n8n-nodes-piapi
Version:
Community n8n nodes for PiAPI - integrate generative AI capabilities (image, video, audio, 3D) into your workflows
343 lines • 13.9 kB
JavaScript
;
Object.defineProperty(exports, "__esModule", { value: true });
exports.LLMConversationalImage = void 0;
const GenericFunctions_1 = require("../shared/GenericFunctions");
const Constants_1 = require("../shared/Constants");
function generatePromptWithAspectRatio(executeFunctions, itemIndex, prompt) {
const aspectRatio = executeFunctions.getNodeParameter('aspectRatio', itemIndex, 'default');
const imageStyle = executeFunctions.getNodeParameter('imageStyle', itemIndex, 'none');
let aspectRatioText = '';
if (aspectRatio !== 'default') {
if (aspectRatio === 'custom') {
const width = executeFunctions.getNodeParameter('width', itemIndex, 1024);
const height = executeFunctions.getNodeParameter('height', itemIndex, 1024);
aspectRatioText = `Image size: ${width}x${height}. `;
}
else if (aspectRatio !== 'square_header' && aspectRatio !== 'landscape_header' && aspectRatio !== 'portrait_header') {
aspectRatioText = `Image size: ${aspectRatio}. `;
}
}
let styleText = '';
if (imageStyle !== 'none') {
styleText = `Image style: ${imageStyle}. `;
}
return `${aspectRatioText}${styleText}${prompt}`;
}
class LLMConversationalImage {
constructor() {
this.description = {
displayName: 'PiAPI GPT-4o Conversational Image',
name: 'llmConversationalImage',
icon: 'file:../piapi.svg',
group: ['transform'],
version: 1,
description: 'Edit and generate images through conversational interaction with GPT-4o',
defaults: {
name: 'GPT-4o Conversational Image',
},
inputs: ["main"],
outputs: ["main"],
credentials: [
{
name: 'piAPIApi',
required: true,
},
],
properties: [
{
displayName: 'Model',
name: 'model',
type: 'options',
options: [
{
name: 'GPT-4o Image Preview',
value: 'gpt-4o-image-preview',
},
],
default: 'gpt-4o-image-preview',
description: 'The model to use for conversational image editing',
},
{
displayName: 'Operation',
name: 'operation',
type: 'options',
options: [
{
name: 'New Conversation',
value: 'new',
},
{
name: 'Continue Conversation',
value: 'continue',
},
],
default: 'new',
description: 'Whether to start a new conversation or continue an existing one',
},
{
displayName: 'Image Source',
name: 'imageSource',
type: 'options',
options: [
{
name: 'None (Text Only)',
value: 'none',
},
{
name: 'URL',
value: 'url',
},
{
name: 'Binary Data',
value: 'binaryData',
},
],
default: 'none',
description: 'The source of the input image (if any)',
displayOptions: {
show: {
operation: ['new'],
},
},
},
{
displayName: 'Image URL',
name: 'imageUrl',
type: 'string',
default: '',
required: true,
displayOptions: {
show: {
imageSource: ['url'],
operation: ['new'],
},
},
description: 'URL of the image to start with',
},
{
displayName: 'Binary Property',
name: 'binaryPropertyName',
type: 'string',
default: 'data',
required: true,
displayOptions: {
show: {
imageSource: ['binaryData'],
operation: ['new'],
},
},
description: 'Name of the binary property containing the image data',
},
{
displayName: 'Prompt',
name: 'prompt',
type: 'string',
typeOptions: {
rows: 4,
},
default: '',
required: true,
description: 'Text prompt for image generation/editing',
},
{
displayName: 'Aspect Ratio',
name: 'aspectRatio',
type: 'options',
options: [
{
name: 'Default (No Specification)',
value: 'default',
},
...Constants_1.ASPECT_RATIO_OPTIONS.filter(option => option.value !== 'square_header' &&
option.value !== 'landscape_header' &&
option.value !== 'portrait_header'),
],
default: 'default',
description: 'Aspect ratio for new image generation (only applies to new conversations or when generating new images)',
},
{
displayName: 'Custom Width',
name: 'width',
type: 'number',
displayOptions: {
show: {
aspectRatio: ['custom'],
},
},
default: 1024,
description: 'Custom width of the generated image',
},
{
displayName: 'Custom Height',
name: 'height',
type: 'number',
displayOptions: {
show: {
aspectRatio: ['custom'],
},
},
default: 1024,
description: 'Custom height of the generated image',
},
{
displayName: 'Image Style',
name: 'imageStyle',
type: 'options',
options: Constants_1.LORA_OPTIONS,
default: 'none',
description: 'Style to apply to the generated image',
},
{
displayName: 'Previous Messages',
name: 'previousMessages',
type: 'json',
typeOptions: {
rows: 4,
},
default: '[]',
required: true,
displayOptions: {
show: {
operation: ['continue'],
},
},
description: 'JSON array of previous conversation messages',
},
{
displayName: 'Previous Images',
name: 'previousImages',
type: 'json',
typeOptions: {
rows: 4,
},
default: '[]',
displayOptions: {
show: {
operation: ['continue'],
},
},
description: 'JSON array of previous image URLs (optional)',
},
{
displayName: 'Wait for Completion',
name: 'waitForCompletion',
type: 'boolean',
default: false,
description: 'Wait for task to complete and return results',
},
],
};
}
async execute() {
var _a;
const items = this.getInputData();
const returnData = [];
for (let i = 0; i < items.length; i++) {
const model = this.getNodeParameter('model', i);
const operation = this.getNodeParameter('operation', i);
const prompt = this.getNodeParameter('prompt', i);
const waitForCompletion = this.getNodeParameter('waitForCompletion', i, true);
let messages = [];
if (operation === 'new') {
const imageSource = this.getNodeParameter('imageSource', i);
if (imageSource === 'none') {
messages = [
{
role: 'user',
content: generatePromptWithAspectRatio(this, i, prompt),
},
];
}
else {
let imageUrl = '';
if (imageSource === 'url') {
imageUrl = this.getNodeParameter('imageUrl', i);
}
else {
const binaryPropertyName = this.getNodeParameter('binaryPropertyName', i);
const binaryData = this.helpers.assertBinaryData(i, binaryPropertyName);
if (binaryData.mimeType && !binaryData.mimeType.includes('image/')) {
throw new Error('The provided binary data is not an image');
}
if (binaryData.data) {
const dataBuffer = Buffer.from(binaryData.data, 'base64');
imageUrl = `data:${binaryData.mimeType};base64,${dataBuffer.toString('base64')}`;
}
else if (binaryData.url) {
imageUrl = binaryData.url;
}
else {
throw new Error('No usable image data found in the provided binary property');
}
}
messages = [
{
role: 'user',
content: [
{
type: 'image_url',
image_url: {
url: imageUrl,
},
},
{
type: 'text',
text: generatePromptWithAspectRatio(this, i, prompt),
},
],
},
];
}
}
else {
try {
const previousMessages = JSON.parse(this.getNodeParameter('previousMessages', i));
if (!Array.isArray(previousMessages)) {
throw new Error('Previous messages must be a valid JSON array');
}
messages = [...previousMessages];
messages.push({
role: 'user',
content: prompt,
});
}
catch (error) {
throw new Error(`Failed to parse previous messages: ${error.message}`);
}
}
const body = {
model,
messages,
stream: true,
};
try {
const response = await GenericFunctions_1.llmApiRequest.call(this, body);
let taskResult = response;
if (waitForCompletion && ((_a = response.data) === null || _a === void 0 ? void 0 : _a.task_id)) {
taskResult = await GenericFunctions_1.waitForTaskCompletion.call(this, response.data.task_id);
}
taskResult.messages = messages;
if (taskResult.choices && taskResult.choices[0] && taskResult.choices[0].message) {
taskResult.messages.push(taskResult.choices[0].message);
}
returnData.push({
json: taskResult,
});
}
catch (error) {
if (this.continueOnFail()) {
returnData.push({
json: {
error: error.message,
},
});
continue;
}
throw error;
}
}
return [returnData];
}
}
exports.LLMConversationalImage = LLMConversationalImage;
//# sourceMappingURL=LLMConversationalImage.node.js.map