
-    withJsonOutput(schema) {
-        // don't override the grammar if it's already set (to allow user to override)
-        if (this.settings.grammar != null) {
-            return this;
-        }
-        const grammar = convertJsonSchemaToGBNF(schema.getJsonSchema());
-        return this.withSettings({
-            grammar: grammar,
-        });
-    }
-    get promptTemplateProvider() {
-        return this.settings.promptTemplate ?? Text;
-    }
-    withTextPrompt() {
-        return this.withPromptTemplate(this.promptTemplateProvider.text());
-    }
-    withInstructionPrompt() {
-        return this.withPromptTemplate(this.promptTemplateProvider.instruction());
-    }
-    withChatPrompt() {
-        return this.withPromptTemplate(this.promptTemplateProvider.chat());
-    }
-    /**
-     * Maps the prompt for the full Llama.cpp prompt template (incl. image support).
-     */
-    withPromptTemplate(promptTemplate) {
-        return new PromptTemplateTextStreamingModel({
-            model: this.withSettings({
-                stopSequences: [
-                    ...(this.settings.stopSequences ?? []),
-                    ...promptTemplate.stopSequences,
-                ],
-            }),
-            promptTemplate,
-        });
-    }
-    withSettings(additionalSettings) {
-        return new LlamaCppCompletionModel(Object.assign({}, this.settings, additionalSettings));
-    }
-}
-const llamaCppTextGenerationResponseSchema = z.object({
-    content: z.string(),
-    stop: z.literal(true),
-    generation_settings: z.object({
-        frequency_penalty: z.number(),
-        ignore_eos: z.boolean(),
-        logit_bias: z.array(z.number()),
-        mirostat: z.number(),
-        mirostat_eta: z.number(),
-        mirostat_tau: z.number(),
-        model: z.string(),
-        n_ctx: z.number(),
-        n_keep: z.number(),
-        n_predict: z.number(),
-        n_probs: z.number(),
-        penalize_nl: z.boolean(),
-        presence_penalty: z.number(),
-        repeat_last_n: z.number(),
-        repeat_penalty: z.number(),
-        seed: z.number(),
-        stop: z.array(z.string()),
-        stream: z.boolean(),
-        temperature: z.number().optional(), // optional for backwards compatibility
-        tfs_z: z.number(),
-        top_k: z.number(),
-        top_p: z.number(),
-        typical_p: z.number(),
-    }),
-    model: z.string(),
-    prompt: z.string(),
-    stopped_eos: z.boolean(),
-    stopped_limit: z.boolean(),
-    stopped_word: z.boolean(),
-    stopping_word: z.string(),
-    timings: z.object({
-        predicted_ms: z.number(),
-        predicted_n: z.number(),
-        predicted_per_second: z.number().nullable(),
-        predicted_per_token_ms: z.number().nullable(),
-        prompt_ms: z.number().nullable().optional(),
-        prompt_n: z.number(),
-        prompt_per_second: z.number().nullable(),
-        prompt_per_token_ms: z.number().nullable(),
-    }),
-    tokens_cached: z.number(),
-    tokens_evaluated: z.number(),
-    tokens_predicted: z.number(),
-    truncated: z.boolean(),
-});
-const llamaCppTextStreamChunkSchema = z.discriminatedUnion("stop", [
-    z.object({
-        content: z.string(),
-        stop: z.literal(false),
-    }),
-    llamaCppTextGenerationResponseSchema,
-]);
-async function createLlamaCppFullDeltaIterableQueue(stream) {
-    const queue = new AsyncQueue();
-    // process the stream asynchonously (no 'await' on purpose):
-    parseEventSourceStream({ stream })
-        .then(async (events) => {
-        try {
-            for await (const event of events) {
-                const data = event.data;
-                const eventData = parseJSON({
-                    text: data,
-                    schema: zodSchema(llamaCppTextStreamChunkSchema),
-                });
-                queue.push({ type: "delta", deltaValue: eventData });
-                if (eventData.stop) {
-                    queue.close();
-                }
-            }
-        }
-        catch (error) {
-            queue.push({ type: "error", error });
-            queue.close();
-        }
-    })
-        .catch((error) => {
-        queue.push({ type: "error", error });
-        queue.close();
-    });
-    return queue;
-}
-export const LlamaCppCompletionResponseFormat = {
-    /**
-     * Returns the response as a JSON object.
-     */
-    json: {
-        stream: false,
-        handler: createJsonResponseHandler(zodSchema(llamaCppTextGenerationResponseSchema)),
-    },
-    /**
-     * Returns an async iterable over the full deltas (all choices, including full current state at time of event)
-     * of the response stream.
-     */
-    deltaIterable: {
-        stream: true,
-        handler: async ({ response }) => createLlamaCppFullDeltaIterableQueue(response.body),
-    },
-};

package/model-provider/llamacpp/LlamaCppCompletionModel.test.cjs

@@ -1,40 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-const streamText_js_1 = require("../../model-function/generate-text/streamText.cjs");
-const StreamingTestServer_js_1 = require("../../test/StreamingTestServer.cjs");
-const arrayFromAsync_js_1 = require("../../test/arrayFromAsync.cjs");
-const LlamaCppCompletionModel_js_1 = require("./LlamaCppCompletionModel.cjs");
-describe("streamText", () => {
-    const server = new StreamingTestServer_js_1.StreamingTestServer("http://127.0.0.1:8080/completion");
-    server.setupTestEnvironment();
-    it("should return a text stream", async () => {
-        server.responseChunks = [
-            `data: {"content":"Hello","multimodal":false,"slot_id":0,"stop":false}\n\n`,
-            `data: {"content":", ","multimodal":false,"slot_id":0,"stop":false}\n\n`,
-            `data: {"content":"world!","multimodal":false,"slot_id":0,"stop":false}\n\n`,
-            `data: {"content":"","generation_settings":{"frequency_penalty":0.0,"grammar":"",` +
-                `"ignore_eos":false,"logit_bias":[],"min_p":0.05000000074505806,"mirostat":0,` +
-                `"mirostat_eta":0.10000000149011612,"mirostat_tau":5.0,"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
-                `"n_ctx":4096,"n_keep":0,"n_predict":-1,"n_probs":0,"penalize_nl":true,"penalty_prompt_tokens":[],` +
-                `"presence_penalty":0.0,"repeat_last_n":64,"repeat_penalty":1.100000023841858,"seed":4294967295,` +
-                `"stop":[],"stream":true,"temperature":0.800000011920929,"tfs_z":1.0,"top_k":40,"top_p":0.949999988079071,` +
-                `"typical_p":1.0,"use_penalty_prompt_tokens":false},"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
-                `"prompt":"hello","slot_id":0,"stop":true,"stopped_eos":true,"stopped_limit":false,` +
-                `"stopped_word":false,"stopping_word":"","timings":{"predicted_ms":1054.704,"predicted_n":69,` +
-                `"predicted_per_second":65.421198743913,"predicted_per_token_ms":15.285565217391303,` +
-                `"prompt_ms":244.228,"prompt_n":5,"prompt_per_second":20.472673075978186,` +
-                `"prompt_per_token_ms":48.845600000000005},"tokens_cached":74,"tokens_evaluated":5,` +
-                `"tokens_predicted":69,"truncated":false}\n\n`,
-        ];
-        const stream = await (0, streamText_js_1.streamText)({
-            model: new LlamaCppCompletionModel_js_1.LlamaCppCompletionModel().withTextPrompt(),
-            prompt: "hello",
-        });
-        // note: space moved to last chunk bc of trimming
-        expect(await (0, arrayFromAsync_js_1.arrayFromAsync)(stream)).toStrictEqual([
-            "Hello",
-            ",",
-            " world!",
-        ]);
-    });
-});

package/model-provider/llamacpp/LlamaCppCompletionModel.test.d.ts

@@ -1 +0,0 @@
-export {};

package/model-provider/llamacpp/LlamaCppCompletionModel.test.js

@@ -1,38 +0,0 @@
-import { streamText } from "../../model-function/generate-text/streamText.js";
-import { StreamingTestServer } from "../../test/StreamingTestServer.js";
-import { arrayFromAsync } from "../../test/arrayFromAsync.js";
-import { LlamaCppCompletionModel } from "./LlamaCppCompletionModel.js";
-describe("streamText", () => {
-    const server = new StreamingTestServer("http://127.0.0.1:8080/completion");
-    server.setupTestEnvironment();
-    it("should return a text stream", async () => {
-        server.responseChunks = [
-            `data: {"content":"Hello","multimodal":false,"slot_id":0,"stop":false}\n\n`,
-            `data: {"content":", ","multimodal":false,"slot_id":0,"stop":false}\n\n`,
-            `data: {"content":"world!","multimodal":false,"slot_id":0,"stop":false}\n\n`,
-            `data: {"content":"","generation_settings":{"frequency_penalty":0.0,"grammar":"",` +
-                `"ignore_eos":false,"logit_bias":[],"min_p":0.05000000074505806,"mirostat":0,` +
-                `"mirostat_eta":0.10000000149011612,"mirostat_tau":5.0,"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
-                `"n_ctx":4096,"n_keep":0,"n_predict":-1,"n_probs":0,"penalize_nl":true,"penalty_prompt_tokens":[],` +
-                `"presence_penalty":0.0,"repeat_last_n":64,"repeat_penalty":1.100000023841858,"seed":4294967295,` +
-                `"stop":[],"stream":true,"temperature":0.800000011920929,"tfs_z":1.0,"top_k":40,"top_p":0.949999988079071,` +
-                `"typical_p":1.0,"use_penalty_prompt_tokens":false},"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
-                `"prompt":"hello","slot_id":0,"stop":true,"stopped_eos":true,"stopped_limit":false,` +
-                `"stopped_word":false,"stopping_word":"","timings":{"predicted_ms":1054.704,"predicted_n":69,` +
-                `"predicted_per_second":65.421198743913,"predicted_per_token_ms":15.285565217391303,` +
-                `"prompt_ms":244.228,"prompt_n":5,"prompt_per_second":20.472673075978186,` +
-                `"prompt_per_token_ms":48.845600000000005},"tokens_cached":74,"tokens_evaluated":5,` +
-                `"tokens_predicted":69,"truncated":false}\n\n`,
-        ];
-        const stream = await streamText({
-            model: new LlamaCppCompletionModel().withTextPrompt(),
-            prompt: "hello",
-        });
-        // note: space moved to last chunk bc of trimming
-        expect(await arrayFromAsync(stream)).toStrictEqual([
-            "Hello",
-            ",",
-            " world!",
-        ]);
-    });
-});

package/model-provider/llamacpp/LlamaCppError.cjs

@@ -1,13 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.failedLlamaCppCallResponseHandler = void 0;
-const zod_1 = require("zod");
-const postToApi_js_1 = require("../../core/api/postToApi.cjs");
-const ZodSchema_js_1 = require("../../core/schema/ZodSchema.cjs");
-const llamaCppErrorDataSchema = zod_1.z.object({
-    error: zod_1.z.string(),
-});
-exports.failedLlamaCppCallResponseHandler = (0, postToApi_js_1.createJsonErrorResponseHandler)({
-    errorSchema: (0, ZodSchema_js_1.zodSchema)(llamaCppErrorDataSchema),
-    errorToMessage: (error) => error.error,
-});

package/model-provider/llamacpp/LlamaCppError.d.ts

@@ -1,11 +0,0 @@
-import { z } from "zod";
-declare const llamaCppErrorDataSchema: z.ZodObject<{
-    error: z.ZodString;
-}, "strip", z.ZodTypeAny, {
-    error: string;
-}, {
-    error: string;
-}>;
-export type LlamaCppErrorData = z.infer<typeof llamaCppErrorDataSchema>;
-export declare const failedLlamaCppCallResponseHandler: import("../../core/api/postToApi.js").ResponseHandler<import("../../index.js").ApiCallError>;
-export {};

package/model-provider/llamacpp/LlamaCppError.js

@@ -1,10 +0,0 @@
-import { z } from "zod";
-import { createJsonErrorResponseHandler } from "../../core/api/postToApi.js";
-import { zodSchema } from "../../core/schema/ZodSchema.js";
-const llamaCppErrorDataSchema = z.object({
-    error: z.string(),
-});
-export const failedLlamaCppCallResponseHandler = createJsonErrorResponseHandler({
-    errorSchema: zodSchema(llamaCppErrorDataSchema),
-    errorToMessage: (error) => error.error,
-});

package/model-provider/llamacpp/LlamaCppFacade.cjs

@@ -1,55 +0,0 @@
-"use strict";
-var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    var desc = Object.getOwnPropertyDescriptor(m, k);
-    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
-      desc = { enumerable: true, get: function() { return m[k]; } };
-    }
-    Object.defineProperty(o, k2, desc);
-}) : (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    o[k2] = m[k];
-}));
-var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
-    Object.defineProperty(o, "default", { enumerable: true, value: v });
-}) : function(o, v) {
-    o["default"] = v;
-});
-var __importStar = (this && this.__importStar) || function (mod) {
-    if (mod && mod.__esModule) return mod;
-    var result = {};
-    if (mod != null) for (var k in mod) if (k !== "default" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);
-    __setModuleDefault(result, mod);
-    return result;
-};
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.prompt = exports.grammar = exports.Tokenizer = exports.TextEmbedder = exports.CompletionTextGenerator = exports.Api = void 0;
-const LlamaCppApiConfiguration_js_1 = require("./LlamaCppApiConfiguration.cjs");
-const LlamaCppCompletionModel_js_1 = require("./LlamaCppCompletionModel.cjs");
-const LlamaCppTextEmbeddingModel_js_1 = require("./LlamaCppTextEmbeddingModel.cjs");
-const LlamaCppTokenizer_js_1 = require("./LlamaCppTokenizer.cjs");
-/**
- * Creates an API configuration for the Llama.cpp server.
- * It calls the API at http://127.0.0.1:8080 by default.
- */
-function Api(settings) {
-    return new LlamaCppApiConfiguration_js_1.LlamaCppApiConfiguration(settings);
-}
-exports.Api = Api;
-function CompletionTextGenerator(settings = {}) {
-    return new LlamaCppCompletionModel_js_1.LlamaCppCompletionModel(settings);
-}
-exports.CompletionTextGenerator = CompletionTextGenerator;
-function TextEmbedder(settings = {}) {
-    return new LlamaCppTextEmbeddingModel_js_1.LlamaCppTextEmbeddingModel(settings);
-}
-exports.TextEmbedder = TextEmbedder;
-function Tokenizer(api = new LlamaCppApiConfiguration_js_1.LlamaCppApiConfiguration()) {
-    return new LlamaCppTokenizer_js_1.LlamaCppTokenizer(api);
-}
-exports.Tokenizer = Tokenizer;
-/**
- * GBNF grammars. You can use them in the `grammar` option of the `TextGenerator` model.
- */
-exports.grammar = __importStar(require("./LlamaCppGrammars.cjs"));
-exports.prompt = __importStar(require("./LlamaCppPrompt.cjs"));

package/model-provider/llamacpp/LlamaCppFacade.d.ts

@@ -1,19 +0,0 @@
-import { ApiConfiguration } from "../../core/api/ApiConfiguration.js";
-import { PartialBaseUrlPartsApiConfigurationOptions } from "../../core/api/BaseUrlApiConfiguration.js";
-import { LlamaCppApiConfiguration } from "./LlamaCppApiConfiguration.js";
-import { LlamaCppCompletionModel, LlamaCppCompletionModelSettings } from "./LlamaCppCompletionModel.js";
-import { LlamaCppTextEmbeddingModel, LlamaCppTextEmbeddingModelSettings } from "./LlamaCppTextEmbeddingModel.js";
-import { LlamaCppTokenizer } from "./LlamaCppTokenizer.js";
-/**
- * Creates an API configuration for the Llama.cpp server.
- * It calls the API at http://127.0.0.1:8080 by default.
- */
-export declare function Api(settings: PartialBaseUrlPartsApiConfigurationOptions): LlamaCppApiConfiguration;
-export declare function CompletionTextGenerator<CONTEXT_WINDOW_SIZE extends number>(settings?: LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>): LlamaCppCompletionModel<CONTEXT_WINDOW_SIZE>;
-export declare function TextEmbedder(settings?: LlamaCppTextEmbeddingModelSettings): LlamaCppTextEmbeddingModel;
-export declare function Tokenizer(api?: ApiConfiguration): LlamaCppTokenizer;
-/**
- * GBNF grammars. You can use them in the `grammar` option of the `TextGenerator` model.
- */
-export * as grammar from "./LlamaCppGrammars.js";
-export * as prompt from "./LlamaCppPrompt.js";

package/model-provider/llamacpp/LlamaCppFacade.js

@@ -1,25 +0,0 @@
-import { LlamaCppApiConfiguration } from "./LlamaCppApiConfiguration.js";
-import { LlamaCppCompletionModel, } from "./LlamaCppCompletionModel.js";
-import { LlamaCppTextEmbeddingModel, } from "./LlamaCppTextEmbeddingModel.js";
-import { LlamaCppTokenizer } from "./LlamaCppTokenizer.js";
-/**
- * Creates an API configuration for the Llama.cpp server.
- * It calls the API at http://127.0.0.1:8080 by default.
- */
-export function Api(settings) {
-    return new LlamaCppApiConfiguration(settings);
-}
-export function CompletionTextGenerator(settings = {}) {
-    return new LlamaCppCompletionModel(settings);
-}
-export function TextEmbedder(settings = {}) {
-    return new LlamaCppTextEmbeddingModel(settings);
-}
-export function Tokenizer(api = new LlamaCppApiConfiguration()) {
-    return new LlamaCppTokenizer(api);
-}
-/**
- * GBNF grammars. You can use them in the `grammar` option of the `TextGenerator` model.
- */
-export * as grammar from "./LlamaCppGrammars.js";
-export * as prompt from "./LlamaCppPrompt.js";

package/model-provider/llamacpp/LlamaCppGrammars.cjs

@@ -1,86 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.fromJsonSchema = exports.list = exports.jsonArray = exports.json = void 0;
-/**
- * GBNF grammar for JSON.
- *
- * @see https://github.com/ggerganov/llama.cpp/blob/master/grammars/json.gbnf
- */
-exports.json = `
-root   ::= object
-value  ::= object | array | string | number | ("true" | "false" | "null") ws
-
-object ::=
-  "{" ws (
-            string ":" ws value
-    ("," ws string ":" ws value)*
-  )? "}" ws
-
-array  ::=
-  "[" ws (
-            value
-    ("," ws value)*
-  )? "]" ws
-
-string ::=
-  "\\"" (
-    [^"\\\\] |
-    "\\\\" (["\\\\/bfnrt] | "u" [0-9a-fA-F] [0-9a-fA-F] [0-9a-fA-F] [0-9a-fA-F]) # escapes
-  )* "\\"" ws
-
-number ::= ("-"? ([0-9] | [1-9] [0-9]*)) ("." [0-9]+)? ([eE] [-+]? [0-9]+)? ws
-
-# Optional space: by convention, applied in this grammar after literal chars when allowed
-ws ::= ([ \t\n] ws)?
-`;
-/**
- * GBNF grammar for JSON array outputs. Restricts whitespace at the end of the array.
- *
- * @see https://github.com/ggerganov/llama.cpp/blob/master/grammars/json_arr.gbnf
- */
-exports.jsonArray = `
-root   ::= arr
-value  ::= object | array | string | number | ("true" | "false" | "null") ws
-
-arr  ::=
-  "[\n" ws (
-            value
-    (",\n" ws value)*
-  )? "]"
-
-object ::=
-  "{" ws (
-            string ":" ws value
-    ("," ws string ":" ws value)*
-  )? "}" ws
-
-array  ::=
-  "[" ws (
-            value
-    ("," ws value)*
-  )? "]" ws
-
-string ::=
-  "\\"" (
-    [^"\\\\] |
-    "\\\\" (["\\\\/bfnrt] | "u" [0-9a-fA-F] [0-9a-fA-F] [0-9a-fA-F] [0-9a-fA-F]) # escapes
-  )* "\\"" ws
-
-number ::= ("-"? ([0-9] | [1-9] [0-9]*)) ("." [0-9]+)? ([eE] [-+]? [0-9]+)? ws
-
-# Optional space: by convention, applied in this grammar after literal chars when allowed
-ws ::= ([ \t\n] ws)?
-`;
-/**
- * GBNF grammar for list outputs. List items are separated by newlines and start with `- `.
- *
- * @see https://github.com/ggerganov/llama.cpp/blob/master/grammars/list.gbnf
- */
-exports.list = `
-root ::= item+
-
-# Excludes various line break characters
-item ::= "- " [^\r\n\x0b\x0c\x85\u2028\u2029]+ "\n"
-`;
-var convertJsonSchemaToGBNF_js_1 = require("./convertJsonSchemaToGBNF.cjs");
-Object.defineProperty(exports, "fromJsonSchema", { enumerable: true, get: function () { return convertJsonSchemaToGBNF_js_1.convertJsonSchemaToGBNF; } });

package/model-provider/llamacpp/LlamaCppGrammars.d.ts

@@ -1,19 +0,0 @@
-/**
- * GBNF grammar for JSON.
- *
- * @see https://github.com/ggerganov/llama.cpp/blob/master/grammars/json.gbnf
- */
-export declare const json: string;
-/**
- * GBNF grammar for JSON array outputs. Restricts whitespace at the end of the array.
- *
- * @see https://github.com/ggerganov/llama.cpp/blob/master/grammars/json_arr.gbnf
- */
-export declare const jsonArray: string;
-/**
- * GBNF grammar for list outputs. List items are separated by newlines and start with `- `.
- *
- * @see https://github.com/ggerganov/llama.cpp/blob/master/grammars/list.gbnf
- */
-export declare const list: string;
-export { convertJsonSchemaToGBNF as fromJsonSchema } from "./convertJsonSchemaToGBNF.js";