ManualTrigger Automate

Target Audience

• Digital Marketers: Those looking to analyze and optimize images for campaigns.
• Data Scientists: Professionals interested in image processing and embedding techniques.
• Developers: Individuals building applications that require image analysis and vector search capabilities.
• Content Creators: Creators who want to enhance their images with metadata for better discoverability.

Problem Solved

This workflow addresses the challenge of extracting meaningful insights from images by automating the process of:

• Downloading images from Google Drive.
• Analyzing color information and generating semantic keywords.
• Creating an embedding document for images, enabling efficient vector searches in databases.

Workflow Steps

1. Manual Trigger: The workflow begins when the user clicks the "Test workflow" button.
2. Google Drive Integration: The workflow fetches an image file from Google Drive for processing.
3. Image Analysis: The image undergoes two main analyses:
   • Color Information Extraction: Retrieves color channel statistics from the image.
   • Image Resizing: Resizes the image to 512x512 pixels if it is larger.
4. Keyword Generation: Generates a list of semantic keywords describing the image using OpenAI's vision model.
5. Combining Analysis Results: Merges the color information and generated keywords into a single document.
6. Embedding Document Creation: Constructs a document that includes metadata for further use.
7. In-Memory Vector Store: Inserts the embedding document into an in-memory vector store for quick retrieval.
8. Search Functionality: Provides a mechanism to search for images based on user-defined prompts, leveraging the stored embeddings.