@huggingface/tasks 0.13.0 → 0.13.1-test

package/src/snippets/js.ts

@@ -1,304 +0,0 @@
-import type { PipelineType } from "../pipelines.js";
-import type { ChatCompletionInputMessage, GenerationParameters } from "../tasks/index.js";
-import { stringifyGenerationConfig, stringifyMessages } from "./common.js";
-import { getModelInputSnippet } from "./inputs.js";
-import type { InferenceSnippet, ModelDataMinimal } from "./types.js";
-
-export const snippetBasic = (model: ModelDataMinimal, accessToken: string): InferenceSnippet => ({
-	content: `async function query(data) {
-	const response = await fetch(
-		"https://api-inference.huggingface.co/models/${model.id}",
-		{
-			headers: {
-				Authorization: "Bearer ${accessToken || `{API_TOKEN}`}"
-				"Content-Type": "application/json",
-			},
-			method: "POST",
-			body: JSON.stringify(data),
-		}
-	);
-	const result = await response.json();
-	return result;
-}
-
-query({"inputs": ${getModelInputSnippet(model)}}).then((response) => {
-	console.log(JSON.stringify(response));
-});`,
-});
-
-export const snippetTextGeneration = (
-	model: ModelDataMinimal,
-	accessToken: string,
-	opts?: {
-		streaming?: boolean;
-		messages?: ChatCompletionInputMessage[];
-		temperature?: GenerationParameters["temperature"];
-		max_tokens?: GenerationParameters["max_tokens"];
-		top_p?: GenerationParameters["top_p"];
-	}
-): InferenceSnippet | InferenceSnippet[] => {
-	if (model.tags.includes("conversational")) {
-		// Conversational model detected, so we display a code snippet that features the Messages API
-		const streaming = opts?.streaming ?? true;
-		const exampleMessages = getModelInputSnippet(model) as ChatCompletionInputMessage[];
-		const messages = opts?.messages ?? exampleMessages;
-		const messagesStr = stringifyMessages(messages, { indent: "\t" });
-
-		const config = {
-			...(opts?.temperature ? { temperature: opts.temperature } : undefined),
-			max_tokens: opts?.max_tokens ?? 500,
-			...(opts?.top_p ? { top_p: opts.top_p } : undefined),
-		};
-		const configStr = stringifyGenerationConfig(config, {
-			indent: "\n\t",
-			attributeValueConnector: ": ",
-		});
-
-		if (streaming) {
-			return [
-				{
-					client: "huggingface.js",
-					content: `import { HfInference } from "@huggingface/inference"
-
-const client = new HfInference("${accessToken || `{API_TOKEN}`}")
-
-let out = "";
-
-const stream = client.chatCompletionStream({
-	model: "${model.id}",
-	messages: ${messagesStr},
-	${configStr}
-});
-
-for await (const chunk of stream) {
-	if (chunk.choices && chunk.choices.length > 0) {
-		const newContent = chunk.choices[0].delta.content;
-		out += newContent;
-		console.log(newContent);
-	}  
-}`,
-				},
-				{
-					client: "openai",
-					content: `import { OpenAI } from "openai"
-
-const client = new OpenAI({
-	baseURL: "https://api-inference.huggingface.co/v1/",
-    apiKey: "${accessToken || `{API_TOKEN}`}"
-})
-
-let out = "";
-
-const stream = await client.chat.completions.create({
-	model: "${model.id}",
-	messages: ${messagesStr},
-	${configStr},
-	stream: true,
-});
-
-for await (const chunk of stream) {
-	if (chunk.choices && chunk.choices.length > 0) {
-		const newContent = chunk.choices[0].delta.content;
-		out += newContent;
-		console.log(newContent);
-	}  
-}`,
-				},
-			];
-		} else {
-			return [
-				{
-					client: "huggingface.js",
-					content: `import { HfInference } from '@huggingface/inference'
-
-const client = new HfInference("${accessToken || `{API_TOKEN}`}")
-
-const chatCompletion = await client.chatCompletion({
-	model: "${model.id}",
-	messages: ${messagesStr},
-	${configStr}
-});
-
-console.log(chatCompletion.choices[0].message);`,
-				},
-				{
-					client: "openai",
-					content: `import { OpenAI } from "openai"
-
-const client = new OpenAI({
-    baseURL: "https://api-inference.huggingface.co/v1/",
-    apiKey: "${accessToken || `{API_TOKEN}`}"
-})
-
-const chatCompletion = await client.chat.completions.create({
-	model: "${model.id}",
-	messages: ${messagesStr},
-	${configStr}
-});
-
-console.log(chatCompletion.choices[0].message);`,
-				},
-			];
-		}
-	} else {
-		return snippetBasic(model, accessToken);
-	}
-};
-
-export const snippetZeroShotClassification = (model: ModelDataMinimal, accessToken: string): InferenceSnippet => ({
-	content: `async function query(data) {
-	const response = await fetch(
-		"https://api-inference.huggingface.co/models/${model.id}",
-		{
-			headers: {
-				Authorization: "Bearer ${accessToken || `{API_TOKEN}`}"
-				"Content-Type": "application/json",
-			},
-			method: "POST",
-			body: JSON.stringify(data),
-		}
-	);
-	const result = await response.json();
-	return result;
-}
-
-query({"inputs": ${getModelInputSnippet(
-		model
-	)}, "parameters": {"candidate_labels": ["refund", "legal", "faq"]}}).then((response) => {
-	console.log(JSON.stringify(response));
-});`,
-});
-
-export const snippetTextToImage = (model: ModelDataMinimal, accessToken: string): InferenceSnippet => ({
-	content: `async function query(data) {
-	const response = await fetch(
-		"https://api-inference.huggingface.co/models/${model.id}",
-		{
-			headers: {
-				Authorization: "Bearer ${accessToken || `{API_TOKEN}`}"
-				"Content-Type": "application/json",
-			},
-			method: "POST",
-			body: JSON.stringify(data),
-		}
-	);
-	const result = await response.blob();
-	return result;
-}
-query({"inputs": ${getModelInputSnippet(model)}}).then((response) => {
-	// Use image
-});`,
-});
-
-export const snippetTextToAudio = (model: ModelDataMinimal, accessToken: string): InferenceSnippet => {
-	const commonSnippet = `async function query(data) {
-		const response = await fetch(
-			"https://api-inference.huggingface.co/models/${model.id}",
-			{
-				headers: {
-					Authorization: "Bearer ${accessToken || `{API_TOKEN}`}"
-					"Content-Type": "application/json",
-				},
-				method: "POST",
-				body: JSON.stringify(data),
-			}
-		);`;
-	if (model.library_name === "transformers") {
-		return {
-			content:
-				commonSnippet +
-				`
-			const result = await response.blob();
-			return result;
-		}
-		query({"inputs": ${getModelInputSnippet(model)}}).then((response) => {
-			// Returns a byte object of the Audio wavform. Use it directly!
-		});`,
-		};
-	} else {
-		return {
-			content:
-				commonSnippet +
-				`
-			const result = await response.json();
-			return result;
-		}
-		
-		query({"inputs": ${getModelInputSnippet(model)}}).then((response) => {
-			console.log(JSON.stringify(response));
-		});`,
-		};
-	}
-};
-
-export const snippetFile = (model: ModelDataMinimal, accessToken: string): InferenceSnippet => ({
-	content: `async function query(filename) {
-	const data = fs.readFileSync(filename);
-	const response = await fetch(
-		"https://api-inference.huggingface.co/models/${model.id}",
-		{
-			headers: {
-				Authorization: "Bearer ${accessToken || `{API_TOKEN}`}"
-				"Content-Type": "application/json",
-			},
-			method: "POST",
-			body: data,
-		}
-	);
-	const result = await response.json();
-	return result;
-}
-
-query(${getModelInputSnippet(model)}).then((response) => {
-	console.log(JSON.stringify(response));
-});`,
-});
-
-export const jsSnippets: Partial<
-	Record<
-		PipelineType,
-		(
-			model: ModelDataMinimal,
-			accessToken: string,
-			opts?: Record<string, unknown>
-		) => InferenceSnippet | InferenceSnippet[]
-	>
-> = {
-	// Same order as in js/src/lib/interfaces/Types.ts
-	"text-classification": snippetBasic,
-	"token-classification": snippetBasic,
-	"table-question-answering": snippetBasic,
-	"question-answering": snippetBasic,
-	"zero-shot-classification": snippetZeroShotClassification,
-	translation: snippetBasic,
-	summarization: snippetBasic,
-	"feature-extraction": snippetBasic,
-	"text-generation": snippetTextGeneration,
-	"image-text-to-text": snippetTextGeneration,
-	"text2text-generation": snippetBasic,
-	"fill-mask": snippetBasic,
-	"sentence-similarity": snippetBasic,
-	"automatic-speech-recognition": snippetFile,
-	"text-to-image": snippetTextToImage,
-	"text-to-speech": snippetTextToAudio,
-	"text-to-audio": snippetTextToAudio,
-	"audio-to-audio": snippetFile,
-	"audio-classification": snippetFile,
-	"image-classification": snippetFile,
-	"image-to-text": snippetFile,
-	"object-detection": snippetFile,
-	"image-segmentation": snippetFile,
-};
-
-export function getJsInferenceSnippet(
-	model: ModelDataMinimal,
-	accessToken: string
-): InferenceSnippet | InferenceSnippet[] {
-	return model.pipeline_tag && model.pipeline_tag in jsSnippets
-		? jsSnippets[model.pipeline_tag]?.(model, accessToken) ?? { content: "" }
-		: { content: "" };
-}
-
-export function hasJsInferenceSnippet(model: ModelDataMinimal): boolean {
-	return !!model.pipeline_tag && model.pipeline_tag in jsSnippets;
-}

package/src/snippets/python.spec.ts

@@ -1,78 +0,0 @@
-import type { ModelDataMinimal } from "./types";
-import { describe, expect, it } from "vitest";
-import { snippetConversational } from "./python";
-
-describe("inference API snippets", () => {
-	it("conversational llm", async () => {
-		const model: ModelDataMinimal = {
-			id: "meta-llama/Llama-3.1-8B-Instruct",
-			pipeline_tag: "text-generation",
-			tags: ["conversational"],
-			inference: "",
-		};
-		const snippet = snippetConversational(model, "api_token");
-
-		expect(snippet[0].content).toEqual(`from huggingface_hub import InferenceClient
-
-client = InferenceClient(api_key="api_token")
-
-messages = [
-	{
-		"role": "user",
-		"content": "What is the capital of France?"
-	}
-]
-
-stream = client.chat.completions.create(
-    model="meta-llama/Llama-3.1-8B-Instruct", 
-	messages=messages, 
-	max_tokens=500,
-	stream=True
-)
-
-for chunk in stream:
-    print(chunk.choices[0].delta.content, end="")`);
-	});
-
-	it("conversational vlm", async () => {
-		const model: ModelDataMinimal = {
-			id: "meta-llama/Llama-3.2-11B-Vision-Instruct",
-			pipeline_tag: "image-text-to-text",
-			tags: ["conversational"],
-			inference: "",
-		};
-		const snippet = snippetConversational(model, "api_token");
-
-		expect(snippet[0].content).toEqual(`from huggingface_hub import InferenceClient
-
-client = InferenceClient(api_key="api_token")
-
-messages = [
-	{
-		"role": "user",
-		"content": [
-			{
-				"type": "text",
-				"text": "Describe this image in one sentence."
-			},
-			{
-				"type": "image_url",
-				"image_url": {
-					"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
-				}
-			}
-		]
-	}
-]
-
-stream = client.chat.completions.create(
-    model="meta-llama/Llama-3.2-11B-Vision-Instruct", 
-	messages=messages, 
-	max_tokens=500,
-	stream=True
-)
-
-for chunk in stream:
-    print(chunk.choices[0].delta.content, end="")`);
-	});
-});

package/src/snippets/python.ts

@@ -1,304 +0,0 @@
-import type { PipelineType } from "../pipelines.js";
-import type { ChatCompletionInputMessage, GenerationParameters } from "../tasks/index.js";
-import { stringifyGenerationConfig, stringifyMessages } from "./common.js";
-import { getModelInputSnippet } from "./inputs.js";
-import type { InferenceSnippet, ModelDataMinimal } from "./types.js";
-
-export const snippetConversational = (
-	model: ModelDataMinimal,
-	accessToken: string,
-	opts?: {
-		streaming?: boolean;
-		messages?: ChatCompletionInputMessage[];
-		temperature?: GenerationParameters["temperature"];
-		max_tokens?: GenerationParameters["max_tokens"];
-		top_p?: GenerationParameters["top_p"];
-	}
-): InferenceSnippet[] => {
-	const streaming = opts?.streaming ?? true;
-	const exampleMessages = getModelInputSnippet(model) as ChatCompletionInputMessage[];
-	const messages = opts?.messages ?? exampleMessages;
-	const messagesStr = stringifyMessages(messages, { attributeKeyQuotes: true });
-
-	const config = {
-		...(opts?.temperature ? { temperature: opts.temperature } : undefined),
-		max_tokens: opts?.max_tokens ?? 500,
-		...(opts?.top_p ? { top_p: opts.top_p } : undefined),
-	};
-	const configStr = stringifyGenerationConfig(config, {
-		indent: "\n\t",
-		attributeValueConnector: "=",
-	});
-
-	if (streaming) {
-		return [
-			{
-				client: "huggingface_hub",
-				content: `from huggingface_hub import InferenceClient
-
-client = InferenceClient(api_key="${accessToken || "{API_TOKEN}"}")
-
-messages = ${messagesStr}
-
-stream = client.chat.completions.create(
-    model="${model.id}", 
-	messages=messages, 
-	${configStr},
-	stream=True
-)
-
-for chunk in stream:
-    print(chunk.choices[0].delta.content, end="")`,
-			},
-			{
-				client: "openai",
-				content: `from openai import OpenAI
-
-client = OpenAI(
-	base_url="https://api-inference.huggingface.co/v1/",
-	api_key="${accessToken || "{API_TOKEN}"}"
-)
-
-messages = ${messagesStr}
-
-stream = client.chat.completions.create(
-    model="${model.id}", 
-	messages=messages, 
-	${configStr},
-	stream=True
-)
-
-for chunk in stream:
-    print(chunk.choices[0].delta.content, end="")`,
-			},
-		];
-	} else {
-		return [
-			{
-				client: "huggingface_hub",
-				content: `from huggingface_hub import InferenceClient
-
-client = InferenceClient(api_key="${accessToken || "{API_TOKEN}"}")
-
-messages = ${messagesStr}
-
-completion = client.chat.completions.create(
-    model="${model.id}", 
-	messages=messages, 
-	${configStr}
-)
-
-print(completion.choices[0].message)`,
-			},
-			{
-				client: "openai",
-				content: `from openai import OpenAI
-
-client = OpenAI(
-	base_url="https://api-inference.huggingface.co/v1/",
-	api_key="${accessToken || "{API_TOKEN}"}"
-)
-
-messages = ${messagesStr}
-
-completion = client.chat.completions.create(
-    model="${model.id}", 
-	messages=messages, 
-	${configStr}
-)
-
-print(completion.choices[0].message)`,
-			},
-		];
-	}
-};
-
-export const snippetZeroShotClassification = (model: ModelDataMinimal): InferenceSnippet => ({
-	content: `def query(payload):
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.json()
-
-output = query({
-    "inputs": ${getModelInputSnippet(model)},
-    "parameters": {"candidate_labels": ["refund", "legal", "faq"]},
-})`,
-});
-
-export const snippetZeroShotImageClassification = (model: ModelDataMinimal): InferenceSnippet => ({
-	content: `def query(data):
-	with open(data["image_path"], "rb") as f:
-		img = f.read()
-	payload={
-		"parameters": data["parameters"],
-		"inputs": base64.b64encode(img).decode("utf-8")
-	}
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.json()
-
-output = query({
-    "image_path": ${getModelInputSnippet(model)},
-    "parameters": {"candidate_labels": ["cat", "dog", "llama"]},
-})`,
-});
-
-export const snippetBasic = (model: ModelDataMinimal): InferenceSnippet => ({
-	content: `def query(payload):
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.json()
-	
-output = query({
-	"inputs": ${getModelInputSnippet(model)},
-})`,
-});
-
-export const snippetFile = (model: ModelDataMinimal): InferenceSnippet => ({
-	content: `def query(filename):
-    with open(filename, "rb") as f:
-        data = f.read()
-    response = requests.post(API_URL, headers=headers, data=data)
-    return response.json()
-
-output = query(${getModelInputSnippet(model)})`,
-});
-
-export const snippetTextToImage = (model: ModelDataMinimal): InferenceSnippet => ({
-	content: `def query(payload):
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.content
-image_bytes = query({
-	"inputs": ${getModelInputSnippet(model)},
-})
-# You can access the image with PIL.Image for example
-import io
-from PIL import Image
-image = Image.open(io.BytesIO(image_bytes))`,
-});
-
-export const snippetTabular = (model: ModelDataMinimal): InferenceSnippet => ({
-	content: `def query(payload):
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.content
-response = query({
-	"inputs": {"data": ${getModelInputSnippet(model)}},
-})`,
-});
-
-export const snippetTextToAudio = (model: ModelDataMinimal): InferenceSnippet => {
-	// Transformers TTS pipeline and api-inference-community (AIC) pipeline outputs are diverged
-	// with the latest update to inference-api (IA).
-	// Transformers IA returns a byte object (wav file), whereas AIC returns wav and sampling_rate.
-	if (model.library_name === "transformers") {
-		return {
-			content: `def query(payload):
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.content
-
-audio_bytes = query({
-	"inputs": ${getModelInputSnippet(model)},
-})
-# You can access the audio with IPython.display for example
-from IPython.display import Audio
-Audio(audio_bytes)`,
-		};
-	} else {
-		return {
-			content: `def query(payload):
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.json()
-	
-audio, sampling_rate = query({
-	"inputs": ${getModelInputSnippet(model)},
-})
-# You can access the audio with IPython.display for example
-from IPython.display import Audio
-Audio(audio, rate=sampling_rate)`,
-		};
-	}
-};
-
-export const snippetDocumentQuestionAnswering = (model: ModelDataMinimal): InferenceSnippet => ({
-	content: `def query(payload):
- 	with open(payload["image"], "rb") as f:
-  		img = f.read()
-		payload["image"] = base64.b64encode(img).decode("utf-8")  
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.json()
-
-output = query({
-    "inputs": ${getModelInputSnippet(model)},
-})`,
-});
-
-export const pythonSnippets: Partial<
-	Record<
-		PipelineType,
-		(
-			model: ModelDataMinimal,
-			accessToken: string,
-			opts?: Record<string, unknown>
-		) => InferenceSnippet | InferenceSnippet[]
-	>
-> = {
-	// Same order as in tasks/src/pipelines.ts
-	"text-classification": snippetBasic,
-	"token-classification": snippetBasic,
-	"table-question-answering": snippetBasic,
-	"question-answering": snippetBasic,
-	"zero-shot-classification": snippetZeroShotClassification,
-	translation: snippetBasic,
-	summarization: snippetBasic,
-	"feature-extraction": snippetBasic,
-	"text-generation": snippetBasic,
-	"text2text-generation": snippetBasic,
-	"image-text-to-text": snippetConversational,
-	"fill-mask": snippetBasic,
-	"sentence-similarity": snippetBasic,
-	"automatic-speech-recognition": snippetFile,
-	"text-to-image": snippetTextToImage,
-	"text-to-speech": snippetTextToAudio,
-	"text-to-audio": snippetTextToAudio,
-	"audio-to-audio": snippetFile,
-	"audio-classification": snippetFile,
-	"image-classification": snippetFile,
-	"tabular-regression": snippetTabular,
-	"tabular-classification": snippetTabular,
-	"object-detection": snippetFile,
-	"image-segmentation": snippetFile,
-	"document-question-answering": snippetDocumentQuestionAnswering,
-	"image-to-text": snippetFile,
-	"zero-shot-image-classification": snippetZeroShotImageClassification,
-};
-
-export function getPythonInferenceSnippet(
-	model: ModelDataMinimal,
-	accessToken: string,
-	opts?: Record<string, unknown>
-): InferenceSnippet | InferenceSnippet[] {
-	if (model.tags.includes("conversational")) {
-		// Conversational model detected, so we display a code snippet that features the Messages API
-		return snippetConversational(model, accessToken, opts);
-	} else {
-		let snippets =
-			model.pipeline_tag && model.pipeline_tag in pythonSnippets
-				? pythonSnippets[model.pipeline_tag]?.(model, accessToken) ?? { content: "" }
-				: { content: "" };
-
-		snippets = Array.isArray(snippets) ? snippets : [snippets];
-
-		return snippets.map((snippet) => {
-			return {
-				...snippet,
-				content: `import requests
-
-API_URL = "https://api-inference.huggingface.co/models/${model.id}"
-headers = {"Authorization": ${accessToken ? `"Bearer ${accessToken}"` : `f"Bearer {API_TOKEN}"`}}
-
-${snippet.content}`,
-			};
-		});
-	}
-}
-
-export function hasPythonInferenceSnippet(model: ModelDataMinimal): boolean {
-	return !!model.pipeline_tag && model.pipeline_tag in pythonSnippets;
-}

package/src/snippets/types.ts

@@ -1,16 +0,0 @@
-import type { ModelData } from "../model-data";
-
-/**
- * Minimal model data required for snippets.
- *
- * Add more fields as needed.
- */
-export type ModelDataMinimal = Pick<
-	ModelData,
-	"id" | "pipeline_tag" | "mask_token" | "library_name" | "config" | "tags" | "inference"
->;
-
-export interface InferenceSnippet {
-	content: string;
-	client?: string; // for instance: `client` could be `huggingface_hub` or `openai` client for Python snippets
-}