the big cats, not the cars
This project mainly consists in a Twitter Bot that searches for, stores and posts jaguar pictures. Written in Node.JS, it is constantly being updated and improved. The implementation brought along many features not proper to a Twitter Bot, making this project more than that.
- Twitter Bot
- Web Scraper
- Computer Vision
- Cloud Storage (not used anymore)
- Species Recognition Neural Network (TBD)
Albeit the project is far from completed, a release is intended to happen simply because the main basic bot features are ready and can be useful to other projects. It is a project of great importance for the author and, as such, needs to be showcased.
- Twitter API
- Bing Image Search JSON API
- Azure Storage Service REST API (not used anymore)
- Image Downloader
- Node Twitter
- MS Cognitive Services Computer Vision Client Library
This project was deployed on Heroku for testing and for learning purposes. It won't be deployed directly from GitHub anymore to avoid releasing API Keys.
The bot is now online and fully operating on its own.
The progress can be tracked on this GitKraken Board or on the GitHub Project Board for the repository. The result can be tracked on this Twitter account
The bot can search for images based on a word of phrase and store its URLs in a JSON file.
The stored URLs then undergo a "filtering" process using the CV API to determine whether the picture represents a Jaguar car or a Jaguar cat, and "select" the URLs for pictures of Jaguar cats. The filter will be improved to also reject content related to sports teams.
When a "selected" URL is to be posted, it is separated from the unused URLs and uploaded to Azure, then it is posted. This will be useful to compare the image to be posted with recently posted ones, avoiding close duplicates. (Not used anymore)
The bot already has many implemented working functions that may be useful for other projects, a tutorial will be compiled when the project is farther in development.