init prd prompt

backend/src/build-system/hanlder-manager.ts

@@ -1,5 +1,6 @@
 import { ProjectInitHandler } from './node/project-init';
 import { BuildHandler } from './types';
+import { PRDHandler } from './node/product_requirements_document/prd';
 
 export class BuildHandlerManager {
   private static instance: BuildHandlerManager;
@@ -10,7 +11,10 @@ export class BuildHandlerManager {
   }
 
   private registerBuiltInHandlers() {
-    const builtInHandlers: BuildHandler[] = [new ProjectInitHandler()];
+    const builtInHandlers: BuildHandler[] = [
+      new ProjectInitHandler(),
+      new PRDHandler(),
+    ];
 
     for (const handler of builtInHandlers) {
       this.handlers.set(handler.id, handler);

backend/src/build-system/node/product_requirements_document/prd.ts

@@ -0,0 +1,58 @@
+import { BuildHandler, BuildResult } from 'src/build-system/types';
+import { BuilderContext } from 'src/build-system/context';
+import { prompts } from './prompt/prompt';
+import { ModelProvider } from 'src/common/model-provider';
+import { StreamStatus } from 'src/chat/chat.model';
+import * as fs from 'fs';
+import * as path from 'path';
+
+export class PRDHandler implements BuildHandler {
+  readonly id = 'op:PRD::STATE:GENERATE';
+
+  async run(context: BuilderContext): Promise<BuildResult> {
+    console.log('Generating PRD...');
+
+    // Extract project data from the context
+    const projectName =
+      context.getData('projectName') || 'Default Project Name';
+    const description = context.getData('description') || 'Default Description';
+    const platform = context.getData('platform') || 'Default Platform';
+
+    // Generate the prompt dynamically
+    const prompt = prompts.generatePRDPrompt(
+      projectName,
+      description,
+      platform,
+    );
+
+    // Send the prompt to the LLM server and process the response
+    const prdContent = await this.generatePRDFromLLM(prompt);
+
+    // Save the PRD content to context for further use
+    context.setData('prdDocument', prdContent);
+
+    return {
+      success: true,
+      data: prdContent,
+    };
+  }
+
+  private async generatePRDFromLLM(prompt: string): Promise<string> {
+    const modelProvider = ModelProvider.getInstance();
+
+    const model = 'gpt-3.5-turbo';
+
+    // Call the chat method with the model specified
+    const chatStream = modelProvider.chat(
+      {
+        content: prompt,
+      },
+      model,
+    ); // Pass the model here
+
+    const prdContent = modelProvider.chunkSync(chatStream);
+
+    console.log('Received full PRD content from LLM server.');
+    return prdContent;
+  }
+}

backend/src/build-system/node/product_requirements_document/prompt/prompt.ts

@@ -0,0 +1,72 @@
+// Define and export the system prompts object
+
+export const prompts = {
+  generatePRDPrompt: (
+    projectName: string,
+    description: string,
+    platform: string,
+  ): string => {
+    return `You are an expert Product Manager. Your job is to analyze and expand upon the details provided, generating a Full Product Requirement Document (PRD) based on the following inputs:  
+       - Project name: ${projectName}
+       - Description: ${description}
+       - Platform: ${platform}
+
+ Follow these guidelines to ensure clarity, thoroughness, and usability in the PRD, which will be directly utilized in development.
+ ### Instructions and Rules:
+
+ 1, Your need to analysis the requirement for the project and covered all the things you can think about.
+ 2, You should focus on core features for the project
+ 3, ask yourself:
+ - what are all expect features require for a {project name} and {Description}?
+ - what are the user stories for the feature?
+ - what are the details description for those features?
+ - Am I cover all expect features ? if not then add new feature.
+ - Are those covered all user expected features ? if not then add new feature.
+ - what are the target audience for the project? Is all the features meet their expectation? if not then add new feature.
+ - Ask your self what are the function requirement and none functional requirement for those features?
+ - Are all features could be agree by others in the team? including product manager, developer....
+
+ ### PRD Structure:
+
+ Start with the following structure. Do not include sections for Milestones, Deliverables, or Technical Requirements.
+
+ ---
+
+### Product Requirement Document
+
+#### 1. Project Overview
+  - **Project Name**: 
+  - **Description**: Provide a brief overview of the project’s objective and purpose.
+  - **Platform**: Indicate the platform(s) (e.g., Web, Mobile).
+
+#### 2. Goals and Objectives
+  - Define the primary goals and purpose of the project.
+  - Describe any key performance indicators or success criteria.
+
+#### 3. Target Audience
+  - Identify the intended users or customer segments.
+  - Describe user needs, pain points, and how the product will address them.
+
+#### 4. User Stories
+  - List user stories that illustrate interactions for each main feature.
+  - Make each story actionable, focusing on user goals.
+
+#### 5. Features
+  - Outline each feature required for the project.
+  - Write the requirement for the feature
+
+#### 6. Functional Requirements
+  - Specify the functional requirements for each feature and sub-feature.
+  - Use clear, concise statements outlining required functionality.
+
+#### 7. Non-Functional Requirements
+  - Describe performance, security, usability, and other quality-related requirements.
+
+#### 8. Additional Analysis
+  - Provide any further insights or considerations to ensure a holistic view of the project requirements.
+---
+
+ Your reply must start with : "\`\`\`ProductDoc" and end with "\`\`\`". Be thorough, and make sure each section is fully developed and ready for implementation.
+`;
+  },
+};

backend/src/build-system/node/project-init.ts

@@ -8,7 +8,9 @@ export class ProjectInitHandler implements BuildHandler {
   async run(context: BuilderContext): Promise<BuildResult> {
     console.log('Setting up project...');
     const result = {
-      projectName: 'example',
+      projectName: 'online shoping',
+      descreption: 'sell products',
+      Platform: 'Web',
       path: '/path/to/project',
     };
     context.setData('projectConfig', result);

backend/src/chat/chat.service.ts

@@ -13,7 +13,6 @@ import {
 import { CustomAsyncIterableIterator } from 'src/common/model-provider/types';
 import { ModelProvider } from 'src/common/model-provider';
 
-
 @Injectable()
 export class ChatProxyService {
   private readonly logger = new Logger('ChatProxyService');

backend/src/common/model-provider/index.ts

@@ -51,7 +51,7 @@ export class ModelProvider {
 
   chat(
     input: ChatInput | string,
-    model?: string,
+    model: string,
     chatId?: string,
   ): CustomAsyncIterableIterator<ChatCompletionChunk> {
     const chatInput = this.normalizeChatInput(input);
@@ -199,6 +199,19 @@ export class ModelProvider {
     }
   }
 
+  async chunkSync(chatStream: AsyncIterableIterator<any>): Promise<string> {
+    let aggregatedContent = '';
+    for await (const chunk of chatStream) {
+      if (chunk.status === StreamStatus.STREAMING) {
+        aggregatedContent += chunk.choices
+          .map((choice) => choice.delta?.content || '')
+          .join('');
+      }
+    }
+    this.logger.log('Aggregated content from chat stream:', aggregatedContent);
+    return aggregatedContent;
+  }
+
   private startChat(input: ChatInput, model: string, chatId?: string) {
     const payload = this.createRequestPayload(input, model, chatId);