Business Model

Project Documentation

Overview

This documentation provides an overview of the project that utilizes Google Cloud services. The project aims to provide a web interface where users can upload memes with text. The frontend sends the image to the backend, which then sends it to a service that uses Weaviate to search for images without text in a database. Additionally, the backend also sends the image to a Google Function, which stores the image in a bucket. There is also a cron job that initiates an analysis of the images uploaded by a user. User authentication is handled using Firebase. The frontend is built using Redux, while the backend is implemented in Node.js with Express.

The project architecture consists of the following components:

Architecture

1. Frontend: The user interacts with the web interface, where they can upload a meme image with text. The frontend is developed using Redux, a predictable state container for JavaScript apps.

2. Backend: The backend handles the requests from the frontend and coordinates the flow of data between different services. It is implemented using Node.js and Express, a web application framework for Node.js.

3. Image Search Engine The image search engine is hosted on a Google Engine. It consists of 3 docker images:

   • weaviate: An open-source search engine that is used to search for images without text in a database.
   • img2vec-pytorch: A docker image that is used to convert images to vectors.
   • web-server: A web server that is used to handle requests from the backend.

   Container management is done using docker compose:

   version: "3.4"
   services:
   weaviate:
       command:
       - --host
       - 0.0.0.0
       - --port
       - "8080"
       - --scheme
       - http
       image: semitechnologies/weaviate:1.19.6
       ports:
       - 8383:8080
       restart: on-failure:0
       environment:
       IMAGE_INFERENCE_API: "http://i2v-neural:8080"
       QUERY_DEFAULTS_LIMIT: 25
       AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: "true"
       PERSISTENCE_DATA_PATH: "/var/lib/weaviate"
       DEFAULT_VECTORIZER_MODULE: "img2vec-neural"
       ENABLE_MODULES: "img2vec-neural"
       CLUSTER_HOSTNAME: "node1"
   i2v-neural:
       image: semitechnologies/img2vec-pytorch:resnet50
       environment:
       ENABLE_CUDA: "0"
   search-engine:
       build:
       context: .
       dockerfile: Dockerfile
       ports:
       - 8081:9200
       restart: on-failure:0
       depends_on:
       - weaviate
       - i2v-neural
       environment:
       WEAVIATE_HOST: "weaviate:8080"

4. Google Cloud Function: A Google Cloud Function is triggered by the backend whenever an image is uploaded. It receives the image from the backend and stores it in a Google Cloud Storage bucket.

5. Cron Job: A cron job is scheduled to run periodically and analyze the images uploaded by a user. It performs any required processing or analysis tasks on the images.

6. User Authentication: Firebase Authentication is used to handle user authentication and authorization. It provides secure user login and registration functionality.

Technologies Used

The project utilizes the following technologies:

• Google Cloud: Provides various services, including Google Cloud Storage and Google Cloud Functions, used for image storage and processing.

• Weaviate: An open-source search engine that is used to search for images without text in a database.

• Firebase Authentication: A service provided by Firebase that handles user authentication and authorization.

• Redux: A predictable state container for JavaScript apps, used for managing the frontend state.

• Node.js: A JavaScript runtime environment used for the backend implementation.

• Express: A fast and minimalist web application framework for Node.js, used to build the backend API.