modelfusion 0.75.0 → 0.77.0

package/README.md

@@ -50,6 +50,8 @@ You can use [prompt formats](https://modelfusion.dev/guide/function/generate-tex
 #### generateText
 
 ```ts
+import { generateText, openai } from "modelfusion";
+
 const text = await generateText(
   openai.CompletionTextGenerator({ model: "gpt-3.5-turbo-instruct" }),
   "Write a short story about a robot learning to love:\n\n"
@@ -61,6 +63,8 @@ Providers: [OpenAI](https://modelfusion.dev/integration/model-provider/openai),
 #### streamText
 
 ```ts
+import { streamText, openai } from "modelfusion";
+
 const textStream = await streamText(
   openai.CompletionTextGenerator({ model: "gpt-3.5-turbo-instruct" }),
   "Write a short story about a robot learning to love:\n\n"
@@ -78,6 +82,8 @@ Providers: [OpenAI](https://modelfusion.dev/integration/model-provider/openai),
 Multi-modal vision models such as GPT 4 Vision can process images as part of the prompt.
 
 ```ts
+import { streamText, openai } from "modelfusion";
+
 const textStream = await streamText(
   openai.ChatTextGenerator({ model: "gpt-4-vision-preview" }),
   [
@@ -95,6 +101,8 @@ Providers: [OpenAI](https://modelfusion.dev/integration/model-provider/openai),
 Generate an image from a prompt.
 
 ```ts
+import { generateImage, openai } from "modelfusion";
+
 const image = await generateImage(
   openai.ImageGenerator({ model: "dall-e-3", size: "1024x1024" }),
   "the wicked witch of the west in the style of early 19th century painting"
@@ -112,9 +120,11 @@ Synthesize speech (audio) from text. Also called TTS (text-to-speech).
 `generateSpeech` synthesizes speech from text.
 
 ```ts
+import { generateSpeech, lmnt } from "modelfusion";
+
 // `speech` is a Buffer with MP3 audio data
 const speech = await generateSpeech(
-  lmnt.Speech({
+  lmnt.SpeechGenerator({
     voice: "034b632b-df71-46c8-b440-86a42ffc3cf3", // Henry
   }),
   "Good evening, ladies and gentlemen! Exciting news on the airwaves tonight " +
@@ -131,10 +141,12 @@ Providers: [Eleven Labs](https://modelfusion.dev/integration/model-provider/elev
 `generateSpeech` generates a stream of speech chunks from text or from a text stream. Depending on the model, this can be fully duplex.
 
 ```ts
-const textStream = await streamText(/* ... */);
+import { streamSpeech, elevenlabs } from "modelfusion";
+
+const textStream: AsyncIterable<string>;
 
 const speechStream = await streamSpeech(
-  elevenlabs.Speech({
+  elevenlabs.SpeechGenerator({
     model: "eleven_turbo_v2",
     voice: "pNInz6obpgDQGcFmaJgB", // Adam
     optimizeStreamingLatency: 1,
@@ -158,8 +170,10 @@ Providers: [Eleven Labs](https://modelfusion.dev/integration/model-provider/elev
 Transcribe speech (audio) data into text. Also called speech-to-text (STT).
 
 ```ts
+import { generateTranscription, openai } from "modelfusion";
+
 const transcription = await generateTranscription(
-  openai.Transcription({ model: "whisper-1" }),
+  openai.Transcriber({ model: "whisper-1" }),
   {
     type: "mp3",
     data: await fs.promises.readFile("data/test.mp3"),

package/model-function/generate-speech/generateSpeech.d.ts

@@ -9,7 +9,7 @@ import { SpeechGenerationModel, SpeechGenerationModelSettings } from "./SpeechGe
  *
  * @example
  * const speech = await generateSpeech(
- *   lmnt.Speech(...),
+ *   lmnt.SpeechGenerator(...),
  *   "Good evening, ladies and gentlemen! Exciting news on the airwaves tonight " +
  *    "as The Rolling Stones unveil 'Hackney Diamonds.'
  * );

package/model-function/generate-speech/streamSpeech.d.ts

@@ -12,7 +12,7 @@ import { SpeechGenerationModelSettings, StreamingSpeechGenerationModel } from ".
  * const textStream = await streamText(...);
  *
  * const speechStream = await streamSpeech(
- *   elevenlabs.Speech(...),
+ *   elevenlabs.SpeechGenerator(...),
  *   textStream
  * );
  *

package/model-function/generate-transcription/generateTranscription.d.ts

@@ -10,7 +10,7 @@ import { TranscriptionModel, TranscriptionModelSettings } from "./TranscriptionM
  * const data = await fs.promises.readFile("data/test.mp3");
  *
  * const transcription = await generateTranscription(
- *   openai.Transcription({ model: "whisper-1" }),
+ *   openai.Transcriber({ model: "whisper-1" }),
  *   { type: "mp3", data }
  * );
  *

package/model-provider/anthropic/AnthropicPromptFormat.cjs

@@ -24,10 +24,7 @@ exports.text = text;
 function instruction() {
     return {
         format: (instruction) => {
-            let text = "";
-            if (instruction.system != null) {
-                text += `${instruction.system}`;
-            }
+            let text = instruction.system ?? "";
             text += "\n\nHuman:";
             text += instruction.instruction;
             text += "\n\nAssistant:";
@@ -46,7 +43,7 @@ function chat() {
     return {
         format: (chatPrompt) => {
             (0, validateChatPrompt_js_1.validateChatPrompt)(chatPrompt);
-            let text = chatPrompt.system != null ? `${chatPrompt.system}\n\n` : "";
+            let text = chatPrompt.system ?? "";
             for (const { role, content } of chatPrompt.messages) {
                 switch (role) {
                     case "user": {

package/model-provider/anthropic/AnthropicPromptFormat.js

@@ -20,10 +20,7 @@ export function text() {
 export function instruction() {
     return {
         format: (instruction) => {
-            let text = "";
-            if (instruction.system != null) {
-                text += `${instruction.system}`;
-            }
+            let text = instruction.system ?? "";
             text += "\n\nHuman:";
             text += instruction.instruction;
             text += "\n\nAssistant:";
@@ -41,7 +38,7 @@ export function chat() {
     return {
         format: (chatPrompt) => {
             validateChatPrompt(chatPrompt);
-            let text = chatPrompt.system != null ? `${chatPrompt.system}\n\n` : "";
+            let text = chatPrompt.system ?? "";
             for (const { role, content } of chatPrompt.messages) {
                 switch (role) {
                     case "user": {

package/model-provider/anthropic/AnthropicTextGenerationModel.cjs

@@ -21,11 +21,14 @@ exports.ANTHROPIC_TEXT_GENERATION_MODELS = {
         contextWindowSize: 100000,
     },
     "claude-2": {
-        contextWindowSize: 100000,
+        contextWindowSize: 200000,
     },
     "claude-2.0": {
         contextWindowSize: 100000,
     },
+    "claude-2.1": {
+        contextWindowSize: 200000,
+    },
 };
 /**
  * Create a text generation model that calls the Anthropic API.

package/model-provider/anthropic/AnthropicTextGenerationModel.d.ts

@@ -20,6 +20,9 @@ export declare const ANTHROPIC_TEXT_GENERATION_MODELS: {
     "claude-2.0": {
         contextWindowSize: number;
     };
+    "claude-2.1": {
+        contextWindowSize: number;
+    };
 };
 export type AnthropicTextGenerationModelType = keyof typeof ANTHROPIC_TEXT_GENERATION_MODELS;
 export interface AnthropicTextGenerationModelSettings extends TextGenerationModelSettings {
@@ -38,7 +41,7 @@ export interface AnthropicTextGenerationModelSettings extends TextGenerationMode
 export declare class AnthropicTextGenerationModel extends AbstractModel<AnthropicTextGenerationModelSettings> implements TextStreamingModel<string, AnthropicTextGenerationModelSettings> {
     constructor(settings: AnthropicTextGenerationModelSettings);
     readonly provider: "anthropic";
-    get modelName(): "claude-instant-1" | "claude-instant-1.2" | "claude-2" | "claude-2.0";
+    get modelName(): "claude-instant-1" | "claude-instant-1.2" | "claude-2" | "claude-2.0" | "claude-2.1";
     readonly contextWindowSize: number;
     readonly tokenizer: undefined;
     readonly countPromptTokens: undefined;

package/model-provider/anthropic/AnthropicTextGenerationModel.js

@@ -18,11 +18,14 @@ export const ANTHROPIC_TEXT_GENERATION_MODELS = {
         contextWindowSize: 100000,
     },
     "claude-2": {
-        contextWindowSize: 100000,
+        contextWindowSize: 200000,
     },
     "claude-2.0": {
         contextWindowSize: 100000,
     },
+    "claude-2.1": {
+        contextWindowSize: 200000,
+    },
 };
 /**
  * Create a text generation model that calls the Anthropic API.

package/model-provider/elevenlabs/ElevenLabsFacade.cjs

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.Speech = void 0;
+exports.SpeechGenerator = void 0;
 const ElevenLabsSpeechModel_js_1 = require("./ElevenLabsSpeechModel.cjs");
 /**
  * Synthesize speech using the ElevenLabs Text to Speech API.
@@ -12,7 +12,7 @@ const ElevenLabsSpeechModel_js_1 = require("./ElevenLabsSpeechModel.cjs");
  *
  * @returns A new instance of {@link ElevenLabsSpeechModel}.
  */
-function Speech(settings) {
+function SpeechGenerator(settings) {
     return new ElevenLabsSpeechModel_js_1.ElevenLabsSpeechModel(settings);
 }
-exports.Speech = Speech;
+exports.SpeechGenerator = SpeechGenerator;

package/model-provider/elevenlabs/ElevenLabsFacade.d.ts

@@ -9,4 +9,4 @@ import { ElevenLabsSpeechModel, ElevenLabsSpeechModelSettings } from "./ElevenLa
  *
  * @returns A new instance of {@link ElevenLabsSpeechModel}.
  */
-export declare function Speech(settings: ElevenLabsSpeechModelSettings): ElevenLabsSpeechModel;
+export declare function SpeechGenerator(settings: ElevenLabsSpeechModelSettings): ElevenLabsSpeechModel;

package/model-provider/elevenlabs/ElevenLabsFacade.js

@@ -9,6 +9,6 @@ import { ElevenLabsSpeechModel, } from "./ElevenLabsSpeechModel.js";
  *
  * @returns A new instance of {@link ElevenLabsSpeechModel}.
  */
-export function Speech(settings) {
+export function SpeechGenerator(settings) {
     return new ElevenLabsSpeechModel(settings);
 }

package/model-provider/lmnt/LmntFacade.cjs

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.Speech = void 0;
+exports.SpeechGenerator = void 0;
 const LmntSpeechModel_js_1 = require("./LmntSpeechModel.cjs");
 /**
  * Synthesize speech using the LMNT API.
@@ -9,7 +9,7 @@ const LmntSpeechModel_js_1 = require("./LmntSpeechModel.cjs");
  *
  * @returns A new instance of {@link LmntSpeechModel}.
  */
-function Speech(settings) {
+function SpeechGenerator(settings) {
     return new LmntSpeechModel_js_1.LmntSpeechModel(settings);
 }
-exports.Speech = Speech;
+exports.SpeechGenerator = SpeechGenerator;

package/model-provider/lmnt/LmntFacade.d.ts

@@ -6,4 +6,4 @@ import { LmntSpeechModel, LmntSpeechModelSettings } from "./LmntSpeechModel.js";
  *
  * @returns A new instance of {@link LmntSpeechModel}.
  */
-export declare function Speech(settings: LmntSpeechModelSettings): LmntSpeechModel;
+export declare function SpeechGenerator(settings: LmntSpeechModelSettings): LmntSpeechModel;

package/model-provider/lmnt/LmntFacade.js

@@ -6,6 +6,6 @@ import { LmntSpeechModel } from "./LmntSpeechModel.js";
  *
  * @returns A new instance of {@link LmntSpeechModel}.
  */
-export function Speech(settings) {
+export function SpeechGenerator(settings) {
     return new LmntSpeechModel(settings);
 }

package/model-provider/openai/OpenAIFacade.cjs

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.Tokenizer = exports.ImageGenerator = exports.Transcription = exports.Speech = exports.TextEmbedder = exports.ChatTextGenerator = exports.CompletionTextGenerator = void 0;
+exports.Tokenizer = exports.ImageGenerator = exports.Transcriber = exports.SpeechGenerator = exports.TextEmbedder = exports.ChatTextGenerator = exports.CompletionTextGenerator = void 0;
 const OpenAICompletionModel_js_1 = require("./OpenAICompletionModel.cjs");
 const OpenAIImageGenerationModel_js_1 = require("./OpenAIImageGenerationModel.cjs");
 const OpenAISpeechModel_js_1 = require("./OpenAISpeechModel.cjs");
@@ -82,10 +82,10 @@ exports.TextEmbedder = TextEmbedder;
  *
  * @returns A new instance of {@link OpenAISpeechModel}.
  */
-function Speech(settings) {
+function SpeechGenerator(settings) {
     return new OpenAISpeechModel_js_1.OpenAISpeechModel(settings);
 }
-exports.Speech = Speech;
+exports.SpeechGenerator = SpeechGenerator;
 /**
  * Create a transcription model that calls the OpenAI transcription API.
  *
@@ -95,7 +95,7 @@ exports.Speech = Speech;
  * const data = await fs.promises.readFile("data/test.mp3");
  *
  * const transcription = await transcribe(
- *   openai.Transcription({ model: "whisper-1" }),
+ *   openai.Transcriber({ model: "whisper-1" }),
  *   {
  *     type: "mp3",
  *     data,
@@ -104,10 +104,10 @@ exports.Speech = Speech;
  *
  * @returns A new instance of {@link OpenAITranscriptionModel}.
  */
-function Transcription(settings) {
+function Transcriber(settings) {
     return new OpenAITranscriptionModel_js_1.OpenAITranscriptionModel(settings);
 }
-exports.Transcription = Transcription;
+exports.Transcriber = Transcriber;
 /**
  * Create an image generation model that calls the OpenAI AI image creation API.
  *

package/model-provider/openai/OpenAIFacade.d.ts

@@ -70,7 +70,7 @@ export declare function TextEmbedder(settings: OpenAITextEmbeddingModelSettings)
  *
  * @returns A new instance of {@link OpenAISpeechModel}.
  */
-export declare function Speech(settings: OpenAISpeechModelSettings): OpenAISpeechModel;
+export declare function SpeechGenerator(settings: OpenAISpeechModelSettings): OpenAISpeechModel;
 /**
  * Create a transcription model that calls the OpenAI transcription API.
  *
@@ -80,7 +80,7 @@ export declare function Speech(settings: OpenAISpeechModelSettings): OpenAISpeec
  * const data = await fs.promises.readFile("data/test.mp3");
  *
  * const transcription = await transcribe(
- *   openai.Transcription({ model: "whisper-1" }),
+ *   openai.Transcriber({ model: "whisper-1" }),
  *   {
  *     type: "mp3",
  *     data,
@@ -89,7 +89,7 @@ export declare function Speech(settings: OpenAISpeechModelSettings): OpenAISpeec
  *
  * @returns A new instance of {@link OpenAITranscriptionModel}.
  */
-export declare function Transcription(settings: OpenAITranscriptionModelSettings): OpenAITranscriptionModel;
+export declare function Transcriber(settings: OpenAITranscriptionModelSettings): OpenAITranscriptionModel;
 /**
  * Create an image generation model that calls the OpenAI AI image creation API.
  *

package/model-provider/openai/OpenAIFacade.js

@@ -76,7 +76,7 @@ export function TextEmbedder(settings) {
  *
  * @returns A new instance of {@link OpenAISpeechModel}.
  */
-export function Speech(settings) {
+export function SpeechGenerator(settings) {
     return new OpenAISpeechModel(settings);
 }
 /**
@@ -88,7 +88,7 @@ export function Speech(settings) {
  * const data = await fs.promises.readFile("data/test.mp3");
  *
  * const transcription = await transcribe(
- *   openai.Transcription({ model: "whisper-1" }),
+ *   openai.Transcriber({ model: "whisper-1" }),
  *   {
  *     type: "mp3",
  *     data,
@@ -97,7 +97,7 @@ export function Speech(settings) {
  *
  * @returns A new instance of {@link OpenAITranscriptionModel}.
  */
-export function Transcription(settings) {
+export function Transcriber(settings) {
     return new OpenAITranscriptionModel(settings);
 }
 /**

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "modelfusion",
   "description": "The TypeScript library for building multi-modal AI applications.",
-  "version": "0.75.0",
+  "version": "0.77.0",
   "author": "Lars Grammel",
   "license": "MIT",
   "keywords": [