vikaswakde commented 1 month ago

Implement Automated Website Analysis and Repair System

one possible fix for #24

Overview

This PR introduces an automated system for analyzing and repairing deployed websites. The system uses image analysis and AI-assisted code generation to detect and fix common issues, ensuring higher quality and consistency in our generated websites.

Approach

Capture screenshots after deployment (using Puppeteer)
Analyze the screenshots (using Google Cloud Vision API)
Repair the website if issues are detected (using a repair service)
Re-deploy the website after repairs
Capture new screenshots after repairs
Re-analyze the website and finalize deployment

Key Changes

1. Screenshot Capture (server/services/screenshotService.js)

Implemented Puppeteer for capturing full-page and element-specific screenshots
Screenshots are saved in a project-specific folder for analysis

2. Image Analysis (server/services/imageAnalysisService.js)

Integrated Google Cloud Vision API for detecting objects, text, and labels in screenshots
Implemented logic to identify common issues like missing headers, navigation, or content

3. Website Repair (server/services/repairService.js)

Created a repair service that generates fixes for detected issues using AI-assisted code generation
Handles various types of issues including missing elements and layout problems

4. Deployment Process (server/controllers/ctoToolController.js)

Updated the deployment process to include the analysis and repair steps
Implemented re-deployment after repairs are made

5. Google Cloud Authentication (server/config/googleCloudAuth.js)

Added secure handling of Google Cloud credentials using environment variables

ptmaroct commented 1 month ago

@vikaswakde @ms4n is already working on this, you can meet him in office and collaborate together on this